Fintech, Digital Finance
& DeFi
13 sessions covering FinTech disruption, blockchain, DeFi, payments, lending, robo-advisors, regulation and scalability — with 10 case studies, 7 interactive simulators, and a 130+ question quiz bank.
Session Map
All 13 SessionsFinTech Introduction
Landscape, evolution, disruption forces. Case: Cutting Through the Fog.
Blockchain & Bitcoin
Distributed ledgers, proof-of-work, crypto. Case: Bitcoin Digital Payments.
DeFi
Decentralised finance, AMMs, stablecoins, yield protocols. Reading: Awakening the Blockchain.
Big Data & Analytics
New data sources, credit scoring, AI. Case: KUESKI Mexico.
Payments
Payment rails, two-sided markets, strategy. Case: PayPal Merchant Services.
Future of Payments
Cashless societies, neobanks, BNPL. Case: Nubank.
P2P Lending
Marketplace lending, credit risk, regulation. Case: Lending Club.
Robo-Advisors
Algorithmic wealth management, democratisation. Case: Wealthfront.
Building a FinTech
Pain points, team, go-to-market, VC funding, unit economics.
Financial Inclusion
Mobile banking, unbanked populations. Case: Mobile Bank for the Unbanked.
Regulation
Regulatory frameworks, sandboxes, compliance. Case: Lending Loop.
Scalability
Exponential growth, platform dynamics. Case: Tencent.
Bank Partnerships
FinTech-bank collaboration, innovation strategy. Case: Scotiabank.
Evaluation & Assessment
Course Structure| Component | Weight | Format |
|---|---|---|
| Final Group FinTech Pitch | 40% | Group project — build a FinTech concept & pitch |
| Individual Case Write-ups | 30% | Written analysis submitted per case session |
| Class Participation | 30% | In-session discussion, questions, debate |
FinTech Introduction: Reshaping the Financial Sector
From Bill Gates' 1994 "banks are dinosaurs" declaration to a $4.7 trillion disruption opportunity. Who are the players, what are the battlegrounds, and what should incumbents do?
Key Numbers
Session 01 DataCast of Characters
Case: Cutting Through the FogWho's who in the FOG (Florida Optimum Group) case
The FinTech Landscape
ContextThe case opens with a deliberately provocative observation: FinTech is easier to describe by what it is not than by what it is. Several competing definitions from the case:
| Source | Definition / Characterisation |
|---|---|
| Common usage | Technology startups disrupting mobile payments, transfers, loans, fundraising, and asset management |
| Industry insider | "The currently held definition is stale and unrepresentative of the opportunity" — the real opportunity is technology serving the clients of financial institutions |
| PwC | "It is the novel application of technology and its speed of evolution that make the current wave unlike any we have seen before" |
| Competitive lens | "FinTechs have two unique selling points: better use of data and frictionless customer experience" |
| 1994 | Bill Gates: "Banks are dinosaurs" — the first wave of digital finance vision, pre-internet scale |
| 1998 | PayPal founded — first major FinTech company, proving internet payments were viable |
| 2005 | Y Combinator's first FinTech: TextPayMe (SMS payments, acquired by Amazon 2006). First significant startup wave begins |
| 2008 | Financial crisis destroys trust in banks. Millennial generation becomes receptive to alternatives. Satoshi publishes Bitcoin whitepaper |
| 2013-15 | Explosive growth: YC FinTech startups double. Funding reaches $11.2B in just 9 months of 2015 — nearly double full-year 2014 |
| 2016 | Case setting. Still only 0.7% penetration in US market — but Goldman estimates $4.7T revenue at risk. Incumbents must decide |
Figure 8 from the case maps bank business lines by likelihood and extent of disruption. Three key bank vulnerabilities identified:
| Business Line | Disruption Likelihood | Disruption Extent | Leading FinTech Attackers |
|---|---|---|---|
| Personal Loans | High | High | Lending Club, SoFi, Kabbage |
| Digital Payments | High | High | PayPal, Square, Stripe, Alipay |
| SMB Loans | High | Medium | OnDeck, Funding Circle, Kabbage |
| Wealth Management | Medium | Medium | Wealthfront, Betterment, SigFig |
| International Remittances | High | High | TransferWise, Xoom, Bitcoin |
| Deposits | Medium | Low | Limited — regulatory moat remains strong |
| Mortgage | Lower | Low-Medium | Complex product protects incumbents near-term |
The Four Strategic Options
Case Decision FrameworkThe counter-evidence Costa finds is damning: digital channels already dominate interactions (internet banking 10+ times/month vs branch 1-2x). Citi's own internal data shows workforce decline of 20%+ expected over 10 years. A "wait and see" posture risks being caught unprepared when threat becomes imminent.
Pros: Immediate access to technology and talent. Proven with examples: Capital One acquired Adaptive Path (UX), Bundle (spending tracker). BBVA acquired Simple. BlackRock acquired FutureAdvisors.
Cons: Pure tech giants (GAFA, Alibaba) are un-acquirable. Integration culture clash is severe — banks and startups have fundamentally different "fail fast" vs risk-averse cultures. Overpaying for hype valuations is likely at peak cycle.
| Acquirer | Target | Strategic Rationale |
|---|---|---|
| Capital One | Adaptive Path, Level Money, Bundle | UX/design capability + personal finance analytics |
| BBVA | Simple (digital bank) | US digital banking presence — but integration challenged |
| BlackRock | FutureAdvisors | Robo-advisory capability for institutional distribution |
| Santander | Cyanogen (mobile OS) | Mobile platform positioning |
The boldest option: transform the entire bank into a technology-first institution. Banks do spend heavily on IT — global bank IT expenses were projected to reach $210B+ by 2017 (North America $70B, Europe $78B, Asia-Pacific $78B). Swedbank's chairman declared "we are an IT company" over a decade before this case.
The Citi-Lending Club partnership is the case's flagship example of this model in action. The division of labour is explicit: Lending Club is better at sourcing and algorithmically evaluating loans; Citi has the deposits to fund them. Each does what it does best. The partnership captures the benefits of FinTech without the integration risks of acquisition.
| Bank | FinTech Partner | Partnership Type |
|---|---|---|
| Citi | Lending Club + Varadero | $150M loan programme for underserved communities |
| BancAlliance (200 community banks) | Lending Club | Co-branded personal loans through LC platform |
| Wells Fargo | Multiple tech startups | Innovation lab — non-banking tech with banking potential |
| Barclays / Santander | Various FinTechs | Fintech accelerators / VC structures (Santander InnoVentures) |
Strategic Frameworks & Exam Angles
Key ConceptsClayton Christensen's framework is the intellectual backbone of Session 1. A bank's entire organisation — its incentive structures, processes, talent profiles, culture, and reporting lines — is optimised for its current high-margin customer base. Investing in lower-margin digital products that would cannibalise existing revenue is rational in the short term and suicidal in the long term.
As of December 2014, FinTech companies traded at a significant premium to the S&P 500 median (19.6x forward P/E):
| FinTech Segment | P/E (LTM) | P/E 2015E | EV/EBITDA (LTM) | EV/Revenue |
|---|---|---|---|---|
| FinTech-Payments | 27.7x | 23.8x | 13.0x | 2.3x |
| FinTech-Solutions | 29.9x | 25.0x | 16.7x | 3.0x |
| FinTech-Technology | 24.4x | 28.2x | 15.6x | 4.0x |
Blockchain & Bitcoin: Much More Than Just Crypto
The distributed ledger breakthrough, proof-of-work mechanics, and why blockchain's potential extends far beyond cryptocurrency into trade finance, identity, and settlement.
Key Numbers
Session 02 DataCast of Characters
Case: Bitcoin Digital Payments (HBS)Who's who in the Bitcoin ecosystem
How Bitcoin Actually Works
Technical FoundationBefore Bitcoin, digital cash was impossible to create without a trusted intermediary. The core problem: a digital file can be copied infinitely. If I send you a digital "coin," what stops me from also sending the same coin to someone else? Every previous attempt required a central mint to prevent this.
| Dimension | Bitcoin | Visa/Mastercard | Cash |
|---|---|---|---|
| Transactions/second | ~7 TPS | ~24,000 TPS | N/A |
| Settlement time | 10-60 min (confirmations) | 2-3 days | Instant |
| Transaction fee | Near zero (2014) | 1.5-3% | Zero |
| Cross-border cost | Same as domestic | High (FX + wire) | Very high (remittance) |
| Volatility | Extreme (-98.5% deflation in 11 months of 2013 in BTC terms) | Stable | Stable |
| Consumer protection | None (irreversible) | Strong (TILA rights) | None for theft |
| Chargeback risk (merchant) | Zero | High | None |
The Regulatory Question: Currency, Commodity or Security?
Case Central Issue| Classification | Regulator | Key Implication | Consequence for Bitcoin |
|---|---|---|---|
| Currency | Central Banks / Treasury | Subject to monetary policy, KYC/AML, capital controls | Most countries: treat as foreign currency for tax. Manageable but adds friction |
| Commodity | CFTC (US) | Derivative contracts possible. Producers taxed on mining gains | Most benign outcome for industry — CFTC's actual position in US |
| Security | SEC | Howey Test: investment + common enterprise + profit from others' efforts = security | Most disruptive: registration requirements, accredited investor limits, exchange licensing |
BitPay's value proposition: merchants accept Bitcoin, BitPay instantly converts to USD and deposits the next business day. Merchant gets all the benefits (lower fees, no chargebacks, global acceptance) with none of the Bitcoin volatility risk.
The session title — "Much More Than Just Bitcoin" — is a deliberate signal. Blockchain's distributed ledger technology has applications wherever there are multiple parties sharing the same data and needing a single, trusted version of truth:
| Use Case | Problem Solved | Example Players |
|---|---|---|
| Trade Finance | Letters of credit take 5-10 days on paper — blockchain reduces to hours | HSBC, we.trade (R3), Contour |
| Cross-border Settlement | SWIFT takes 1-5 days; blockchain enables near-real-time settlement | Ripple (XRP), JPM Coin |
| Digital Identity | Single verified identity shared across institutions — eliminates repeated KYC | Sovrin, uPort, Microsoft ION |
| Supply Chain | Immutable provenance tracking from origin to consumer | IBM Food Trust, Maersk TradeLens |
| Securities Settlement | T+2 settlement could become T+0 — eliminating counterparty risk window | DTCC, ASX CHESS replacement |
DeFi: Decentralised Finance
How smart contracts on Ethereum enable lending, exchange, and yield without banks, intermediaries, or identity requirements — and the risks this creates.
Key Numbers
Session 03 DeFi DataThe DeFi Stack: Three Core Applications
HBS Note StructureStablecoins solve DeFi's fundamental usability problem: crypto is too volatile to be useful as a unit of account or medium of exchange. Three types exist:
| Type | Mechanism | Example | Risk |
|---|---|---|---|
| Fiat-backed | 1:1 USD held in bank reserves | USDC, Tether (USDT) | Counterparty/bank risk. Trust in issuer's reserve claims |
| Crypto-collateralised | Over-collateralise with ETH, lock in vault, mint DAI | DAI (MakerDAO) | Liquidation cascade if ETH crashes. CR must stay above 150% |
| Algorithmic | Smart contract expands/contracts supply to maintain peg. Seigniorage model | TerraUSD (UST) — collapsed 2022 | Catastrophic: death spiral if confidence breaks (UST/LUNA -99.99%) |
Compound is a decentralised money market. No loan officers, no credit scores, no identity requirements. Anyone with crypto can lend or borrow algorithmically.
Yield Farming & Governance Tokens
Advanced DeFiYield farming is the practice of moving capital across DeFi protocols to maximise return — a strategy analogous to a currency carry trade. The HBS note explicitly uses this analogy: "arbitrageurs take advantage of potential mispricing... borrow at a low interest rate and invest in another that yields a higher return."
Governance tokens (MKR, COMP, UNI) resemble equity: they represent percentage ownership of the protocol and confer voting rights on protocol parameters. But they are fundamentally different from corporate equity:
| Dimension | Corporate Equity | Governance Token |
|---|---|---|
| Cash flow rights | Dividends, buybacks | Protocol fee revenue (varies by design) |
| Voting rights | One share = one vote (typically) | One token = one vote = plutocracy risk |
| Legal protection | Securities law, fiduciary duties | None in most jurisdictions |
| Regulatory status | Clearly a security | Ambiguous — SEC scrutiny ongoing |
| Dilution risk | Requires shareholder vote | Protocol can mint new tokens via governance vote |
| Risk Type | Description | Example Event |
|---|---|---|
| Smart Contract Risk | Bugs in immutable code can drain protocol funds | The DAO hack (2016): $60M drained — led to Ethereum hard fork |
| Liquidation Cascade | Rapid price falls trigger liquidations, which depress price further — self-reinforcing spiral | March 2020 "Black Thursday": ETH -50% in hours, mass MakerDAO liquidations |
| Oracle Manipulation | DeFi protocols rely on price feeds from oracles — manipulable via flash loans | Multiple flash loan attacks 2020-21 exploiting oracle price manipulation |
| Algorithmic Stablecoin Failure | If confidence in peg breaks, death spiral is instantaneous | TerraUSD/LUNA collapse May 2022: $40B evaporated in 72 hours |
| Governance Attack | Attacker borrows governance tokens, passes malicious proposal | Beanstalk Farms 2022: $182M drained via governance flash loan attack |
Lecture Slides: DeFi & the Banking System
Session 3 SlidesBanks engaging with crypto — whether through custody, lending, or market-making — face five structurally distinct risk categories. Each maps to a different part of the balance sheet and regulatory framework:
| Risk Category | What It Means in Practice | Example Trigger |
|---|---|---|
| Operational | Custody of private keys, irreversibility of crypto transactions, heightened fraud and cyber-attack exposure | Key compromise → irrecoverable asset loss; no central authority to reverse |
| Legal | Ambiguous regulatory classification across jurisdictions; AML/KYC compliance gaps in pseudonymous networks | Regulator reclassifies crypto holdings as securities → capital surcharge |
| Market / Balance Sheet | Crypto volatility passes through to asset values; mark-to-market losses on holdings or collateralised loans | BTC falls 50% → collateral backing crypto-backed loans drops below thresholds |
| Liquidity / Funding | Deposit migration to stablecoins or crypto platforms reduces stable funding; rapid outflows in stress events | Stablecoin yield spike draws retail deposits away from bank accounts |
| Reputational / Strategic | Association with crypto collapses damages brand; failure to engage loses the next generation of customers | Bank seen as FTX custodian → reputational contagion; bank ignoring DeFi → loses digital-native clients |
Stablecoins represent a structural threat to bank funding that goes beyond price volatility. The mechanism is deposit substitution: retail and institutional holders park liquidity in USDC or USDT instead of bank accounts — reducing the stable deposit base that banks rely on for cheap funding.
| Dimension | Bank Deposit | Stablecoin (e.g. USDC) |
|---|---|---|
| Yield to holder | Near-zero (retail) / SOFR-linked (institutional) | DeFi lending rates (historically 3–10%) |
| Regulatory protection | FDIC-insured up to $250K (US) / DGS in EU | No deposit insurance |
| Funding cost to bank | Low (deposits are cheap) | N/A — bank loses the deposit entirely |
| AML / KYC | Full KYC required | Pseudonymous; on-chain analytics only |
| Stress scenario | FDIC backstop prevents run | Peg break can trigger instant digital bank run |
Banks have offered securities custody for over a century. Crypto custody looks superficially similar — holding assets on behalf of clients — but the underlying mechanics create entirely new risk exposures:
| Dimension | Traditional Securities Custody | Crypto Custody |
|---|---|---|
| What is "held" | Legal claim on shares held at CSD (e.g. DTCC, Euroclear) | Private cryptographic key — mathematical proof of ownership |
| Transaction reversibility | T+2 settlement; errors correctable | Irreversible once confirmed on-chain — no recourse |
| Key loss scenario | Share register can re-issue certificates | Lost private key = permanently lost assets (no recovery mechanism) |
| Regulatory framework | Mature: MiFID II, Rule 15c3-3, UCITS | Nascent: SAB 121 (US) requires on-balance-sheet recognition — capital cost |
| Cyber risk surface | Centralised — one target but hardened | Keys are bearer instruments — stolen = gone; HSM / cold storage required |
Blockchain for Cross-Border Payments
S3 Slides 7–9International wire transfers today rely on the SWIFT correspondent banking network — a chain of bilateral relationships that adds cost, delay, and opacity at every hop. Blockchain collapses this stack:
| Dimension | SWIFT Correspondent | Blockchain Settlement |
|---|---|---|
| Settlement time | 1–5 business days | Seconds to minutes |
| Cost (sender) | $25–$50 + 1–3% FX | <$0.01 on-chain + FX |
| Liquidity requirement | Pre-funded Nostro accounts in every corridor | On-demand — no pre-funding needed |
| Transparency | Opaque — no real-time tracking | Full on-chain audit trail |
| Counterparty risk | Each correspondent adds counterparty risk | Protocol / validator risk instead |
| Regulatory compliance | Mature AML / sanctions screening | On-chain analytics still maturing; travel rule compliance complex |
Digital Assets Regulation
S3 Slides 10–12Regulators are not trying to kill crypto — they are trying to manage its externalities. The slide deck identifies five core risk categories that drive the regulatory agenda globally:
| Risk Category | Regulator Concern | Primary Tool |
|---|---|---|
| Consumer / Investor Protection | Retail investors buying highly volatile or fraudulent assets without adequate disclosure | Prospectus requirements, suitability rules, exchange licensing |
| Financial Stability | Large-scale crypto collapses spilling into the real economy via bank contagion or stablecoin runs | Systemic designation, reserve requirements for stablecoins, bank crypto capital rules |
| Market Integrity | Wash trading, front-running, insider trading on unregulated exchanges | Market abuse rules extended to crypto (MiCA Art. 90+); exchange surveillance |
| AML / CFT | Use of pseudonymous transactions to launder proceeds or finance terrorism | Travel Rule (FATF), VASP registration, on-chain analytics (Chainalysis) |
| Tax Evasion | Unreported gains and cross-border crypto income streams | DAC8 (EU): automatic exchange of crypto tax data from 2026; IRS Form 1099-DA (US) |
The Future of Crypto: CBDCs, Legal Tender & Big Tech
S3 Slides 13–16CBDCs are the state's response to private digital money — a digital form of central bank liability that is programmable, traceable, and directly issued to citizens or institutions. They co-opt blockchain's benefits while preserving monetary sovereignty:
| Dimension | Decentralised Crypto (BTC/ETH) | Stablecoin (USDC) | CBDC (e.g. Digital Euro) |
|---|---|---|---|
| Issuer | No issuer (protocol) | Private company (Circle) | Central Bank |
| Legal tender | No | No | Yes |
| Monetary policy lever | None | None | Programmable (negative rates, expiry dates, spending constraints) |
| Privacy | Pseudonymous | KYC at on/off ramp | Full transaction visibility to state |
| Bank disintermediation risk | Partial | Partial | Severe — retail CBDC held directly at CB bypasses commercial banks |
In September 2021, El Salvador became the first country to adopt Bitcoin as legal tender alongside the US dollar — requiring all merchants to accept it. The experiment has been closely watched as a live stress test of crypto as a monetary system:
| Dimension | Outcome |
|---|---|
| Remittance cost reduction | Potential to cut $400M/year in remittance fees (remittances = 24% of GDP) — the primary use case |
| Adoption by population | Chivo wallet distributed with $30 bonus — initial uptake driven by incentive, not organic use |
| Merchant acceptance | Large majority of merchants still refuse Bitcoin in practice; law poorly enforced |
| IMF / bond market reaction | IMF refused $1.3B loan until Bitcoin legal tender status was modified; sovereign bond spreads widened |
| Volatility management | BTC's 50–80% drawdowns are incompatible with price-setting for everyday goods — a fundamental unresolved problem |
| Government BTC holdings | El Salvador accumulated ~2,500 BTC — reported unrealised gains when price recovered above $60K |
Facebook's 2019 Libra announcement (later renamed Diem) was the most significant regulatory stress test the crypto industry has faced — and its rapid collapse reveals the limits of private digital money at scale:
Web 3.0: The Internet of Value
S3 Slides 17–20| Era | Period | Core Paradigm | Value Capture | Financial Analogue |
|---|---|---|---|---|
| Web 1.0 | 1991–2004 | Read-only. Static HTML pages. Users consume content. No interactivity. | Content publishers (media cos); ISPs | Electronic information delivery (Bloomberg terminals, online banking portals) |
| Web 2.0 | 2004–2020 | Read-write. User-generated content. Platforms aggregate users and monetise via data/advertising. | Platform companies (FAANG). Data is the product. | PayPal, Stripe, Robinhood — financial rails built on platform infrastructure. Network effects → winner-take-all |
| Web 3.0 | 2020–present | Read-write-own. Users own their data and assets via cryptographic proofs. Trustless protocols replace platforms. | Protocol token holders; liquidity providers; validators | DeFi protocols — financial logic encoded in smart contracts, no corporate intermediary. Ownership via tokens. |
Crypto Market Phases & the DeFi Ecosystem
S3 Slides 21–22The course material frames crypto's evolution as three distinct phases — each building on the last, each opening new use-cases and business models:
| Phase | Core Innovation | Representative Protocols | Financial Analogue |
|---|---|---|---|
| Phase 1: Digital Money | Censorship-resistant, peer-to-peer store of value and payment rail. Proof-of-work consensus. No programmability. | Bitcoin (BTC), Litecoin (LTC), Zcash (ZEC) | Digital gold / alternative currency. No financial services built on top. |
| Phase 2: Economy Platforms | Programmable blockchains. Smart contracts enable arbitrary logic. Turing-complete execution environments. NFTs, tokens, DAOs. | Ethereum (ETH), Cardano (ADA), Solana (SOL) | Operating systems for financial applications — the "App Store" for finance. Value in platform adoption. |
| Phase 3: Finance Ecosystems (DeFi) | Full financial stack on-chain. Lending, exchange, derivatives, insurance, asset management — without intermediaries. | MakerDAO, Yearn Finance, Uniswap, Aave, Compound, Curve | A parallel financial system: money markets, FX, wealth management — recreated as open-source protocols. |
Ethereum, Smart Contracts & DeFi Primitives
S3 Slides 23–24Ethereum (launched 2015) solved the limitation of Bitcoin's intentionally constrained scripting language by introducing a Turing-complete virtual machine — the EVM (Ethereum Virtual Machine). This allowed arbitrary programmes to run on a decentralised, tamper-proof computer:
| Concept | Definition | Financial Analogue | Key Exam Point |
|---|---|---|---|
| Smart Contract | A programme stored on a blockchain that self-executes the terms of an agreement — no intermediary, no discretion, code is law | A legal contract + an automatic enforcement mechanism, combined | Immutable once deployed — bugs cannot be patched without a new contract; upgrade paths require governance votes |
| DeFi (Decentralised Finance) | The ecosystem of financial applications built on public blockchains using smart contracts — permissionless, non-custodial, transparent | The entire financial system (banking, FX, wealth management) recreated as open-source software | Not regulated like banks: no FDIC, no KYC, no recourse — pure code risk. Over-collateralisation is the only credit mechanism. |
| DApp (Decentralised App) | An application whose backend logic runs on a blockchain smart contract rather than a centralised server | Think of it as a fintech app where the "bank" is replaced by a protocol — no corporate owner can shut it down | Front-end can still be centralised (a website); true decentralisation requires the smart contract to be the canonical interface |
| DAO (Decentralised Autonomous Org) | An organisation governed by smart contract rules and token-holder votes rather than a board of directors or management team | A company where the shareholders vote directly on every operational decision via their tokens — no CEO, no CFO | Legal status is ambiguous in most jurisdictions — DAOs may be treated as general partnerships with unlimited personal liability for token holders |
DeFi by the Numbers & CeFi vs DeFi
S3 Slides 25–32Total Value Locked (TVL) is DeFi's headline metric — the sum of all assets deposited into smart contracts across all protocols. It serves as a proxy for the ecosystem's economic footprint:
| DeFi Category | What It Does | Dominant Protocol | TVL Share (approx.) |
|---|---|---|---|
| Lending / Borrowing | Deposit collateral, borrow against it. Interest rates set algorithmically by utilisation ratio. | Aave, Compound, MakerDAO | ~50% — the largest category |
| DEX / AMM | Permissionless token exchange via liquidity pools. No order book; constant product formula sets price. | Uniswap, Curve, SushiSwap | ~25% |
| Yield / Asset Mgmt | Auto-compounds yield across protocols to maximise APY. Abstracts complexity for users. | Yearn Finance, Convex | ~10% |
| Derivatives | On-chain perpetual futures, options, synthetic assets (e.g. synthetic S&P 500 exposure) | dYdX, Synthetix, GMX | ~8% |
| Other (Bridges, NFT-Fi) | Cross-chain liquidity, NFT collateralisation, insurance protocols | Various | ~7% |
| Property | What It Means | Why It Matters |
|---|---|---|
| Permissionless | Anyone with a crypto wallet can access any DeFi protocol — no account opening, no KYC, no geography restrictions | 1.7B unbanked adults could theoretically access lending and savings without a bank account |
| Non-custodial | Users retain control of their assets at all times — the protocol never holds your keys | Eliminates custodial risk (FTX collapse: users lost assets held "at" the exchange); "Not your keys, not your coins" |
| Transparent | All protocol rules, interest rates, liquidity positions, and transactions are visible on-chain in real time | No hidden fees, no information asymmetry between protocol insiders and retail users — complete audit trail |
| Composable | DeFi protocols are "money legos" — any protocol can call any other's smart contract, enabling instant integration | A new protocol can instantly use Uniswap's liquidity, Compound's rates, and Chainlink's prices with no API agreements or business development |
| Programmable | Financial logic is code — interest rates, liquidations, governance rules execute automatically without human discretion | Flash loans (borrow and repay in one transaction) are only possible because execution is atomic and trustless |
| Dimension | CeFi (Centralised Finance) | DeFi (Decentralised Finance) |
|---|---|---|
| Control of funds | Exchange / bank holds assets ("custodial") | User holds via private keys ("non-custodial") |
| Identity | KYC/AML mandatory — government ID required | Pseudonymous — wallet address only |
| Access | Restricted by geography, credit history, legal status | Permissionless — internet connection + crypto wallet sufficient |
| Counterparty risk | Platform default risk (e.g. FTX, Celsius, BlockFi) | Smart contract risk (code bugs, exploits) |
| Regulatory status | Licensed — regulated entity with recourse | Unregulated in most jurisdictions — no recourse if funds lost |
| Product range | Full financial services: fiat on/off ramp, leverage, OTC, staking | Growing but limited: lending, DEX, yield, derivatives — no fiat |
| Transparency | Opaque — audits periodic, reserve proof rare (pre-Proof-of-Reserves) | Fully on-chain — all positions and transactions visible in real time |
| Innovation speed | Constrained by compliance, regulatory approval cycles | Permissionless — anyone can fork a protocol and deploy in hours |
Big Data & Analytics: The New "Oil"
How FinTechs use alternative data to score the previously un-scoreable, and the strategic, ethical and regulatory tensions this creates.
Key Numbers
Session 04 DataCast of Characters
Case: KUESKI Mexico (Kellogg)Who's who in Kueski's journey
The Alternative Data Playbook
Kueski's Core InnovationTraditional credit scoring (FICO in US, Buró de Crédito in Mexico) works by tracking repayment history on previous credit products. In Mexico, this creates a Catch-22: you need credit history to get credit, but you need credit to build history. The case documents Mexico's structural credit gap:
| Indicator | Mexico (2014) | Implication |
|---|---|---|
| Population with bank account | 39% | 61% have no formal financial relationship at all |
| Population with credit/debit card | 18% | 82% cannot make online purchases |
| Credit needs met by formal institutions | 25% | 75% rely on moneylenders, pawn shops, family |
| Informal economy participation | 59% | Cannot document income -> banks won't serve them |
| E-commerce lost to credit gap | $4B/year | Flores's original discovery — the business opportunity |
Kueski built a proprietary algorithm using hundreds of non-traditional data points to answer three questions: (1) Is this person who they say they are? (identity fraud) (2) Do they intend to repay? (willingness) (3) Can they repay? (capacity)
| Data Category | Specific Signals Used | What It Predicts |
|---|---|---|
| Behavioural / Device | Typing speed, device type, browser, application time, session duration, mouse movement patterns | Faster-than-normal typing = fraud; slower = late payment risk |
| Digital Reputation | Age of email account, email provider (Gmail vs Hotmail vs Prodigy), email usage patterns | Older email + premium provider correlates with repayment intention |
| Social Demographic | Geolocation, IP address history, social connections | Location stability, community embeddedness |
| Identity Verification | Photo ID match + selfie comparison | KYC / anti-fraud verification |
| Traditional Bureau | Buró de Crédito (accessed only for those WITH history) | Repayment history — used when available, not required |
The Bank Partnership Question
Case Central DecisionKUESKI is a specific case of a general trend: the application of Big Data and machine learning to transform financial services. The session uses KUESKI to illustrate four big-data value drivers in finance:
| Value Driver | Mechanism | Example |
|---|---|---|
| Personalisation | Individual pricing of risk replaces population-level averages | Kueski: each applicant scored individually vs bank's broad risk buckets |
| New Market Access | Previously unscoreable populations become serviceable | 61% unbanked Mexicans; thin-file Americans; gig economy workers |
| Fraud Detection | Real-time pattern recognition catches anomalies humans miss | Typing speed anomaly = fraud flag. IP address change = risk signal |
| Credit Cycle Prediction | Real-time monitoring of borrower behaviour enables early intervention | Payment behavior change -> proactive collections outreach |
Case Discussion: Model Answers
KUESKI — Class PrepFlores identified the pain point while trying to replicate Netflix's subscription model in Mexico: he realised the problem wasn't lack of demand for digital content, but lack of credit to pay for it. 61% of Mexicans had no bank account; 82% had no credit or debit card; $4B in annual e-commerce was lost because consumers couldn't transact online. The credit gap was the root constraint.
Critically, Flores did not build immediately. He spent six months testing four distinct credit concepts via landing pages — a lean startup approach to validating demand before committing engineering resources. Each concept tested a different customer segment and product form. Only after seeing which landing page drove the strongest conversion signal did he commit to online consumer credit. This is the case's most important strategic lesson: systematic hypothesis testing before scaling capital and talent.
| Dimension | Strengths | Risks / Where It Breaks |
|---|---|---|
| Data sources | Typing speed, device type, email age, IP address, session behaviour — captures signals invisible to traditional bureaus. Enables fast approval (minutes, not weeks). | These signals have not been validated through a full credit cycle. A benign economic environment inflates apparent predictive power — untested in recession where behaviours change. |
| Thin-file problem | Breaks the Catch-22: customers who have never had credit can be scored and given a first loan. Builds a credit relationship from zero. | First loan defaults are higher because there is genuinely no credit history. The model must absorb these costs while building the dataset — requires patient capital. |
| Proxy discrimination | Avoids explicitly using race, gender, or religion — complies with letter of fair lending law. | Email provider, device type, and IP location can correlate with protected characteristics. The model may discriminate by proxy without intent — a regulatory and ethical liability as scrutiny increases. |
| Fraud signals | Typing speed anomalies (too fast = bot/fraud; too slow = coached), browser/device mismatch, session duration — catches fraud patterns invisible to document review. | Fraudsters adapt. Once the signals are known or guessed, they can be gamed — particularly by organised fraud rings who reverse-engineer approval criteria through iterative testing. |
Scaling a lending business differs fundamentally from scaling a software business. Revenue is linear with loan volume, but risk is non-linear — concentrated credit deterioration can wipe out years of origination income. The conditions Flores must satisfy before scaling aggressively:
| Condition | What It Requires | Kueski's Status (at case) |
|---|---|---|
| CLV/CAC > 3x | Lifetime value of repeat customers must cover acquisition cost by a comfortable margin. First loan may lose money — repeat loans must be profitable. | Demonstrated in early cohorts but untested at scale and through a credit cycle. |
| Default rate stability | Alternative data model must produce consistent default rates across different origination vintages, not just the first cohort tested during model development. | Positive early signs but insufficient vintage data to be confident. |
| Marginal cost of underwriting declining | Each additional loan should cost less to underwrite (model improves with data), not more — confirming the data flywheel thesis. | Plausible but unproven at case writing date. |
| Capital source scalability | Can Kueski raise enough debt capital to fund the loan book as it scales? Without a bank licence or deposit base, it relies on equity and debt raises — expensive and uncertain. | Binding constraint — bank partnership directly addresses this. |
Lecture Slides: Big Data Foundations
S4 Slides 2–19Session 4 situates Big Data within the wider technology disruption landscape. Four converging forces are enabling FinTech to challenge traditional financial services:
| Technology | What It Enables in Finance | Session Coverage |
|---|---|---|
| 1. Big Data & Analytics | Credit scoring for thin-file customers; real-time fraud detection; personalised pricing; alternative data underwriting | Slides 4–27 — core topic of this session |
| 2. Artificial Intelligence | Automated credit decisions; chatbots; algorithmic trading; RegTech compliance; Generative AI for document generation | Slides 28–44 — increasingly intertwined with Big Data |
| 3. Cloud Computing | Elastic infrastructure for FinTechs without data-centre capital expenditure; enables API banking; real-time processing at scale | Slides 45–48 — the infrastructure layer |
| 4. Internet of Things (IoT) | Telematics for insurance pricing; in-car payments; connected device data as new underwriting signal | Slides 49–52 — the emerging data frontier |
| The "V" | Meaning | Financial Services Example |
|---|---|---|
| Volume | Scale of data generated — petabytes to exabytes | A large bank processes billions of transactions daily; each generates metadata: timestamp, location, device, merchant category, amount |
| Velocity | Speed at which data is generated and must be processed in real time | Fraud detection must evaluate a card transaction in <100ms — real-time streaming processing, not overnight batch |
| Variety | Diversity of data types — structured, semi-structured, unstructured | Structured: transaction records. Semi-structured: JSON app logs. Unstructured: call recordings, social media posts, contract documents |
| Veracity | Trustworthiness and accuracy of data — noise, bias, completeness | Alternative credit data (social media signals) may be noisy or manipulable; data quality directly affects model reliability and fair lending compliance |
| Technique | How It Works | Key Financial Use Case |
|---|---|---|
| Text Mining | Extracts structured information from unstructured text. Three steps: text selection → information extraction → mining relations between facts. 80% of available information exists in text form. | Automated contract clause review; regulatory document parsing; earnings call sentiment; KYC document processing |
| Graph Analysis | Represents connections and information flows. Measures reputation, models rumour propagation, identifies influencers and anomalies in networks. | Fraud ring detection (connected accounts sharing devices/addresses); AML transaction flow mapping; counterparty risk networks |
| Collaborative Filtering | "People who bought X also bought Y" — recommends based on behavioural similarity across the user population without explicit user input. | Cross-selling financial products; personalised investment recommendations; mortgage customers offered home insurance |
| Clustering (K-means) | Discovers hidden structures by grouping observations with similar characteristics without predefined labels — unsupervised learning. | Customer segmentation for personalised pricing; identifying risk cohorts in a loan portfolio; anomalous transaction clusters (fraud) |
| Pattern Recognition | Tags data and applies classification algorithms. Underpins biometrics (facial recognition), voice transcription, spam detection. | Facial recognition for KYC onboarding; voice authentication in call centres; fraudulent spending pattern detection |
| Predictive Modelling | 3-stage pipeline: (1) model training using objective + predictor variables; (2) model evaluation and adjustment; (3) deployment. Uses historical events to predict future outcomes. | Credit default prediction; customer churn early warning; delinquency prediction; propensity-to-buy models |
| Sentiment Analysis | Links events (what people do) with feelings (what they think). NLP-based. Key challenge: irony and sarcasm are hard to capture — language is ambiguous. | Customer satisfaction monitoring from reviews; call centre improvement; brand risk monitoring; earnings surprise prediction from news flow |
| Risk Analysis | Quantifying and modelling probability and impact of adverse events across portfolios using statistical and ML models. | Credit portfolio stress testing; market risk VaR modelling; operational risk scoring; insurance actuarial modelling |
Banks sit on one of the world's richest datasets — but struggle to exploit it due to legacy system fragmentation. The critical distinction is between data volume and data usability:
| Data Source | Type | Analytics Value |
|---|---|---|
| Core banking transactions | Structured | Highest — but siloed by product line; mortgage system doesn't talk to current account system |
| Card and payment data | Structured (MCCs, location, amount, time) | Rich behavioural signal — spending patterns reveal lifestyle, income, risk profile, health |
| Customer service calls/chat | Unstructured (audio, text) | Requires NLP/voice recognition; reveals churn signals, product complaints, financial stress |
| Digital/mobile app logs | Semi-structured (clickstream) | Navigation patterns reveal financial anxiety, engagement depth, product confusion |
| Loan application forms | Semi-structured (stated income) | Self-reported — veracity risk; basis for traditional underwriting but easily manipulated |
| Third-party bureau data | Structured (credit scores, public records) | Foundation of traditional credit decisioning; fails for thin-file customers — the gap FinTechs exploit |
How FinTechs Use Big Data: Company Profiles
S4 Slides 20–27| # | Strategy | Mechanism | Company |
|---|---|---|---|
| 1 | Predictive Analytics | Use historical patterns for better forward-looking credit, fraud, and attrition decisions — faster and more accurately than human underwriters | Prosper (Prosper Score) |
| 2 | Customer Insights | Aggregate and visualise spending to help customers understand their finances — increasing engagement and loyalty | ClarityMoney (Spend Analytics) |
| 3 | Credit Risk Analysis | Score previously un-scoreable populations using alternative data — enabling new market access banks cannot reach cost-effectively | KUESKI, LendGenius |
| 4 | Compliance & KYC | Automate identity verification, fraud detection, AML screening — reducing cost and false-positive rate vs manual review | IdentityMind |
| 5 | Cost Reduction | Automate manual underwriting, document review, customer service — FinTechs operate at 60–70% lower cost per loan than banks | All FinTech lenders (structural) |
Prosper (US marketplace lender) uses a two-score hybrid credit model that goes beyond traditional bureau data — an early, clean example of Big Data applied to consumer credit:
| Component | Detail |
|---|---|
| Prosper Score | Proprietary model focusing on debt-to-income ratio, payment history patterns, and "soft checks" from credit bureaus not included in standard FICO scores. Captures behavioural signals invisible to traditional scoring. |
| Bureau Score | Standard credit reporting agency score used as the second input — Prosper augments, not replaces, traditional data. |
| Borrower Grade | Combines both scores into a letter grade (AA through HR) that determines the interest rate offered. A strong Prosper Score can offset a thin bureau score. |
| Origination Fee | 0.5% to 5% of loan amount depending on grade — higher risk = higher fee charged to borrower at disbursement. |
From Big Data to Artificial Intelligence
S4 Slides 28–44The course material traces AI's historical arc (1950s → 1980s → present) to explain why the current wave succeeds where prior ones failed. Three converging factors made the difference:
| Factor | Why It Changed Everything |
|---|---|
| Exponential data volume | AI models require massive labelled datasets to train. The internet, smartphones, and IoT generated this data at unprecedented scale from the 2010s onward. Earlier AI attempts failed partly because training data was scarce and expensive to label. |
| Compute power (GPUs / TPUs) | Neural network training is embarrassingly parallel — GPUs reduced training time from months to hours. Cloud computing (AWS, GCP, Azure) democratised access: any FinTech startup can rent GPU clusters at variable cost with no upfront capital. |
| Algorithm breakthroughs | Deep learning (convolutional and recurrent networks), the transformer architecture (2017, "Attention Is All You Need"), and reinforcement learning breakthroughs enabled qualitative leaps — translating to finance via fraud detection, document processing, and quantitative trading. |
| Issue | The Risk in Finance | Regulatory Response |
|---|---|---|
| Workforce Impact | Baker McKenzie analysis (Slide 34): AI's impact on financial services employment is significant — roles involving data processing, compliance review, basic underwriting, and call-centre service are highly exposed. White-collar finance jobs are more vulnerable than previously assumed. | Social policy, retraining programmes; not yet directly regulated in financial services |
| Algorithmic Bias | AI models trained on historical data replicate historical discrimination. A credit model trained on past approvals will encode the biases of those approvals — amplifying rather than eliminating unfair lending patterns against protected groups. | EU AI Act: high-risk AI systems (including credit scoring) must be transparent, auditable, explainable. "Right to explanation" for automated decisions. |
| Opacity / Black Box | Deep learning models cannot explain their decisions in human terms — making regulatory compliance (ECOA in US, GDPR Art. 22 in EU) and customer fairness extremely difficult. | Explainable AI (XAI) requirements; SHAP values and LIME techniques for model interpretability becoming standard in high-risk applications |
| Category | Specific Applications | Maturity / Disruption Level |
|---|---|---|
| Customer-Focused | Credit scoring with ML models; dynamic pricing (insurance, mortgages, personal loans); personalised marketing; insurance underwriting; AI chatbots for 24/7 service | High — already deployed at scale. AI chatbots handle 60–80% of routine queries at major banks. Real-time credit decisions eliminate days-long underwriting. |
| Operations-Focused | Capital optimisation; risk management automation; stress testing (scenario generation); market impact analysis for large order execution | Medium-High — regulatory requirements (SR 11-7 model risk guidance) require human oversight. AI assists rather than replaces at this stage. |
| Trading & Portfolio Management | Algorithmic trading execution; systematic portfolio management; factor model generation; earnings prediction from alternative data (satellite imagery, credit card aggregates, job postings) | Very High — quantitative funds deploy AI extensively. Alternative data + ML is the new competitive edge in asset management. Human portfolio managers under structural pressure. |
| Compliance & Supervision | RegTech: automated regulatory reporting; macroprudential surveillance; data quality assurance; fraud detection at transaction scale. SupTech: regulators using AI to monitor banks in real time. | High and growing fast — ML reduces AML false-positive rates by 30–70%. SupTech means regulators can analyse entire transaction populations rather than sampling. |
Cloud Computing & IoT: The Infrastructure Layer
S4 Slides 45–52| Benefit | What It Means for FinTech | Incumbent Disadvantage |
|---|---|---|
| Elasticity | Scale compute up (peak season) and down automatically — pay only for what you use | Banks built fixed-capacity data centres sized for peak load — massive idle cost at non-peak times |
| Speed to market | Launch a new product in weeks without procuring physical servers. AWS/GCP/Azure provision in minutes. | Bank IT hardware procurement cycles: 6–18 months |
| Global reach | Deploy in any geography via cloud region — no local data-centre investment required | Banks negotiate hosting jurisdiction-by-jurisdiction for each regulatory regime |
| API ecosystem | Cloud platforms bundle pre-built ML tools, databases, data pipelines — FinTechs build sophisticated products from components | Banks maintain bespoke legacy tools that don't connect to modern ML stacks without expensive middleware |
| OPEX vs CAPEX | Variable operating cost replaces fixed capital expenditure — lowers break-even dramatically, extends capital runway | Banks carry heavy fixed IT CAPEX that must be amortised regardless of revenue |
IoT connects physical devices to the internet, generating real-world behavioural data that can transform financial underwriting, payments, and insurance. The slide projects an exponentially growing number of connected devices — each a potential financial data source:
| IoT Application | Financial Services Use | Companies |
|---|---|---|
| Telematics / Connected Cars | Pay-per-mile or behaviour-based auto insurance — driving speed, braking, time of day, mileage determine premium. Direct signal of actual risk, not proxy (age, postcode). | Progressive Snapshot, Metromile, Root Insurance |
| In-Car Payments (Slide 52) | Frictionless payment at pumps, parking garages, drive-throughs — car identifies itself, payment debited automatically. No wallet, phone, or card interaction needed. | Shell + BMW pilots; Honda / GM connected car payment trials; Visa/Mastercard automotive OEM partnerships |
| Wearables | Health data from smartwatches as underwriting signal for life/health insurance — real-time risk adjustment based on actual health behaviour, not actuarial tables | John Hancock Vitality (Apple Watch integration), AIA Vitality (Asia) |
| Smart Home | Property sensor data for home insurance pricing — water leak detectors, smart locks, security systems as risk-reduction signals that earn premium discounts | Hippo Insurance, Lemonade, Aviva |
Payments: A Large Business with Very Different Players
The economics of payment rails, two-sided network effects, and how PayPal — the original FinTech — built a dominant position by solving trust, not just technology.
Key Numbers
Session 05 DataCast of Characters
Case: PayPal Merchant Services (HBS)Who's who in PayPal's 2006 strategic decision
The Payment Value Chain
EconomicsPayPal is a textbook two-sided market (Rochet & Tirole). To create value, it must simultaneously attract two distinct user groups whose presence makes the platform more valuable to the other side.
| Dimension | PayPal | Google Checkout |
|---|---|---|
| Consumer base | 148M registered PayPal accounts | Hundreds of millions of Gmail/Google users |
| Merchant fee | 2.9% + $0.30 | Free to buyers; cheap for sellers (subsidised by ads) |
| Revenue model | Transaction fees (must be profitable) | Can cross-subsidise from search advertising |
| Trust anchor | eBay transaction history — proven escrow | Google brand + Gmail familiarity |
| Fraud advantage | Both-sides verification (0.17% fraud rate) | Google identity (Gmail) provides some verification |
Strategic Frameworks & Exam Angles
Key ConceptsNetwork effects in payments are among the strongest of any industry. A payment method's value to any user is a direct function of how many other users (and merchants) accept it. This creates self-reinforcing dynamics:
| Network Effect Type | Mechanism | Example |
|---|---|---|
| Direct (same-side) | More consumers with PayPal = easier P2P transfers between them | Venmo's social feed creates direct consumer-to-consumer value |
| Cross-side (two-sided) | More consumers = more valuable to merchants; more merchants = more valuable to consumers | Visa: 40M+ accepting merchants makes each Visa card more useful |
| Data network effects | More transactions = better fraud models = lower loss rate = more competitive pricing | PayPal's fraud rate advantage (0.17% vs 1.8%) comes from transaction history |
The case is set in 2006 — PayPal is still owned by eBay (acquired 2002 for $1.5B). This creates a strategic constraint: eBay has an incentive to keep PayPal as a captive service, limiting its ability to partner with Amazon, Alibaba, or Shopify. The off-eBay strategy is partially constrained by eBay's interests.
Case Discussion: Model Answers
PayPal Merchant Services — Class PrepPayPal's early success was driven by four converging factors — notably, none of them were planned:
| Factor | What Happened | Strategic Lesson |
|---|---|---|
| Unanticipated demand | PayPal launched as a P2P payment tool for Palm Pilots. eBay sellers — who couldn't accept credit cards and were sending cash in envelopes — adopted it spontaneously as a payment solution. PayPal didn't target eBay; eBay adopted PayPal. | The best early traction often comes from customers who were not in the original plan. Founders who observe and follow demand signals rather than forcing their original thesis succeed. |
| eBay's existing friction | The incumbent payment method for eBay was personal cheques — which took days to clear, could bounce, and required physical mailing. The bar PayPal had to clear was very low: faster, cheaper, and more reliable than cheques. | Incumbents' weaknesses define the FinTech opportunity. The worse the status quo, the more forgiving customers are of a new product's own flaws. |
| Credit card access democratisation | Auction sellers without merchant accounts (small traders, individuals) gained access to credit card payments for the first time through PayPal — dramatically expanding their buyer pool. | When a product unlocks access for previously excluded participants, adoption is driven by necessity rather than preference — much faster and stickier growth. |
| Viral mechanics | Every PayPal transaction sent between two people introduced the recipient to the product. Payment itself was the distribution channel — free viral growth at zero marginal cost. | Inherent virality (product functionality drives adoption) is more powerful and capital-efficient than paid marketing. PayPal's CAC on eBay was effectively zero. |
Winner-takes-all or winner-takes-most? Payments exhibits winner-takes-most dynamics, not winner-takes-all. The coexistence of Visa and Mastercard — two nearly identical products with roughly equal market shares — proves the market does not naturally collapse to a single winner. But it does heavily favour incumbents: switching costs on both sides (consumer: stored cards; merchant: integrated checkout) make displacement extremely difficult once density is established.
| Option | Argument For | Argument Against | Priority |
|---|---|---|---|
| Reduce transaction fees | Directly addresses the merchant's number-one objection; signals competitive intent against Google; could accelerate off-eBay merchant adoption at scale | Compresses already-thin gross margins; creates a floor problem — difficult to raise fees later; signals desperation to the market | Low-medium — only if targeted at strategic merchant segments, not blanket |
| Encourage non-eBay volume | Diversifies away from eBay dependency (the captivity problem); builds sustainable, multi-platform network density; reduces regulatory and competitive risk from eBay itself | eBay may retaliate by promoting competing payment options; requires merchant acquisition investment at scale; slower than eBay's organic growth | High — the long-term survival imperative. eBay-only PayPal has existential concentration risk. |
| Rewards programme | Increases consumer stickiness and frequency; reduces churn; can be structured as a merchant-funded marketing tool (merchants pay for visibility) | Expensive to fund; rewards attract value-seekers who churn when a better offer appears; complex to manage across currencies and markets | Medium — worthwhile but not the primary lever |
| Large merchant accounts | High-volume accounts provide scale and credibility; a signed deal with a major retailer (e.g. Gap, Best Buy) validates the enterprise proposition and opens mid-market doors | Large merchants have leverage — they can negotiate fees to near-zero, destroying unit economics; they also have IT procurement cycles that slow implementation | Medium — essential for brand credibility but not where volume profitability comes from |
| Acquisitions | Buying VeriSign Payment Services gave 144,000 merchant relationships overnight; acquiring competing checkout tools eliminates them as threats and adds network density | Integration risk; cultural disruption; overpaying in competitive auction processes; acquired networks may churn post-acquisition once PayPal's culture is imposed | High for specific, complementary assets (VeriSign was exactly right) |
Level of concern: high, but not existential — yet. Google Checkout in April 2006 is a genuine competitive threat, but its structural weaknesses are as significant as its strengths:
Lecture Slides: The Payments Landscape
S5 Slides 3–14The course material frames FinTech's disruption of payments as a fundamental business model shift — from banks organising themselves around products to organising around customer needs:
| Dimension | Product-Driven Model (Traditional Banks) | Customer-Driven Model (FinTechs) |
|---|---|---|
| Organisational logic | Product silos — mortgage division, card division, current account division. Each optimised independently. Customer is secondary. | Customer journey first — all products designed around a single, unified customer experience. Data flows across products. |
| Pricing | Standard rates with broad risk buckets — same mortgage rate for very different risk profiles; hidden fees embedded in product terms | Dynamic, personalised pricing — risk-based, real-time, transparent. Customers understand what they pay and why. |
| Distribution | Physical branches as primary channel — high fixed cost, limited hours, geographic constraint | Mobile-first, 24/7, global — customer chooses channel. Cost per customer interaction: <$0.01 vs $4+ for branch |
| Data use | Data siloed by product — mortgage system doesn't talk to current account system. No single customer view. | Unified data layer — every interaction improves credit model, fraud detection, and personalisation simultaneously |
| Revenue model | Fee-heavy: account maintenance fees, overdraft fees, transfer fees, minimum balance requirements | Fee-light or free at entry level — revenue from interchange, premium subscriptions, lending spread, and data insights |
Despite the FinTech narrative, the slides ground the discussion in data: cash remains the most-used payment method globally — but the structural trend is clear and accelerating:
| Trend | Data from Slides | Implication |
|---|---|---|
| Non-cash transaction growth | Non-cash transactions were growing steadily pre-2020. COVID-19 caused a step-change acceleration — consumers and merchants adopted contactless and digital payment out of hygiene and necessity. | The pandemic compressed 5 years of adoption into 12 months. Habits formed during COVID have proven sticky post-pandemic. |
| Cash use declining | Cash's share of total transactions has declined in every major economy, consistently, since 2015. Sweden approaches fully cashless (<1% of transactions). UK at ~15% cash by volume. | But high-cash markets (Germany, Italy, large parts of Asia/Africa/LatAm) will persist for decades — regulatory and cultural dimensions, not just technology. |
| Contactless acceleration (COVID) | Mastercard April 2020 survey: majority of respondents reported using cash less frequently due to hygiene concerns — driven by fear of coin/note transmission. | Contactless (NFC) became the default at POS globally. Contactless payment limit increases (e.g. UK: £45 → £100) made the shift permanent. |
| High-cash market persistence | Markets with high structural cash use: Germany (cultural distrust of surveillance), India (pre-demonetisation), Mexico, Indonesia, Egypt, Nigeria. | FinTech opportunity in these markets is largest — but requires solving trust, infrastructure, and connectivity challenges, not just UX. |
The course material presents a stakeholder map of the full payments ecosystem, categorising the traditional and emerging players by function:
| Player Category | Role in the Ecosystem | Examples |
|---|---|---|
| Card Networks | Set the rules, operate the authorisation and clearing infrastructure, collect assessment fees from both sides | Visa, Mastercard, American Express, UnionPay |
| Issuers | Issue cards to consumers, extend credit, collect interchange fees, bear fraud risk on consumer-side | JPMorgan Chase, Citi, Barclays, Revolut, Monzo |
| Acquirers | Sign up merchants, process transactions, collect merchant discount rate, pay interchange to issuers | Worldpay, Adyen, Stripe, Square, Paymentech |
| Payment Gateways | Technical layer connecting merchants' websites/POS to the acquirer — encryption, tokenisation, routing | Stripe, Braintree (PayPal), Checkout.com, Cybersource |
| Digital Wallets | Store multiple cards/bank accounts in one interface; add a security/convenience layer above the card network | PayPal, Apple Pay, Google Pay, Samsung Pay, AliPay, WeChat Pay |
| P2P / Transfer | Direct person-to-person or cross-border money movement — bypassing or riding on card rails | Venmo, Zelle, Cash App, TransferWise (Wise), Revolut, Remitly |
| Fraud / Security | Real-time transaction monitoring, identity verification, device fingerprinting to prevent losses at scale | Forter, Signifyd, Kount, Stripe Radar |
New Payment Players: Eight Company Profiles
S5 Slides 20–29| Dimension | Detail |
|---|---|
| What it does | Global payments platform allowing businesses to accept payments from major credit/debit cards, mobile apps, Apple Pay, Google Pay — in a single integration across countries and currencies |
| Volume | Processed close to $150 billion in payments at the time of the slide (now significantly higher as a public company) |
| Client base | Enterprise-focused: Microsoft, Uber, Spotify, Sephora — multinationals that need a single payment partner across dozens of markets |
| Strategic partnerships | Expanded into Chinese market through partnership with Alipay — enabling European merchants to accept China's dominant payment method |
| Business model | Interchange++ pricing (transparent component breakdown) + processing fee. Unlike Stripe's developer-first SME focus, Adyen targets large merchants and processes both online and in-store (unified commerce) |
Mobile Payments: China & India as Case Studies
S5 Slides 30–36China leapfrogged cards entirely — going from cash to mobile payments without the decades-long credit card infrastructure build that Western markets had. The result: two super-apps dominating a near-cashless society:
| Dimension | Alipay (Ant Group / Alibaba) | WeChat Pay (Tencent) |
|---|---|---|
| Origin | Escrow service for Taobao e-commerce (2004) — solving merchant/buyer trust problem, analogous to PayPal's eBay origin | Payment feature inside WeChat social messaging app (2013) — leveraging 1B+ existing social graph |
| Distribution strategy | Merchant-first: built acceptance network through Alibaba's commerce ecosystem | Consumer-first: viral via WeChat "Red Envelopes" (digital money gifts) during Chinese New Year 2014 — 40M sent in 24 hours |
| Market position | Stronger in e-commerce, rural areas, financial services (MYbank, Yu'e Bao money market) | Stronger in social, peer-to-peer, offline merchant QR payments |
| Super-app model | Alipay = payment + insurance + loans + wealth management + utilities in one app | WeChat = messaging + social + payments + mini-programs (embedded apps within WeChat) |
| Combined reach | Together >90% of China's mobile payment market. A foreign tourist in Shanghai can go days spending nothing but QR codes. China is the closest major economy to being cashless. | |
| Dimension | UPI (Unified Payments Interface) |
|---|---|
| What it is | A real-time payment system built by NPCI (National Payments Corporation of India) on top of the existing IMPS (Immediate Payment Service) infrastructure. Launched April 2016 — months before demonetisation made it essential. |
| Architecture | Open-API interoperable layer: any bank can connect, any app can build on it. A consumer's UPI ID (like an email address) routes payment to their bank — no card numbers needed. |
| Key innovation | Interoperability by design — any UPI app can pay any UPI merchant regardless of which bank or app either uses. This solved the network fragmentation that plagued earlier mobile payment attempts. |
| Scale achieved | By 2023, UPI processed 100+ billion transactions annually, surpassing credit card volumes. India became the world's largest real-time payment market by volume — ahead of China, USA, EU. |
| Key players | PhonePe (Flipkart/Walmart), Google Pay, Paytm, BHIM (government app), WhatsApp Pay. Competition is on UX and features — the underlying UPI rail is shared infrastructure. |
The Future of Payments: Cashless Societies & Neobanks
How Nubank built Latin America's most valuable bank from a purple credit card — using referral virality, real-time data, and radical customer empathy to crack Brazil's oligopoly.
Key Numbers
Session 06 DataCast of Characters
Case: Nubank (HBS)Who's who in Nubank's story
Nubank's Playbook
Growth ModelNubank launched its credit card in September 2014 — deliberately purple because market research showed it was the colour least associated with a credit card. This single design decision made the card a social object, not just a financial product.
| Nubank Feature | What It Solved | Incumbent Comparison |
|---|---|---|
| No annual fee (ever) | Brazil's banks charged high annual fees as standard — a major pain point | Big Five: mandatory annual fees regardless of spend |
| 2-minute app approval | Traditional banks required multiple branch visits and document submissions | Big Five: days to weeks for credit card approval |
| Real-time app notifications | Instant spending visibility — customers knew exactly where money went | Big Five: monthly paper statements only |
| Customer can lower own credit limit | Gave customers control over their own financial behaviour | Unheard of at traditional banks |
| No foreign transaction fees | Major saving for young professionals who travel | Big Five: standard 4-6% FX surcharge |
The Mexico Expansion Decision
Case Central QuestionSession 6 uses Nubank as the case entry point but broadens to the future of payments landscape. Two major trends:
| Model | How it Works | Revenue Source | Key Risk |
|---|---|---|---|
| BNPL (Klarna, Afterpay, Affirm) | Split purchase into 3-6 interest-free instalments for consumers | Merchant fees (2-8%): merchants pay because BNPL boosts conversion 20-30% | Consumer over-indebtedness; credit cycle exposure; regulation |
| Neobanks (Revolut, N26, Chime) | Digital-only, no branches, superior UX, typically lower fees | Interchange fees + premium subscriptions + FX spread | Customer acquisition cost; path to profitability; regulatory licence |
| Embedded Finance (Shopify, Uber, Amazon) | Financial products built into non-financial apps and flows | Revenue share from BaaS partner + data monetisation | Regulatory compliance for non-financial companies |
Case Discussion: Model Answers
Nubank — Class PrepThe Brazilian banking market in 2013 was a textbook oligopoly: five banks controlled 79.5% of assets and 82.8% of deposits — with the market power to extract extraordinary rents from customers who had no meaningful alternative. David Vélez experienced this personally when opening his own account: trapped behind bulletproof glass, treated as a suspect, offered a credit card at 450% APR. His personal experience was the market's experience at scale.
| Pain Point | Specific Evidence | Why Incumbents Didn't Fix It |
|---|---|---|
| Predatory pricing | Credit card APR: 272% annually (December 2018). Annual card fees charged regardless of spend. Overdraft fees compounding daily. | Five banks with ~80% market share face no competitive pressure to reduce margins. Oligopoly equilibrium is stable as long as no entrant can reach scale. |
| Hostile UX | Branch visits required for basic tasks. Paper statements only. Account opening required extensive documentation and multiple in-person visits. No real-time transaction visibility. | Legacy IT systems; branch network represents sunk cost that must be utilised; no competitive reason to invest in digital UX that might reduce branch visits. |
| Gatekeeping culture | Consumers describe banks as treating them as adversaries rather than clients. Vélez's bulletproof-door experience is emblematic: security designed to intimidate, not to serve. | Culture inherited from decades of high-inflation Brazil, when bank clients were routinely in arrears and fraud was rampant. The culture calcified even as the macroeconomic context improved. |
| High barriers to first credit | First-time credit card applicants routinely rejected despite sufficient income; incumbents require documented employment history and a prior credit relationship to grant credit — a circular trap. | Conservative underwriting culture; cost of manual underwriting makes thin-file customers unprofitable at small loan sizes; preference for established relationships with lower default risk. |
After identifying the pain point, Vélez faced the two foundational challenges every FinTech founder encounters before the first product exists: convincing investors to fund an unproven idea in a high-risk emerging market, and recruiting technical talent willing to bet on a startup against entrenched banking incumbents.
The purple Mastercard credit card (launched September 2014) was chosen deliberately as the first product. Vélez's logic: credit cards are the single product Brazilians hate most about their banks — highest fees, worst UX, most opaque pricing. Win on the most hated product and you win the customer relationship entirely.
| Product Decision | What It Was | Why It Worked |
|---|---|---|
| Mastercard partnership | Nubank is not a bank — it needed a licensed network partner. Mastercard provided the card network infrastructure and international acceptance. | Enabled Nubank to launch without a full banking licence. Mastercard gained a fast-growing FinTech issuer in a market where it was underrepresented. |
| Easy and fast onboarding | 2-minute app-based application. No branch visit. Decision delivered in minutes, card arrives by mail within days. | Directly attacked the most common complaint — the friction of opening a traditional bank credit card. Speed became the product. |
| 2,000-point credit model | Proprietary underwriting combining traditional bureau data with alternative signals. Real-time decision. Dynamic credit limit adjustments as repayment data accumulates. | Approved customers incumbents would decline (thin-file, young professionals); lower default rates than expected because the model found genuine creditworthy signals beyond bureau scores. |
| 70% referral acquisition | Waitlist model: existing customers received referral codes; referred applicants jumped the waitlist. 70% of applications came via referral. | Near-zero CAC. Self-selection: referred customers had lower default rates because they were vouched for by creditworthy existing customers. Virality built trust simultaneously. |
| No annual fee + merchant fee revenue | No annual fee — ever. Revenue from interchange (merchant fee) and interest on revolving balances. | Removed the most viscerally resented cost. Forced Nubank to compete purely on engagement and spend volume — which it won by designing a card worth using daily. |
| Target: pay-on-time customers | Deliberately went after customers who pay balances in full each month — not the high-interest revolvers banks prefer. | Lower credit risk; stronger NPS; higher referral rate; longer lifetime. Revolvers churn when they find a cheaper rate; on-time payers stay for the relationship. |
| Credit limit increases + dynamic analysis | System continuously monitored repayment behaviour and automatically raised limits for reliable customers. | Created a positive engagement loop: customers checked the app to see their growing limit. Loyalty through progressive trust — a fundamentally different relationship than the adversarial incumbent model. |
This is the case's central decision — presented to the Nubank board in January 2019. The question is not simply "is Mexico a good market?" but "is this the right time, for this company, to take this step?"
Lecture Slides: COVID-19 as FinTech Accelerator
S6 Slides 4–26The course material frames the session around six structural trends that the COVID-19 pandemic accelerated beyond their pre-2020 trajectories — together creating the "contactless society" that is the platform for Digital Banking growth:
| # | Trend | Pre-COVID State | COVID Acceleration |
|---|---|---|---|
| 1 | Digital Payments | Growing steadily — cash declining gradually, contactless adoption slow in many markets | Step-change: hygiene concerns eliminated cash preference; contactless became default at POS globally; limits raised (UK: £45→£100) |
| 2 | E-commerce | Growing at ~15% annually — established in electronics, fashion, travel | 5 years of growth in 12 months. Categories previously resistant (grocery, pharmacy, furniture) forced online by lockdowns. Platforms that couldn't ship struggled; pure-play e-commerce exploded. |
| 3 | Digital User Experience | Banks invested in apps but many customers still preferred branch for complex transactions | Branch closures during lockdown forced even reluctant customers (elderly, SME) online. Banks that had invested in digital UX gained; those that hadn't lost customers permanently. |
| 4 | Financial Inclusion | World Bank 2014→2020: adult account ownership increased from ~62% to ~76% globally — progress, but slower than needed | Government stimulus payments (direct to accounts) forced account opening in many developing countries. Mobile money (M-Pesa, bKash) saw transaction volume surge as in-person alternatives closed. |
| 5 | Remote Working | ~5% of workers worked from home regularly (pre-2020). Videoconferencing niche. | Zoom, Teams, Google Meet went from niche to essential. Knowledge workers globally demonstrated remote productivity. Zoom revenue: $623M (2019) → $2.65B (2020). Permanently reshuffled real estate, talent markets, and digital infrastructure investment. |
| 6 | Distance Learning | EdTech growing but constrained by institutional resistance. Online courses seen as inferior. | 187M students globally affected by school closures. Virtual learning platforms forced at scale. Resilient habit formation, especially in higher education. EdTech funding surged: $16.1B in 2020, up 5x from 2015. |
The course material highlights the COVID-driven surge in retail stock trading as an example of e-commerce penetrating financial markets — using Robinhood (US) and its European equivalents as case studies:
Nubank: Lecture Deep-Dive — Credit Model & Culture
S6 Slides 27–34The course material provides more detail on Nubank's credit product mechanics than the HBS case. Three innovations made the credit card defensible, not just desirable:
| Innovation | Detail | Why It Mattered |
|---|---|---|
| 2,000 data point credit model | Nubank's underwriting used approximately 2,000 data points per applicant — combining traditional bureau data with alternative signals from the application process, device data, and referral network quality | Enabled real-time credit decisions in minutes; allowed credit extension to thin-file customers incumbent banks rejected; dynamic credit limit increases as repayment data accumulated |
| Referral-powered acquisition (70%) | 70% of credit card applications came via referral from existing customers. Nubank deployed "Referral Lists" — invited customers could refer friends who jumped the waitlist | Near-zero CAC vs ~R$300 ($75) for traditional bank credit card customer acquisition. Referrals also self-selected: referred customers had lower default rates because they were vouched for by creditworthy existing customers |
| Dynamic credit limit increases | Rather than periodic manual reviews, Nubank's system continuously monitored repayment behaviour and automatically increased limits for customers demonstrating reliability | Builds customer loyalty and increases spending (higher interchange revenue); creates positive engagement loop — customers check the app to see their growing limit |
| Strategic target: "pay-on-time" customers | Deliberately went after customers who pay their full balance on time — not the revolvers (those who carry balances and pay high interest) that traditional banks optimise for | Lower default risk; builds sustainable relationship; interchange income is sufficient at scale; revolvers churn when rates drop — on-time payers stay loyal |
The Future of Payments: 12 Key Trends
S6 Slides 35–49| # | Trend | Key Insight from Slides |
|---|---|---|
| 1 | Understanding Gen Z, Not Just Millennials | Gen Z (born 1997–2012) entered adulthood during COVID. Defining characteristics: digital natives (no memory of pre-smartphone world), higher financial anxiety than Millennials, distrustful of institutions, prefer embedded finance over standalone banking apps. Different from Millennials in risk tolerance and platform loyalty — requires distinct product design, not just younger marketing. |
| 2 | UX Is the Key Differentiator | In a world where any payment can be made from a phone, product experience determines market share. The slide argues UX is now the primary competitive dimension — not price, not features, not brand heritage. NPS (Net Promoter Score) correlates directly with wallet share in payments. Banks with best UX: Monzo, Revolut, Nubank, N26. Worst UX incumbents lose customers to challengers despite rate and product advantages. |
| 3 | Mobile Is King | Worldpay Global Payments Report 2025: over 50% of the world's population used a mobile wallet at least monthly in 2025. Wallets gaining momentum across all markets — including previously cash-dominant ones. Mobile is not just a channel — it is the product for the next generation of financial services consumers. |
| 4 | Consumers Love Rewards | Rewards (points, cashback, miles) remain a dominant factor in payment method choice — particularly for credit cards. FinTechs have democratised rewards: Nubank (rewards), Revolut (premium cashback), Amex strategy. Key dynamic: rewards programmes create switching costs and drive spending behaviour — the rational consumer optimises around rewards, shifting spend to higher-reward instruments. |
| 5 | Player Collaboration Increasing | The payment industry is moving from competitive silos to collaborative ecosystems. Banks partner with FinTechs for technology; FinTechs partner with banks for regulatory access and balance sheets; card networks partner with digital wallets. BaaS (Banking as a Service) makes collaboration modular — any player can provide any layer of the stack to any other. |
| 6 | Great Symbiosis with FinTechs | The course material highlights the Visa Fast Track European programme (Slide 42) — Visa's initiative to onboard and certify FinTech partners quickly, giving them access to the Visa network in weeks rather than the years-long traditional certification process. This is the card network's response to disintermediation risk: co-opt the disruptors by making them partners rather than enemies. |
| # | Trend | Key Insight from Slides |
|---|---|---|
| 7 | Tokenisation Has Arrived — and Will Stay | Tokenisation = replacing sensitive card data (PAN) with a non-sensitive token. Already standard in Apple Pay/Google Pay (device tokenisation). Expanding to: asset tokenisation (real estate, bonds, funds on blockchain); loyalty points; digital identity. The slide argues tokenisation is not a buzzword — it is the architectural shift enabling every frictionless payment use case of the next decade. |
| 8 | Crypto Not Yet Getting Traction in Payments | Despite the technology and narrative, crypto (BTC, ETH) has not achieved meaningful penetration as a payment medium. Reasons: volatility makes it unsuitable as a unit of account; merchant acceptance is minimal; regulatory uncertainty. Stablecoins show more promise for B2B settlement but face their own regulatory challenges (MiCA). The slide is candid: the payments use case for crypto remains largely aspirational in 2024. |
| 9 | Payments Everywhere (Embedded Payments) | The shift from "going to pay" to "payment happens invisibly" — Uber, Amazon Go, Shopify, Airbnb. The best payment is the one you don't notice. This requires deep merchant integration, data sharing agreements, and often proprietary payment rails (Amazon's ACH push, Uber's global payment orchestration). The endpoint: all consumer and B2B transactions are embedded in workflows, not separate payment acts. |
| 10 | BNPL on the Rise | Three provider types (Slide 47): Direct Providers (Klarna, Afterpay — own the customer relationship and credit risk); Facilitators (Affirm embedded at checkout — merchant-integrated, brand-visible); Retroactive Providers (Splitit — converts existing credit card purchases to instalments after the fact). Regulatory pressure increasing globally: UK Financial Conduct Authority, EU Consumer Credit Directive revision, CFPB scrutiny in US. |
| 11 | Fraud Is Also Innovative | As payment fraud defences improve, so does the sophistication of fraud. The slide highlights that fraud is itself an innovation ecosystem — evolving from card skimming to synthetic identity fraud, to AI-generated deepfake authorisation calls, to social engineering at scale. The arms race between fraud and fraud prevention is permanent. FinTechs that win on fraud (PayPal's 0.17%, Forter's ML) have a sustainable cost and experience advantage. |
| 12 | Adapting to Regulation | The payments regulatory environment is intensifying globally — PSD2/Open Banking in EU, FedNow in US, UPI in India, faster payments mandates globally. The slide frames regulation as a force that can either enable (Open Banking creating API mandates) or constrain (GDPR, SCA friction). FinTechs that embed regulatory compliance into product design from day one have structural advantage over those who treat it as an afterthought. |
PSD2, Open Banking & Cyber Security
S6 Slides 50–56| Dimension | Detail |
|---|---|
| What PSD2 is | Payment Services Directive 2 — an EU regulation that entered into force January 2018. An update of the original PSD (2007). Mandates banks to open their payment infrastructure and customer data (with customer consent) to licensed Third Party Providers (TPPs) via standardised APIs. |
| Two core objectives | (1) End fragmentation in European payment services — create a unified EU payment market; (2) Increase competition and innovation — force banks to share their data moat with FinTechs |
| Third Party Providers (TPPs) | Two types: AISPs (Account Information Service Providers — read-only access to aggregate account data, e.g. Yolt, Moneyhub) and PISPs (Payment Initiation Service Providers — trigger payments directly from customer accounts, e.g. GoCardless, Truelayer) |
| Open Banking impact | Banks must provide APIs giving TPPs access to account data and payment initiation — with customer consent. This enables: aggregated account views, automated switching, cashflow-based credit scoring, and A2A (Account-to-Account) payments that bypass card networks entirely |
| Strong Customer Authentication (SCA) | PSD2 also mandated SCA — two-factor authentication for most online payments (something you know + something you have/are). Added friction to checkout, creating implementation challenges for merchants during 2019–2021 rollout. |
The session closes by asking: who will be the dominant payment players in the future? And what are the systemic risks of an increasingly digital payment infrastructure?
| Future Big Payment Players (Slide 52) | Rationale |
|---|---|
| Big Tech (Apple, Google, Meta, Amazon) | Hundreds of millions of existing consumer relationships, device control, and data. Apple Pay at 1B+ capable devices. The question is regulatory appetite: post-Libra, regulators are watching closely. |
| Super-App platforms | WeChat/Alipay model: if you control the social/commerce interface, payments are a natural extension. Southeast Asia (Grab, Sea), LatAm (Nubank), and Africa (M-Pesa ecosystem) show this works outside China. |
| Established card networks | Visa and Mastercard are evolving from card networks to "network of networks" — processing real-time payments, crypto settlements, and A2A transfers. Their durability is higher than often assumed. |
| Central banks (CBDCs) | If retail CBDCs roll out at scale, they could displace card and bank-account payment rail dominance. Direct competition with commercial payment infrastructure — the systemic wildcard. |
Peer-to-Peer Lending
Lending Club showed the world that marketplace lending could work at scale — until governance failure, funding volatility, and the credit cycle exposed the model's structural vulnerabilities.
Key Numbers
Session 07 DataCast of Characters
Case: Lending Club (Stanford GSB)Who's who in Lending Club's rise and fall
How the Marketplace Works
P2P Mechanics| Grade | Approx Rate Range | Default Risk | Investor Appeal |
|---|---|---|---|
| A | 5-8% | Lowest | Safety-first investors |
| B-C | 9-14% | Low-Medium | Core retail investor target |
| D-E | 15-22% | Medium | Yield-seeking investors |
| F-G | 23-30% | High | Institutional only (high yield) |
| Issue 1 | Altered loan dates: $22M of near-prime loans sold to Jefferies (a major investor) had incorrect application dates. The investor had specific date requirements. Someone on Laplanche's team changed the dates to meet these criteria and the loans were sold. This is fraud — misrepresentation of asset characteristics to investors. |
| Issue 2 | Undisclosed conflict of interest: Laplanche had a personal investment in Cirrix Capital — a fund that LC had been considering investing in as a way to provide liquidity for its platform. He did not disclose this to the board. The board found out and demanded his resignation. |
| Impact | Stock fell ~35% on the announcement. Institutional investors pulled back. LC's origination volume collapsed in Q2 2016 — demonstrating the platform's funding fragility. The scandal exposed the fundamental tension of running a marketplace: LC needed to act as a neutral intermediary but had incentives to keep investors happy. |
| Dimension | Traditional Bank | Lending Club (Marketplace) |
|---|---|---|
| Funding source | Sticky deposits (FDIC-insured, slow to move) | Investor appetite (performance-sensitive, fast to move) |
| Credit risk holder | Bank balance sheet (requires capital buffer) | Investors (no capital required by LC) |
| During market stress | Can continue lending from deposit base | Originations collapse if investors withdraw |
| Business model | Counter-cyclical potential (deposit rates fall in stress) | Pro-cyclical (investor risk appetite contracts in stress) |
| Regulatory capital | Heavy (Basel III: 8%+ of RWA) | Minimal (no balance sheet lending) |
Case Discussion: Model Answers
Lending Club — Class PrepThis is the case's most analytically demanding question — it is fundamentally a question about the nature of Lending Club's business model, not just a valuation argument. The answer determines which comparables are appropriate, which risks investors must price, and what the long-run equilibrium looks like.
| Dimension | Technology / Marketplace View | Specialty Finance View | Evidence |
|---|---|---|---|
| Revenue sensitivity | Fee income from originations is recurring and volume-driven — like a SaaS platform charging per transaction | Revenue collapses in credit downturns (investors withdraw) — exactly like a finance company losing its funding line | Finance: 2016 scandal caused 30%+ origination collapse in one quarter. 2020 COVID caused similar. Revenue is not recurring — it is cycle-sensitive. |
| Balance sheet risk | LC holds no loans — investors bear all credit risk. Asset-light = tech multiple justified | LC's platform viability depends entirely on investor credit appetite — which is itself a form of balance sheet risk, just off-balance-sheet | Finance: the risk did not disappear — it was transferred to investors whose behaviour directly determines LC's revenue. This is hidden leverage. |
| Moat | Network effects: more borrowers → better data → better risk grades → more investor appetite → more borrowers. Compounding flywheel. | Underwriting quality is the moat — but this is indistinguishable from a specialty lender's competitive advantage | Ambiguous: LC's underwriting was good but not uniquely defensible. SoFi, Prosper, and eventually banks replicated the model. |
| Regulatory treatment | Technology intermediary — lighter-touch regulation, no capital requirements | The SEC classified LC's notes as securities in 2008. FCA and international regulators treat P2P platforms as financial intermediaries. | Finance: regulatory trajectory is toward treating P2P as finance, not technology. This is a permanently resolving ambiguity — in the wrong direction for tech valuations. |
More concerned than he appeared to be — for a specific structural reason. The threat from private competitors is not primarily about product or pricing. It is about information asymmetry and strategic patience.
| Competitive Dimension | Public LC's Position | Private Competitor Advantage |
|---|---|---|
| Strategic visibility | LC must disclose origination volumes, default rates, underwriting changes, and strategic initiatives quarterly. Every weakness is public. | SoFi, Prosper, Avant operated in the dark — could see LC's strategy while hiding their own. Could pivot without public narrative management. |
| Time horizon | Quarterly EPS pressure forces origination volume optimisation — cannot accept short-term volume decline even if it would improve long-term credit quality | Private companies can sacrifice 2 years of growth to invest in underwriting improvement, product expansion, or geographic entry without analyst scrutiny |
| Capital flexibility | Post-IPO equity is expensive to issue (dilution is visible, priced in real time); debt capital requires credit ratings | Private equity and VC investors can deploy patient capital into competitive initiatives on a multi-year horizon without market repricing |
| Talent | Public company stock options are priced in real time — if the stock underperforms, compensation deteriorates and talent leaves | Private company equity is illiquid but has asymmetric upside — attracts talent willing to accept illiquidity for larger potential gain |
Lecture Slides: P2P Lending — Case Discussion & Evolution
S7 Slides 2–19The slides contain the course material's case discussion questions and key takeaways — critical for exam preparation:
| Discussion Question | The Analytical Frame |
|---|---|
| Marketplace technology co. vs specialty finance co.? (Slide 4) | Tech company valuation: revenue multiple (15–20x), driven by growth potential and asset-light model. Specialty finance valuation: book value multiple (1–2x), driven by credit quality of loan book. Lending Club wanted tech-company treatment but its revenue was fundamentally credit-cycle-sensitive. The stock market eventually forced the reclassification — LC traded at finance company multiples post-IPO. |
| Costs and benefits of Laplanche's growth strategy? (Slide 5) | Benefits: rapid origination growth, brand leadership, first-mover institutional investor relationships. Costs: pressure to relax underwriting standards to maintain volume; investor perception of growth-at-any-cost; institutional concentration risk; public company pressure amplified all these tensions. Scaling responsibly in lending means accepting slower growth — a tension with equity market expectations. |
| Competition concern: private vs public competitors? (Slide 6) | Private FinTech competitors (SoFi, Prosper) didn't face quarterly earnings pressure — could invest in product without managing analyst expectations. LC's public status created an information asymmetry disadvantage: competitors could see LC's strategy while hiding their own. Laplanche's challenge: compete against better-funded, less visible rivals while managing public market narrative. |
One of the session's most important structural insights: the P2P label is now a misnomer. The course material identifies four consequences of the shift from retail to institutional funding:
| Consequence | What It Means |
|---|---|
| Platform-to-Institution | "Peer-to-peer" lending has become "platform to hedge fund / pension fund / CLO". The retail investor is largely gone (LC discontinued retail investors in 2020). The democratisation narrative — ordinary people lending to ordinary borrowers — no longer accurately describes the business. |
| Liquidity cycle risk becomes critical | Institutional investors are performance-sensitive and can withdraw entire funding allocations at quarter-end. When credit conditions deteriorate, institutional appetite contracts simultaneously and rapidly — far faster than retail investor withdrawals historically. This is the mechanism that caused LC's Q2 2016 origination collapse after the governance scandal. |
| Concentration risk increases | A platform funded by 5 large hedge funds has far higher concentration risk than one funded by 100,000 retail investors. The loss of one institutional relationship can reduce originations by 20–30% immediately — retail investors churn gradually, not in lumps. |
| Underwriting pressure | Institutional investors have yield targets and volume requirements that create pressure on platforms to maintain origination volume — sometimes at the expense of credit quality. This is the mechanism behind the "underwriting drift" risk: loosening standards to keep institutional relationships happy and volume metrics up. |
| Risk Type | Description | Example / Signal |
|---|---|---|
| 1. Credit Risk | Model accuracy and underwriting drift — alternative data models tested in benign credit environments may fail in recessions | COVID-19 caused massive defaults in BNPL and marketplace lending books that had no recession data in their training sets |
| 2. Liquidity Risk | Institutional investor pullback — platforms that depend on institutional funding face sudden origination collapse when credit markets tighten | LC Q2 2016: governance scandal → institutional withdrawal → 30%+ origination collapse in one quarter |
| 3. Regulatory Risk | Classification as lender vs marketplace — determines capital requirements, investor protection obligations, and supervisory regime | SEC classified LC's notes as securities (2008): major compliance burden. UK FCA brought P2P platforms under its regulatory perimeter in 2019 |
| 4. Operational Risk | Fraud, borrower verification failures, data integrity — particularly acute at high application volumes with automated decision-making | Identity fraud, income misrepresentation, synthetic identity fraud at scale without manual review |
| 5. Reputational Risk | Platform failures harm trust for the entire sector — one high-profile failure raises investor and borrower risk perception across all platforms | Prosper's early near-collapse (2009); Funding Circle's post-IPO underperformance; WeLend (China) fraud — all depressed sector investor appetite |
The Full Alternative Finance Landscape
S7 Slide 20The course material presents alternative finance as a broad ecosystem — far beyond P2P lending. The slide provides a definitive landscape map:
| Category | Description | Key Players |
|---|---|---|
| P2P & Marketplace Consumer Lending | Unsecured personal loans funded by investors or institutions through a digital marketplace platform. Classic P2P model. | LendingClub, Prosper, Funding Circle (historical consumer), Zopa (historical) |
| Digital Balance Sheet Lenders | Consumer lending and refinancing funded directly from the lender's own balance sheet — credit risk held internally, not passed to investors. | SoFi, Upstart, Earnest, Marcus (Goldman, historical), Avant, Laurel Road |
| BNPL & Embedded Checkout Lending | Short-to-medium-term instalment payments integrated at checkout (online or POS). Merchant-subsidised — consumer often pays zero interest. | Klarna, Affirm, Afterpay, Sezzle, PayPal Pay in 4, Apple Pay Later |
| Neobanks with Lending | Digital banks offering accounts, payments, cards, and credit products within one integrated app. | Revolut, N26, Chime, Monzo, Varo |
| Digital Mortgage & Property Lending | Online mortgage origination, refinancing, home equity loans, or digital real estate lending — removing the branch-based, document-heavy process. | Better Mortgage (Better.com), Rocket Mortgage, Blend (infrastructure), LendInvest |
| SME & Merchant Working Capital | Loans or cash advances for small businesses and merchants — often repaid as a percentage of sales or through automated debits from business accounts. | Kabbage, PayPal Working Capital, Square Loans, Amazon Lending, Shopify Capital, OnDeck, BlueVine |
| Microfinance & Microcredit | Small short-term digital loans via mobile apps — targeting underserved or thin-file customers in emerging markets. The financial inclusion use case. | Tala, Branch, FairMoney, Migo, Kiva (social lending) |
| Crowdfunding Platforms | Reward-based or investment-based funding for projects, entrepreneurs, or startups — aggregating many small contributions to reach capital targets. | Kickstarter (rewards), Indiegogo (rewards), Fundable (equity/rewards), Seedrs (equity), Crowdcube (equity), GoFundMe (donations) |
Company Profiles: Balance Sheet Lenders, BNPL & Mortgages
S7 Slides 21–30| Dimension | Detail |
|---|---|
| Origins | Founded 2005 in Sweden as an online payments solution. Evolved from a checkout payment method into a global BNPL leader with strong merchant integrations across Europe, US, and Australia. |
| Core product | Pay-in-4 and pay-later products embedded directly at checkout — increasing merchant conversion rates and average order value. The "ghost card" model: Klarna generates a one-time virtual card the consumer uses at any merchant. |
| Business model | Primarily a balance-sheet lender — takes credit risk directly and funds loans through debt facilities, securitisation, and banking licences in Europe. Revenue: merchant fees (2–5%) + interchange on Klarna card + interest on longer-term financing products. Consumers pay zero if they pay on time; Klarna earns entirely from merchants. |
| AI underwriting | Uses advanced AI-driven underwriting and alternative data to make real-time credit decisions in milliseconds — no credit check, no hard inquiry, instant approve/decline. Klarna's model: high approval rates + low fraud loss + merchant fee revenue = positive unit economics at scale. |
| Strategic position | Klarna went from a $45B peak valuation (2021) to a $6.7B down-round (2022) to an IPO process at recovering valuation (2024–2025) — one of the starkest FinTech valuation cycles. Lessons: BNPL economics are highly interest-rate sensitive; "growth at all costs" during zero-rate era didn't survive the rate normalisation. |
SME Lending, Microloans & Banks Entering the Space
S7 Slides 31–38| Dimension | Detail |
|---|---|
| What Marcus was | Goldman Sachs's consumer banking venture, launched 2016. Named after founder Marcus Goldman. Offered no-fee, fixed-rate personal loans (debt consolidation) and high-yield online savings accounts — digitally, with no branches. |
| Why it was significant | Marcus reached $1 billion in loans before any FinTech P2P lending platform reached the same milestone — demonstrating that a traditional bank with a strong brand and access to cheap deposit funding could out-execute pure-play FinTechs on cost of capital. |
| Competitive advantage | Goldman's AAA credit rating enabled borrowing at lower cost than any FinTech; its brand reassured deposit customers; its risk management infrastructure was world-class. What it lacked: consumer brand recognition, UX intuition, and patience for slow consumer relationship building. |
| What happened | Goldman wound down the consumer ambition (2022–2024) after billions in losses — not because the lending was bad, but because building a consumer bank is expensive, slow, and culturally alien to a firm built on institutional and investment banking. The lesson: access to capital is necessary but not sufficient for FinTech success. Customer acquisition, UX, and cultural alignment matter enormously. |
Microloans, Crowdfunding & the Future of Alternative Credit
S7 Slides 39–44| Dimension | Detail |
|---|---|
| What Tala does | Approves developing-world borrowers who lack a credit history for micro-loans of $10 to $500 — using 10,000+ data points extracted from the applicant's smartphone |
| The data model | Analyses: financial transactions (M-Pesa, bank SMS alerts), mobile game behaviour (decision-making patterns), communication patterns (frequency, consistency), app usage (productivity apps correlate with reliability), location patterns, and hundreds of other device signals |
| Why mobile game data matters | Counterintuitive insight: mobile game behaviour (how someone plays, their patience, whether they follow through, whether they accept losses gracefully) is genuinely predictive of loan repayment behaviour — demonstrating that the boundary between "financial data" and "behavioural data" is increasingly irrelevant |
| Markets | Kenya, Tanzania, Philippines, India, Mexico — countries where 60–80% of adults have smartphones but no credit history. Tala's total addressable market: the 1.7B unbanked adults globally who own smartphones. |
| Impact | Over $2B in loans disbursed; 90%+ repayment rates; repeat borrowers use Tala credit to build small businesses, cover health expenses, and smooth income volatility — the financial inclusion use case in practice |
| Model | Mechanism | Return to Backer | Key Platforms |
|---|---|---|---|
| Rewards-Based | Creator raises funds in exchange for a product, experience, or recognition — no financial return | Early access to product, special edition, acknowledgement | Kickstarter, Indiegogo |
| Equity Crowdfunding | Backers receive equity stake in the company in exchange for investment — regulated as securities offering | Potential capital gain / dividend if company succeeds | Seedrs, Crowdcube (UK/EU), Republic (US), Fundable |
| Donation-Based | Backers donate with no expectation of financial or product return — charitable or community motivation | Social impact, recognition | GoFundMe, JustGiving |
| Debt Crowdfunding (P2P) | Backers lend money expecting repayment with interest — this is the Lending Club model at a more retail/community level | Interest payments over loan term | Funding Circle (SME), Zopa (historical) |
The course material closes Session 7 by positioning the P2P lending story as chapter one of a longer story — the transformation of credit infrastructure:
| Evolution Stage | Key Characteristics | Representative Players |
|---|---|---|
| P2P 1.0 (2005–2015) | Retail investors fund retail borrowers; platforms as pure marketplaces; democratisation narrative; no balance-sheet risk | Early LendingClub, Prosper, Zopa, Funding Circle |
| P2P 2.0 (2015–2020) | Institutional investors replace retail; platform-to-institution model; underwriting standardisation; regulatory scrutiny | Late LendingClub, OnDeck, Kabbage, Avant |
| Credit Infrastructure (2020–) | Platforms become embedded credit rails for larger ecosystems; balance-sheet evolution or bank acquisition; BaaS-style credit APIs for non-financial platforms | Affirm (embedded at Shopify/Amazon), Stripe Capital (embedded SME credit), Shopify Capital, Klarna (full banking services), PayPal Working Capital |
Robo-Advisors: Democratising Asset Management
Wealthfront built an industry-defining automated investment service on MPT foundations — then discovered that fee compression and incumbents' distribution advantages are existential threats.
Key Numbers
Session 08 DataCast of Characters
Case: The Wealthfront Generation (HBS)Who's who in Wealthfront's story
The Investment Philosophy
MPT in Practice| Wealthfront Asset Class | Rationale | Typical ETF |
|---|---|---|
| US Stocks | Core growth engine — long-run equity premium | VTI (Vanguard Total Market) |
| International Stocks | Diversification — low correlation to US | VEU (Vanguard All World ex-US) |
| Emerging Markets | Higher growth potential, higher risk | VWO |
| Dividend Stocks | Value tilt + income | VIG |
| US Bonds | Ballast — negative correlation to equities in crashes | BND |
| TIPS | Inflation protection | VTIP |
| Real Estate (REIT) | Inflation hedge + income + low stock correlation | VNQ |
| Natural Resources | Commodity inflation hedge | XLE / DJP |
| Provider | Annual Fee | AUM Minimum | Differentiation |
|---|---|---|---|
| Wealthfront | 0.25% | $500 | TLH, direct indexing, tech UX |
| Betterment | 0.25% | $0 | Human advisor hybrid option |
| Vanguard VPAS | 0.30% | $50,000 | Brand trust + human advisor access |
| Schwab Intelligent | 0% | $5,000 | Free — cross-sell from brokerage |
| Fidelity Go | 0% | $10 | Free + Fidelity brand |
Exam Frameworks
Key ConceptsThe Wealthfront case implicitly raises a challenge that became explicit in the March 2020 COVID crash: stated risk tolerance (from a questionnaire) diverges dramatically from revealed risk tolerance (under real market stress).
| Investor Behaviour | MPT Assumption | Reality (DALBAR data) |
|---|---|---|
| Market timing | Investors hold through cycles | Average investor underperformed S&P by 4.32%/yr (1992-2011) from buying high/selling low |
| Risk tolerance | Stable across market conditions | Self-reported "aggressive" investors panic-sold in March 2020 crashes |
| Diversification adherence | Portfolios held as constructed | Investors override algorithms during crises — sell the "risky" assets at worst times |
How to Create a FinTech
From pain point to product-market fit to Series A — a practical framework for building a FinTech company that uses all the lessons from the course's previous cases.
The FinTech Creation Framework
Step by StepEvery successful FinTech in this course started with a genuine, observable pain point — not a feature idea. The pattern is consistent:
| Founder | Pain Point Observed | FinTech Built |
|---|---|---|
| Renaud Laplanche (LC) | 18% credit card rate vs 6.7% bank deposit — the spread made no sense | Lending Club: marketplace to eliminate the spread |
| David Vélez (Nubank) | Trapped in bulletproof door, treated as criminal, 450% credit card APR | Nubank: no-fee, digital-first credit card |
| Adalberto Flores (Kueski) | $4B in Mexican e-commerce lost to lack of credit access | Kueski: alternative data credit scoring for unbanked |
| Rachleff + Carroll (WF) | Sophisticated investing only accessible to the wealthy ($1M+ minimum) | Wealthfront: MPT-based investing from $500 |
| Peter Thiel (PayPal) | eBay sellers couldn't accept credit cards — friction in online commerce | PayPal: P2P payment network |
Every FinTech requires a rare combination of competencies that almost never exist in one person. The successful FinTechs in this course all show deliberate team architecture:
| Role | Why Essential | Example from Cases |
|---|---|---|
| Domain Expert (Finance) | Understands regulatory constraints, product design, risk management | Rachleff (VC/endowment), Laplanche (finance background), Flores (Ooyala finance ops) |
| Technical Architect (Engineer) | Builds the core product — must be exceptional, not just competent | Wible (Princeton CS + PE), Carroll (bond trader turned coder), Levchin (Palantir-level engineer) |
| Growth / Product | Understands user psychology, CAC, conversion, viral loops | Adam Nash (Facebook/LinkedIn growth experience) |
| Regulatory Navigator | FinTech is uniquely regulated — someone must own this from day 1 | LC's WebBank relationship, Nubank's BCB engagement, Kueski's CNBV licence |
FinTech GTM strategies differ fundamentally from traditional bank marketing (TV + branch + mass mail). The successful cases all use digital-first, low-CAC approaches:
| GTM Type | Mechanism | Example | Why it Works |
|---|---|---|---|
| Viral / Referral | Existing users invite friends; product scarcity creates desire | Nubank waitlist — invitations sold on Mercado Libre; 70% of apps via referral | Near-zero CAC; social proof; better credit quality from referred customers |
| Freemium | Free up to a threshold; premium features beyond | Wealthfront: first $25K free + $5K/referral; Dropbox model applied to finance | Removes risk for trial; viral when users refer to extend their own free tier |
| Content / Community | Educational content attracts target demographic organically | Wealthfront blog for tech workers; Nubank's transparency content | SEO + organic trust building; attracts exactly the right audience |
| Developer-Led | API-first product adopted bottom-up by developers | Stripe: documentation so good developers integrated it themselves, then recommended to employers | B2D2C: developers as distribution channel inside companies |
| Employer / Payroll | Integrate into payroll or HR systems to access entire workforces | Wealthfront 401K integration; employer financial wellness programmes | Captive distribution; large batch acquisition; trusted context |
| Regulatory Path | Pros | Cons | Used By |
|---|---|---|---|
| Banking as a Service (BaaS) — partner with licensed bank | Launch fast, no capital requirements, no banking licence | Dependency on partner, share economics, partner can exit | Early Nubank (Mastercard partner), many neobanks |
| E-Money Licence | Lighter than full banking licence, faster to obtain | Cannot offer interest-bearing deposits, limited product set | Revolut (initially), N26, Wise |
| Full Banking Licence | Full product flexibility, deposit insurance, lower funding cost | Capital requirements (Basel III), board/management requirements, years to obtain | WeBank (Tencent), LC (Radius acquisition), Nubank (Brazil) |
| Regulatory Sandbox | Test live product with real customers, supervised | Time-limited, jurisdiction-specific | UK FCA sandbox graduates including Monzo, Starling |
Financial Inclusion & Mobile Banking
2 billion unbanked adults. 1 billion of them with a mobile phone. For the first time in history, technology makes universal financial access achievable — but the unit economics and last-mile challenges are brutally hard.
Key Numbers
Session 10 DataCast of Characters
Case: Mobile Banking for the Unbanked (HBS)Who's who in the mobile money ecosystem
The Financial Inclusion Opportunity
Context & Economics| Barrier | Traditional Banking Response | Mobile Money Solution |
|---|---|---|
| Distance / Access | Physical branches — only economical in urban centres | Agent network at existing local shops; no branch needed |
| Minimum Balance | Minimum deposits exclude low-income users | No minimum balance on mobile money accounts |
| Documentation | Requires tax ID, proof of address, employer letter | Simplified KYC — national ID only for small amounts |
| Cost | Monthly fees, transaction fees, ATM fees | 35-85% cheaper per transaction than traditional banking |
| Literacy | Paper forms, written contracts, formal processes | Simple SMS/USSD interface; WIZZkid helps with setup |
| Trust | Banks seen as institutions of the elite; intimidating | Peer referral through WIZZkids; community distribution |
Agent banking (using non-bank agents as distribution points) is the key enabler of financial inclusion — but creates its own risks:
| Risk | Description | WIZZIT's Experience |
|---|---|---|
| KYC Compliance | Regulators require identity verification for every account — hard to enforce in rural areas | Compliance officers demanded hard copy ID photocopies; no photocopy machines in rural areas. 4 compliance officers in 4 years, each with different interpretations |
| Liquidity Risk | Agents may run out of cash, making withdrawals impossible on payday | Agents must maintain float; under-capitalised agents fail customers at worst moment |
| Agent Fraud | Agents may charge unauthorised fees or provide incorrect change to low-literacy customers | Community-based WIZZkids mitigated this — but not eliminated |
| Literacy | Customers cannot read; cannot verify their own transactions | WIZZkid training was expensive and slow; marketing agencies "of little help" reaching illiterate populations |
Strategic Frameworks
Key ConceptsThe single most powerful lever for financial inclusion is linking government benefit payments (G2P) to digital financial accounts. When governments pay wages, pensions, and social transfers digitally, they create millions of new account holders in one policy decision.
| Country | Programme | Impact |
|---|---|---|
| India | Jan Dhan Yojana + Aadhaar biometric ID + Mobile (JAM Trinity) | 450M+ new bank accounts opened; enabled direct benefit transfer to the poorest citizens |
| Kenya | Government payroll via M-Pesa | Eliminated cash wage packets (theft risk); forced agent network expansion into rural areas |
| Brazil | Bolsa Familia social transfer via Caixa digital accounts | Millions of low-income Brazilians' first formal financial account — then cross-sold to Nubank |
| Philippines | GCash (G-Xchange) + government disbursements | Telecoms-led mobile money achieving 60M+ registered users |
Regulation: Enabler or Constraint?
Lending Loop's collision with Canadian securities law reveals how a single regulatory interpretation can halt a FinTech — and how strategic regulatory navigation can become a competitive moat.
Key Numbers
Session 11 DataCast of Characters
Case: Lending Loop (Ivey)Who's who in Lending Loop's regulatory battle
The Regulatory Landscape
Core Concepts| Regulation | Jurisdiction | What it Does | FinTech Impact |
|---|---|---|---|
| PSD2 | EU (2018) | Requires banks to open payment APIs to authorised third parties (AISPs, PISPs) | Created Open Banking ecosystem — budget apps, instant account verification, bank-to-bank payments |
| Consumer Data Right (CDR) | Australia (2020) | Gives consumers right to share their data with authorised third parties | Equivalent of PSD2 for banking; extends to energy and telecoms |
| FCA Regulatory Sandbox | UK (2016) | Live testing with real customers under bespoke regulatory conditions | Launched 50+ companies including Monzo, Starling, Revolut — copied by 50+ regulators globally |
| FinTech Law (Mexico, 2018) | Mexico | First comprehensive FinTech law in LatAm — created regulatory framework for e-money, P2P lending, crowdfunding | Enabled Kueski and others to operate with legitimacy; also clarified regulatory expectations |
| MiCA (EU, 2024) | EU | Markets in Crypto-Assets Regulation — first comprehensive EU crypto framework | Creates legal certainty for crypto issuers and exchanges operating in EU |
Know Your Customer (KYC) and Anti-Money Laundering (AML) compliance are required of ALL financial services providers — but disproportionately burden small FinTechs.
| Dimension | Traditional Bank | Lending Loop (Marketplace) |
|---|---|---|
| Loan processing time | 2-6 weeks (Canada SBF Programme) | Days (digital application) |
| Branch overhead | 30-35% of operational costs | Near zero (fully digital) |
| Credit scoring | Traditional bureau data only | Alternative data + proprietary algorithms |
| Capital requirements | Basel III: significant regulatory capital | No balance sheet lending = no capital requirement |
| Regulatory compliance | Full banking regulation | Lighter (EMD only) — but this is a fragile advantage |
| Funding stability | Stable deposits | Investor-dependent — withdraws in stress |
Scalability in the Era of Exponential Growth
Tencent is the most dramatic example of platform-led scalability in financial services — from a messaging app to a $272B super-app handling 1B+ users' financial lives, all without a traditional bank licence.
Key Numbers
Session 12 DataCast of Characters
Case: Tencent (HBS)Who's who in Tencent's evolution
Platform Scalability: Tencent's Model
Core Framework| 1998-2003 | QQ messaging launched. "Closed garden" strategy — keep users on-platform. 135M active users by end of period. |
| 2004-2009 | IPO on Hong Kong Stock Exchange. Expands into online gaming (massive; becomes revenue dominant), social networking (Qzone). Strained telco relationships as state-owned operators see Tencent as competitor. |
| 2010-2013 | Begins "opening the walled garden." WeChat launched 2011 for smartphones. Strategic investments in JD.com ($215M for 15% stake) and Sogou search. Acknowledges it cannot compete everywhere — better to invest in leaders and integrate. |
| 2014-2016 | "Connection" strategy. Red Envelope CNY moment. WeChat Pay explosive growth. Mobile QQ Wallet. Mini-programs strategy conceived. 2016 revenue: ¥152B (+48%). Becomes most valuable company in Asia at $272B. |
| 2015+ | WeBank digital banking licence. WeStar credit scoring. Microlending at scale. Financial services become a major revenue segment alongside gaming and social. |
| Dimension | Traditional Bank | Tencent / WeBank |
|---|---|---|
| Customer acquisition | $100-300 CAC (branch + marketing) | Near zero — WeChat users already exist |
| Credit data | Bureau data (formal history only) | WeChat social graph + spending + behavioural data (360° view) |
| Branch infrastructure | Enormous fixed cost (30-35% of ops) | Zero (digital-only WeBank) |
| NPL ratio | Big-four Chinese banks: ~1.5-2% | WeBank reportedly ~0.3% (better data = better underwriting) |
| Product cross-sell | Requires separate engagement per product | Single app: pay, borrow, invest, insure — all in one session |
Partnerships with Banks: A Win-Win Game
Scotiabank's journey from dismissing FinTechs to building the Digital Factory and partnering with Kabbage reveals how established banks can innovate without the full risks of acquisition or in-house build.
Key Numbers
Session 13 DataCast of Characters
Case: Scotiabank Innovation (Ivey)Who's who in Scotiabank's FinTech journey
The Partnership Spectrum
Strategic Options| Model | Structure | Pros | Cons | Example |
|---|---|---|---|---|
| Technology Licensing | Bank pays FinTech a licence fee to use their technology/platform | Fast to deploy; no equity dilution for FinTech; bank controls brand | Bank dependent on FinTech; FinTech can sell same tech to competitors | Scotiabank + Kabbage; JPMorgan + OnDeck; Banco Santander + Ripple |
| Minority Investment | Bank takes 5-20% equity stake + commercial agreement | Alignment of interests; information rights; acquisition option; FinTech retains independence | Bank cannot direct strategy; minority can be diluted in later rounds | Scotiabank + QED (VC co-investment); most bank "innovation fund" investments |
| Joint Venture | Bank and FinTech create a new legal entity together | Shared risk; combines strengths; can access markets neither can alone | Complex governance; potential conflict on strategy and exit | Singapore MAS + Government of Andhra Pradesh (blockchain JV) |
| Acquisition | Bank buys 100% of FinTech | Full control; IP ownership; talent retention; rapid integration if done well | Cultural clash; integration complexity; acqui-hire risk; premium price | Scotiabank acquires Tangerine (ING Canada); BBVA acquires Simple; Blackrock acquires FutureAdvisors |
The case is candid about the organisational barriers that make bank-FinTech partnerships harder than they look on paper:
| Challenge | Bank Perspective | FinTech Perspective |
|---|---|---|
| Procurement Speed | Vendor due diligence: 6-18 months (security review, legal, committee) | 12-month runway cannot survive 18-month sales cycle — may pivot or run out of cash |
| Risk Culture | "Avoid failures, protect the institution" — slow, deliberate, consensus-driven | "Fail fast, iterate" — speed is survival; failure is learning |
| Technology Stack | Legacy core banking systems — integration is expensive and slow | Cloud-native, API-first — designed to integrate, frustrated by legacy barriers |
| Success Metrics | Measured on risk-adjusted return, compliance, long-term profitability | Measured on growth rate, user acquisition, product velocity |
| Contract Terms | Requires exclusivity, IP ownership, liability caps, extensive data clauses | Cannot give exclusivity without losing other bank clients; IP is the core asset |
| Business Line | Recommended Approach | Rationale |
|---|---|---|
| Digital SMB Lending | License + partner (Kabbage model) | FinTech has proven superior technology; bank has balance sheet and distribution. No need to build or acquire. |
| Digital Retail Payments | Build in-house + minority invest | Core customer relationship; cannot outsource. Invest in FinTechs for insights and optionality. |
| LatAm FinTech exposure | VC co-investment (QED model) | High uncertainty; portfolio approach appropriate. Bank gets insights without concentrated bets. |
| Digital banking platform | Acquire and preserve (Tangerine model) | Proven product + customer base; integrate distribution while preserving culture. |
| Blockchain / Trade Finance | Proof of concept + consortium | Still early stage; industry consortia (R3, we.trade) share development cost. |
FinTech Simulators & Calculators
Seven interactive tools to build intuition for the mathematical concepts underlying FinTech — from blockchain hashing to DeFi yield, AMM pricing, and portfolio construction.
Blockchain Hash Demo
See how SHA-256 changes with a 1-bit flip. Visualise immutability and the avalanche effect.
AMM / Uniswap Price Calculator
Explore the x·y=k constant-product formula and visualise price impact & slippage.
BNPL vs Credit Card Cost
Compare the true cost of BNPL instalment plans against revolving credit card debt.
Stablecoin Collateral Ratio
Model how crypto-backed stablecoins maintain their peg and the liquidation cascade risk.
Robo-Advisor Allocator
Build a Modern Portfolio Theory portfolio across risk profiles and see the efficient frontier.
DeFi APY ↔ APR Converter
Convert between APR and APY with compounding frequency — understand yield farming returns.
Tokenomics Supply Calculator
Model token inflation schedules, vesting cliffs, and circulating supply over time.
Blockchain Hash Demo
Type any message and see its SHA-256 hash. Change a single character and observe how the entire hash changes — demonstrating the avalanche effect and immutability.
AMM / Uniswap Price Calculator
The constant-product formula x · y = k governs all trades in Uniswap V2. Explore price impact, slippage, and liquidity depth.
BNPL vs Credit Card Cost Comparison
Compare the total cost of "Buy Now Pay Later" instalment plans against revolving credit card debt — revealing hidden fees and the true APR.
Stablecoin Collateralisation Ratio
Model how crypto-backed stablecoins like DAI maintain their peg, and see at what collateral ratio liquidation is triggered.
Robo-Advisor Portfolio Allocator
Select a risk profile and see how Modern Portfolio Theory drives the asset allocation. Adjust expected returns and volatility to understand efficient frontier logic.
DeFi APY ↔ APR Converter
Understand the difference between APR (simple annual rate) and APY (effective annual rate with compounding). DeFi protocols often advertise APY — this matters.
Tokenomics Supply & Inflation Calculator
Model token emission schedules, vesting cliffs, team unlock events and their impact on circulating supply and price pressure.
FinTech Glossary
Key terms organised by topic — matching the course's 13 sessions. Use this for quick exam reference and case preparation.
Blockchain & Cryptocurrency
Sessions 02-03| Term | Definition |
|---|---|
| Blockchain | A distributed, append-only ledger of transactions organised into cryptographically linked blocks. Tamper-evident because changing any block invalidates all subsequent blocks, requiring re-doing more proof-of-work than the honest network produces. |
| Proof-of-Work (PoW) | A consensus mechanism where miners compete to find a nonce that makes the block's SHA-256 hash begin with a required number of zero bits. Computationally expensive to produce, trivially cheap to verify — the security asymmetry that makes Bitcoin tamper-resistant. |
| Proof-of-Stake (PoS) | Alternative consensus mechanism where validators are chosen based on the amount of cryptocurrency they "stake" (lock up as collateral). Far more energy-efficient than PoW. Ethereum switched from PoW to PoS in "The Merge" (2022). |
| SHA-256 | Cryptographic hash function used in Bitcoin. Takes any input and produces a deterministic 256-bit output. Even a 1-bit change in input produces a completely different output (avalanche effect) — making blockchain records tamper-evident. |
| Nonce | A "number used once" — the variable miners increment when searching for a valid block hash. The mining process is essentially: increment nonce, hash the block header, check if result meets difficulty target; repeat until it does. |
| Mining / Miner | The process and participants who validate Bitcoin transactions by finding valid proof-of-work. Rewarded with block subsidy (new Bitcoin) + transaction fees. Halvings reduce the subsidy every ~4 years until ~2140. |
| Halving | Scheduled reduction of the Bitcoin block reward by 50% every 210,000 blocks (~4 years). 2009: 50 BTC; 2012: 25 BTC; 2016: 12.5 BTC; 2020: 6.25 BTC; 2024: 3.125 BTC. Hard-coded into the protocol — maximum 21M Bitcoin ever. |
| Public / Private Key | Cryptographic key pair. Public key = Bitcoin address (shared freely). Private key = proof of ownership (never shared). To spend Bitcoin, the owner proves possession of the private key that unlocks the public key — without revealing the private key itself. |
| Double-Spending | The core problem Bitcoin solved without a trusted third party. Sending the same digital token to two recipients simultaneously. Bitcoin's blockchain provides a shared, tamper-resistant history of all transactions — making double-spending computationally infeasible if honest miners control majority CPU power. |
| 51% Attack | If a single entity controls more than 50% of a blockchain's hash rate, they can rewrite recent transaction history (e.g., reverse their own payments). Probability of success drops exponentially as more blocks are added — the Nakamoto security model. |
| Smart Contract | Self-executing code stored on a blockchain that automatically enforces the terms of an agreement when predetermined conditions are met. No intermediary required. Enables DeFi protocols, NFTs, DAOs. Ethereum is the dominant smart contract platform. |
| Gas / Gas Fees | The fee paid to Ethereum validators for executing a transaction or smart contract. Denominated in "gwei" (fractions of ETH). Gas price fluctuates with network demand — high during peak DeFi activity, making small transactions uneconomical. |
DeFi (Decentralised Finance)
Session 03| Term | Definition |
|---|---|
| DeFi | Decentralised Finance — financial services (lending, exchange, yield, insurance) built on open blockchains (primarily Ethereum), accessible to anyone with a wallet, governed by smart contracts rather than institutions. No KYC, no banks, no custodians. |
| AMM (Automated Market Maker) | A type of decentralised exchange that uses a mathematical formula (typically x·y=k) to set prices and enable trades without a traditional order book. Liquidity providers deposit token pairs; traders pay a small fee on each swap. |
| Constant Product Formula | x · y = k — the invariant used by Uniswap V2. x and y are the reserves of two tokens in a pool; k is constant. Any trade moves along the curve, with larger trades producing worse execution prices (price impact). |
| Liquidity Provider (LP) | A user who deposits equal-value amounts of two tokens into an AMM pool. Earns a share of trading fees proportional to their contribution. Exposed to impermanent loss if token prices diverge significantly. |
| Impermanent Loss | The opportunity cost LPs face compared to simply holding the underlying tokens. When token prices diverge, the AMM's constant-product formula auto-rebalances in a way that leaves LPs worse off than holding. "Impermanent" because it disappears if prices return to entry levels. |
| Stablecoin | A cryptocurrency designed to maintain a stable value relative to a reference asset (usually USD). Three types: fiat-backed (USDC — 1:1 USD reserve), crypto-collateralised (DAI — over-collateralised with ETH), algorithmic (e.g., TerraUSD — collapsed 2022). |
| DAI / MakerDAO | DAI is a crypto-collateralised stablecoin. Users lock ETH (or other tokens) worth at least 150% of the DAI minted in a "vault" (CDP). If the collateral ratio falls below 150%, automated liquidators can seize and sell the collateral at a 13% discount. |
| Collateralisation Ratio (CR) | CR = (Collateral Value / Stablecoin Debt) × 100%. MakerDAO minimum: 150%. A CR of 200%+ is considered safe. Below the liquidation ratio, the vault is automatically liquidated. |
| Yield Farming | Actively moving capital across DeFi protocols to maximise return — combining lending rates, liquidity mining rewards, and leveraged positions. The crypto equivalent of a currency carry trade. High potential returns come with compounded risks: smart contract, liquidation, governance, and gas. |
| Governance Token | A cryptocurrency representing ownership of a DeFi protocol and voting rights on its parameters (fees, risk levels, upgrades). Examples: MKR (MakerDAO), COMP (Compound), UNI (Uniswap). Resembles equity but without legal shareholder protections. Risk: token-weighted voting creates plutocracy. |
| Total Value Locked (TVL) | The total value of crypto assets deposited in DeFi protocols — used as a market size/adoption metric. Subject to double-counting (same assets used as collateral multiple times). Peaked at ~$180B before the May 2021 crypto crash. |
| Flash Loan | An uncollateralised loan that must be borrowed and repaid within a single blockchain transaction. If not repaid, the entire transaction is reversed. Used by arbitrageurs and — notoriously — attackers to manipulate DeFi protocol prices. |
| DAO (Decentralised Autonomous Organisation) | An organisation governed by smart contracts and token-holder votes rather than a board of directors. Rules are encoded in code; decisions are made by governance token holders. MakerDAO governs the DAI stablecoin protocol. |
| CBDC (Central Bank Digital Currency) | A digital form of central bank money — state-issued, programmable, and directly distributed. Retains monetary sovereignty while co-opting blockchain's properties. Key risk: a retail CBDC held directly at the central bank disintermediates commercial banks by removing their retail deposit funding base. The ECB's Digital Euro design includes a €3,000 holding cap per person to prevent this. |
| MiCA (Markets in Crypto-Assets Regulation) | The EU's comprehensive digital assets framework (in force 2023–24) — the world's first unified crypto regulation. Covers stablecoins (asset-referenced tokens, e-money tokens) and general crypto-assets. Imposes reserve requirements and issuance caps on "significant" stablecoins. Contrasts with the US's fragmented, enforcement-led approach, creating regulatory arbitrage opportunities. |
| Nostro / Vostro Accounts | Pre-funded accounts banks hold at each other to lubricate the SWIFT correspondent banking network. Collectively estimated at ~$27 trillion globally — capital that earns no return and creates liquidity risk. Blockchain's atomic settlement eliminates the pre-funding requirement, freeing this trapped liquidity. |
| Web 3.0 / Internet of Value | The third internet generation where users read, write, AND own their data and assets via on-chain cryptographic proofs. Web 1.0 = read-only. Web 2.0 = read-write (platforms own the data). Web 3.0 = trustless ownership — value accrues to protocol token holders, not corporate shareholders. DeFi is the financial infrastructure of Web 3.0. |
| Libra / Diem | Facebook's failed 2019 attempt at a global basket-backed stablecoin. Regulatory collapse driven by three simultaneous threats: data sovereignty (Facebook combining payment + social data), monetary sovereignty (private money at central bank scale), and systemic risk (reserve mismanagement implications). Consortium dissolved 2019–21; project sold to Silvergate Bank for ~$200M in January 2022. |
| SAB 121 | US SEC Staff Accounting Bulletin 121 (2022): banks holding crypto in custody must recognise it as both an on-balance-sheet asset AND liability at fair value — dramatically increasing capital requirements. Puts regulated banks at a disadvantage vs unregulated crypto custodians. Congress passed a resolution to overturn it in 2024; status remains contested. |
| DApp (Decentralised Application) | An application whose backend logic runs on blockchain smart contracts rather than centralised servers. The contract is the tamper-resistant backend; front-end interfaces may remain centralised. True decentralisation requires the contract itself to be the canonical interface. |
Payments
Sessions 05-06| Term | Definition |
|---|---|
| Payment Rail | The underlying infrastructure over which money moves — the "pipes" of the financial system. Key rails: ACH (batch, low-cost, 1-2 day), SWIFT (international wire), card networks (Visa/MC), and real-time payments (RTP, FedNow, SEPA Instant). FinTechs build on top of rails or build new ones. |
| Interchange Fee | The fee paid by the merchant's bank (acquirer) to the cardholder's bank (issuer) on each card transaction. Typically 1.5-2.0% of transaction value for credit cards. The primary revenue source for card issuers. Funded by merchant discount fees — ultimately a cost passed to all consumers via higher prices. |
| Merchant Discount Rate | The total fee merchants pay to accept card payments — typically 2-3%. Comprises interchange (~1.5-2%), network assessment (~0.10%), and acquirer margin (~0.2-0.5%). PayPal charges 2.9% + $0.30 for standard transactions. |
| ACH (Automated Clearing House) | US batch payment system connecting financial institutions via the Federal Reserve. Used for payroll, bill payments, P2P transfers. Cost: ~$0.05 per transaction. Delay: 1-2 business days. PayPal actively incentivises ACH over card funding to reduce its cost of funds. |
| Two-Sided Market | A platform serving two distinct user groups whose presence makes the platform more valuable to the other side. PayPal: consumers and merchants. Visa: cardholders and merchants. Classic chicken-and-egg problem: neither side joins without the other. |
| Network Effect | The property where a product or service becomes more valuable as more people use it. Direct: P2P payments (Venmo). Cross-side: Visa card (more merchants = more valuable to consumers; more consumers = more valuable to merchants). Creates winner-takes-most dynamics. |
| Chargeback | A disputed transaction that a consumer reverses through their card issuer — without the merchant's agreement. Merchants bear chargeback risk for card-not-present transactions. PayPal's fraud rate (0.17%) vs card-not-present (1.8%) is a key merchant value proposition. |
| BNPL (Buy Now Pay Later) | An instalment product allowing consumers to split a purchase into 3-6 equal payments, typically interest-free if paid on time. Revenue model: merchant fees (2-8%) + late payment fees. Providers: Klarna, Afterpay, Affirm. Key risk: consumer over-indebtedness; regulatory scrutiny increasing. |
| Neobank / Challenger Bank | A digital-only financial institution with no physical branches. Lower cost structure enables competitive pricing. Often targets underserved demographics. Examples: Revolut, N26, Chime, Nubank, Monzo. May operate under e-money licence (lighter) or full banking licence. |
| Embedded Finance | Financial products (payments, insurance, lending) integrated directly into non-financial products and platforms. Shopify merchant loans, Uber driver insurance, Amazon seller working capital. Enabled by Banking-as-a-Service (BaaS) platforms. |
| Super-App | A single mobile application combining multiple services — messaging, payments, shopping, investing, insurance — to make it the primary interface for users' digital lives. Examples: WeChat (China), Grab (Southeast Asia). Creates near-infinite switching costs. |
Lending & Credit
Sessions 04, 07| Term | Definition |
|---|---|
| P2P Lending (Marketplace Lending) | A platform that matches borrowers with investors who fund those loans. The platform earns origination and servicing fees without holding credit risk on its balance sheet. Key structural risk: investor appetite is pro-cyclical — withdraws in stress, destabilising originations. |
| Thin File / No File | A consumer with no or minimal formal credit history — making it impossible for traditional credit bureaus to generate a reliable score. Affects the underbanked, young adults, immigrants, and gig workers. FinTechs like Kueski use alternative data to score these "invisible" populations. |
| Alternative Data | Non-traditional information used in credit scoring when bureau data is absent or thin. Examples: device metadata, email age/provider, social network patterns, GPS mobility, typing speed, rental payment history. Risk: may correlate with protected characteristics (proxy discrimination). |
| CAC (Customer Acquisition Cost) | Total sales and marketing spend divided by new customers acquired in a period. For FinTechs: CAC must be recovered by Customer Lifetime Value (CLV). Nubank achieved near-zero CAC via referral virality — the gold standard. |
| CLV (Customer Lifetime Value) | The total profit a company expects to earn from a customer over the duration of the relationship. Key formula: (ARPU × Gross Margin) / Churn Rate. Healthy FinTech: CLV/CAC > 3x. Below 1x = destroying value. |
| Origination Fee | An upfront fee charged to borrowers when a loan is issued, typically 1-6% of the loan amount. Lending Club charged 1-5%; Lending Loop charged 3.5-6.5%. Deducted from loan proceeds. Primary revenue source for marketplace lenders alongside servicing fees. |
| NIM (Net Interest Margin) | The difference between interest income earned on loans and interest paid on deposits, expressed as a percentage of earning assets. Banks earn NIM; pure marketplace lenders earn fees (no NIM, no credit risk). Higher NIM implies more balance-sheet risk. |
| NPL (Non-Performing Loan) Ratio | The percentage of a loan portfolio where borrowers have missed payments beyond a threshold (typically 90 days). Key indicator of credit quality. WeBank: ~0.3% (data-advantage). Chinese big four: ~1.5-2%. High NPL ratios erode bank capital and profitability. |
| Proxy Discrimination | When an alternative data variable correlates with a protected characteristic (race, gender, religion) even if that characteristic is not used directly. Using zip code or email provider can inadvertently produce racially discriminatory lending outcomes — illegal under fair lending law even without intent. |
Wealth Management & Investing
Session 08| Term | Definition |
|---|---|
| MPT (Modern Portfolio Theory) | Harry Markowitz's 1952 framework showing that diversification reduces portfolio variance without proportionally reducing expected return — because assets with low/negative correlation offset each other's volatility. The mathematical foundation of all robo-advisor allocation. |
| Efficient Frontier | The set of portfolios offering the highest expected return for a given level of risk (variance), or the lowest risk for a given expected return. Portfolios below the frontier are inefficient — dominated by better alternatives. Robo-advisors place users on the frontier based on their risk score. |
| Sharpe Ratio | S = (E(Rp) - Rf) / σp. Measures risk-adjusted return: excess return above the risk-free rate per unit of volatility. Higher = better. A portfolio with 10% return, 3% Rf, and 5% σ has Sharpe = 1.4 — generally considered good. |
| Tax-Loss Harvesting (TLH) | Selling assets trading below cost to realise a tax loss, then immediately buying a "substantially similar" replacement to maintain exposure. Generates tax savings that compound over time. Robo-advisors execute this daily at scale — far more efficiently than human advisors who review quarterly. |
| Wash-Sale Rule | IRS rule prohibiting the purchase of a "substantially identical" security within 30 days before or after selling it at a loss. FinTechs navigate this by substituting similar but not identical ETFs (e.g., Vanguard Total Market for iShares Core S&P Total Market). |
| Direct Indexing | Holding the individual stocks comprising an index (e.g., all 500 S&P 500 stocks) directly in an account, rather than via an ETF. Enables granular tax-loss harvesting on individual positions, custom ESG screens, and factor tilts. Wealthfront's premium offering; minimum ~$100K AUM. |
| Robo-Advisor | An automated investment service that builds and rebalances diversified portfolios based on a client's risk tolerance questionnaire. Uses MPT and index ETFs. Charges 0.25-0.50% AUM — vs 1%+ for human advisors. Examples: Wealthfront, Betterment, Schwab Intelligent Portfolios. |
| AUM (Assets Under Management) | The total market value of assets a financial institution manages on behalf of clients. Fee income = AUM × annual fee rate. Primary revenue metric for wealth managers and robo-advisors. Wealthfront grew from <$100M to $1.5B AUM in under 2 years (2013-14). |
| Fiduciary Standard | The legal obligation to act in a client's best interest. Registered Investment Advisors (RIAs) — including robo-advisors — are fiduciaries. Traditional broker-dealers operated under the weaker "suitability" standard. Eliminates commission-based product conflicts of interest. |
Regulation & Compliance (RegTech)
Sessions 09, 11| Term | Definition |
|---|---|
| KYC (Know Your Customer) | Regulatory requirement to verify customer identity, source of funds, and assess money laundering risk before establishing a financial relationship. The compliance foundation of every licensed financial institution. Costly for FinTechs — KYC-as-a-Service (Jumio, Onfido) helps amortise the cost. |
| AML (Anti-Money Laundering) | Regulatory framework requiring financial institutions to monitor transactions for suspicious activity patterns and report them to authorities. Ongoing compliance obligation — not just at account opening. Heavy fines for violations (HSBC: $1.9B in 2012; Standard Chartered: $1.1B in 2019). |
| PSD2 (Payment Services Directive 2) | EU regulation (2018) requiring banks to open their payment infrastructure to authorised third parties via APIs. Created the Open Banking ecosystem. Two key roles: AISP (Account Information Service Provider — read-only data access) and PISP (Payment Initiation Service Provider — initiate payments from bank accounts). |
| Open Banking | The framework enabling customers to share their bank account data and initiate payments with authorised third parties via APIs. With customer consent. Creates infrastructure for budgeting apps, instant account verification, and bank-to-bank payments. Mandated by PSD2 (EU), CDR (Australia), and emerging regulations globally. |
| Regulatory Sandbox | A controlled environment where FinTechs can test innovative products with real customers under supervised, bespoke regulatory conditions — without full compliance with normal requirements. Pioneered by the UK FCA (2016). Enables innovation while protecting consumers. Copied by 50+ regulators globally. |
| Exempt Market Dealer (EMD) | Canadian securities registration category. Allows firms to sell "securities" (including loan notes) to investors without filing a full prospectus — but restricts investor eligibility to "accredited investors" or those covered by specific exemptions (e.g., Offering Memorandum). Lending Loop's forced regulatory category. |
| Accredited Investor | A regulatory designation for individuals deemed sophisticated enough to bear higher investment risk. Canada/US typically require $1M+ in financial assets or $200K+ annual income. Restricts who can invest in exempt market securities — undermining marketplace lending's democratisation mission. |
| Regulatory Arbitrage | Exploiting gaps or inconsistencies between regulatory frameworks to gain competitive advantage. Example: operating as an e-money institution (lighter regulation) while offering products functionally equivalent to bank deposits. Regulators increasingly closing these gaps with "same activity, same risk, same regulation." |
| Howey Test | US Supreme Court test for whether an instrument is a "security": (1) investment of money, (2) in a common enterprise, (3) with expectation of profits, (4) from the efforts of others. If all four prongs satisfied, SEC jurisdiction applies. Used to evaluate whether crypto tokens are securities. |
| MiCA | Markets in Crypto-Assets Regulation — EU's comprehensive crypto regulatory framework (effective 2024). Creates licensing requirements for crypto asset issuers and service providers. First major jurisdiction to provide clear, comprehensive rules for the crypto industry. |
Financial Inclusion & Mobile Money
Session 10| Term | Definition |
|---|---|
| Unbanked | Adults who do not have an account at a financial institution or through a mobile money provider. Estimated 1.4 billion globally (World Bank, 2021). Concentrated in Sub-Saharan Africa, South Asia, and Latin America. Primary barriers: distance, cost, documentation requirements, distrust. |
| Mobile Money | A service that allows users to store, send, and receive money using a basic mobile phone — without a bank account. Works via SMS/USSD. Enabled by a network of cash-in/cash-out agents. M-Pesa (Kenya) is the global benchmark: launched 2007, reached 40M+ users. |
| Agent Banking | The use of non-bank agents (shops, kiosks, post offices) as distribution points for financial services — account opening, deposits, withdrawals, bill payments. Enables financial services to reach remote populations without building bank branches. Key risks: agent liquidity, fraud, regulatory compliance. |
| Last Mile | The challenge of reaching the most remote, low-income, or marginalised populations who remain outside the formal financial system even after mainstream FinTech expansion. Requires fundamentally different business models: agent networks, feature phones, voice interfaces, community intermediaries. |
| G2P (Government-to-Person) | Government payments to individuals — salaries, pensions, social transfers. Linking G2P disbursements to digital accounts is the fastest path to financial inclusion (India's JAM Trinity: Jan Dhan + Aadhaar + Mobile). Creates the "first transaction" that activates new accounts at scale. |
| BOP (Base of the Pyramid) | The 4+ billion people living on less than $8/day, representing the lowest economic tier globally. Historically excluded from formal finance. FinTech mobile money and alternative credit scoring represent the first genuine pathway to serving this population profitably. |
| M-Pesa | Mobile money service launched by Safaricom/Vodafone in Kenya (2007). Repurposed Safaricom's airtime agent network as cash-in/cash-out points. Works on basic SMS. Benchmark case for mobile money success. By 2012, processing more transactions per day than the entire Western Union global network. |
| Remittance | Money sent by migrant workers to family in their home countries. Global flows: ~$700B/year to developing countries. Average cost: ~6% of transaction value (World Bank target: <3% by 2030). FinTechs (Wise, Remitly, WorldRemit) reduced costs by 60-70% vs traditional providers. |
Platform Economics & Strategy
Sessions 01, 12, 13| Term | Definition |
|---|---|
| Platform Business | A business that creates value by facilitating interactions between two or more distinct user groups. Does not produce the product/service itself — creates the infrastructure for others to transact. Visa, PayPal, Lending Club, Uniswap are all platforms. Scale advantages compound through network effects. |
| Winner-Takes-Most | Market structure where the leading platform captures a disproportionate share of value due to network effects and switching costs. Not necessarily winner-takes-all — multiple platforms can coexist (Visa + Mastercard). FinTech payments exhibit strong winner-takes-most dynamics. |
| BaaS (Banking as a Service) | A model where licensed banks provide their banking infrastructure (accounts, payments, compliance) via APIs to non-bank FinTechs — who build financial products on top without their own banking licence. Enables rapid product launch; creates bank dependency. Providers: Railsbank, Synapse, Green Dot, Cross River Bank. |
| API (Application Programming Interface) | A standardised interface that allows different software systems to communicate. API-first banks expose their infrastructure to third parties. PSD2 mandated bank APIs in Europe. Stripe's developer-friendly API was itself the product — making payment integration trivial for merchants. |
| Innovator's Dilemma | Clayton Christensen's framework: successful companies are rationally optimised for their current customers — making it structurally difficult to invest in lower-margin disruptive innovations that threaten existing revenue. Perfectly describes why banks struggle to self-disrupt. |
| Disruptive Innovation | Innovation that initially serves a low-end or overlooked market segment with a simpler, cheaper product — then improves until it disrupts the mainstream market. FinTechs targeting the unbanked, thin-file borrowers, or young investors are classic disruptive innovators. |
| Technology Licensing | A partnership model where a bank pays a FinTech a fee to use their technology/platform under the bank's brand. Example: Scotiabank + Kabbage (SMB lending). JPMorgan + OnDeck. Fast to deploy; bank controls brand; FinTech can sell same technology to competitors. |
| Digital Factory | An organisational structure separating innovation teams from the main bank — with different physical space, talent profiles, and processes (agile/lean). Enables innovation velocity while maintaining connection to the bank's distribution and balance sheet. Scotiabank's Digital Factory is the case example. |
Quiz Bank
130+ questions across all 13 sessions · Easy / Medium / Hard