How the market values a fintech

Key terms:

• TPV (Total Payment Volume): the total dollar amount processed. Adyen, for example, reports TPV in the hundreds of billions of euros annually.
• Take rate: the fee the processor keeps, expressed as a percentage of TPV. Take rates are typically well under 1 percent for large processors.

Revenue = TPV x take rate. Because the processor does not hold credit risk, its revenue is high quality and recurring. Volume grows with commerce. Margins can be very high once the infrastructure is built.

So the market prices these companies on EV/Revenue (enterprise value divided by revenue) or EV/EBITDA (a proxy for cash operating profit). The logic: each dollar of revenue is durable, scalable, and carries little downside risk. Investors will pay a premium multiple for that.

This is why a company like Adyen or Visa can trade at revenue multiples far above what a bank would ever fetch. Visa's operating margins routinely exceed 60 percent, a level no lender can approach because a lender must reserve capital against losses.

Why lenders trade on book value

Now flip to the lender. A digital lender's value is not really about revenue. It is about the assets it owns and the risk attached to them.

Key terms:

• Book value: the accounting value of a company's equity, meaning assets minus liabilities. For a lender, the loan portfolio is the core asset.
• P/B (Price-to-Book): market value divided by book value. This is the primary lens for banks and lenders.
• ROE (Return on Equity): net profit divided by equity. It measures how efficiently the lender earns on its capital base.

Here is the core rule for balance-sheet businesses:

> A lender trades above book value (P/B above 1) only if it earns an ROE above its cost of equity. Otherwise it trades at or below book.

Why? Because a bank's revenue is not clean. Interest income can be wiped out by loan losses. In a recession, a "profitable" lender can swing to a loss overnight as borrowers default. Revenue tells you almost nothing about risk. Book value, adjusted for expected losses, tells you what the equity is actually worth.

This is why most lenders, including fintech lenders, trade at low single-digit revenue multiples and are judged instead on P/B and ROE. The market is asking one question: can this company earn a good return on the capital it has tied up?

The FDIC's quarterly banking profile is a free way to see aggregate industry ROE and asset quality trends, which anchor how the whole lending sector gets valued.

The comps, side by side

Let's make the gap concrete with real, well-known public fintechs. Exact multiples move daily, so treat these as illustrative of the *pattern*, not live quotes.

| Company | Model | Primary multiple | Why |

|---|---|---|---|

| Visa / Mastercard | Payments network | High EV/Revenue, high EV/EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.Voir la définition complète → | Asset-light, no credit risk, huge margins |

| Adyen | Payments processor | High EV/Revenue | Growth plus asset-light take-rate model |

| PayPal | Payments plus some credit | Moderate EV/Revenue | Mostly asset-light, some lending drags multiple |

| SoFi | Digital lender plus bank | P/B and P/E | Holds loans and deposits, credit risk on books |

| LendingClub | Digital bank / lender | P/B | Balance-sheet lender, valued on equity and ROE |

Notice PayPal in the middle. Companies that blend both models get a blended valuation. Analysts often value the segmentssegmentsDividing a market into distinct groups of customers who share similar needs, characteristics or behaviours, so each group can be served with a tailored approach.Voir la définition complète → separately (a "sum of the parts" approach): the payments arm on revenue multiples, the credit arm on book value, then add them together. When a payments company starts lending heavily, the market can actually mark down its multiple, because it is importing credit risk into a previously clean model.

🎬 [VIDEO: "How to Value a Fintech Company" — youtube.com — a walkthrough of revenue-multiple versus book-value approaches for financial businesses]

The deeper reason: risk and capital

Step back. The valuation gap is really about two things.

1. Capital intensity. An asset-light processor needs little capital to grow. Double the volume, and you barely add cost. A lender must hold regulatory capital against every loan. To double the loan book, it must roughly double its equity or funding. Growth is expensive and constrained.

Regulatory capital is the minimum equity cushion regulators require a lender to hold against potential losses (the Basel framework governs this globally). This requirement is precisely why lenders cannot scale like software: the balance sheet is the bottleneck.

2. Earnings quality. Processor revenue is recurring and low-risk. Lender revenue is cyclical and loss-prone. Markets pay premium multiples for predictable cash flows and discount volatile ones. A dollar of Visa earnings is simply worth more than a dollar of a subprime lender's earnings, because it is far more likely to still be there next year.

Put simply: the market pays for certainty and scalability. Asset-light models offer both. Balance-sheet models offer neither, so they get priced on the hard floor of what their equity is worth.

How to apply this as an analyst

When you look at any fintech, run this quick diagnostic:

1. Does it hold financial risk on its own balance sheet? Look for a loan book, deposits, or insurance reserves. If yes, lean toward P/B and ROE. If no, lean toward EV/Revenue and EV/EBITDAEBITDAEBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) measures a company's operating profitability before financing and accounting decisions, used to compare core performance across firms.Voir la définition complète →.

2. What is the revenue quality? Recurring take-rate fees justify a premium. Interest spread income that depends on credit performance justifies a discount.

3. For blended models, split it. Value the asset-light and balance-sheet segmentssegmentsDividing a market into distinct groups of customers who share similar needs, characteristics or behaviours, so each group can be served with a tailored approach.Voir la définition complète → separately, then sum. Do not slap one multiple on the whole thing.

4. Sanity-check with a comp set. Pull three or four public peers with the same model and compare. A lender priced like a payments network, or vice versa, is either mispriced or misunderstood.

A common trap: a neobank markets itself as a "tech company" to earn a software multiple, but if the bulk of its revenue is net interest income from lending, the market will eventually value it as a bank. Read the revenue mix, not the marketing.

Key Takeaways

• Asset-light fintechs (payments) trade on revenue and EBITDA multiples because their fee income is recurring, high-margin, and carries no credit risk.
• Balance-sheet fintechs (lenders) trade on price-to-book and ROE because their revenue is loss-prone and their equity, not their revenue, defines their real worth.
• A lender only trades above book value when its ROE exceeds its cost of equity. Below that, it trades at or under book.
• Blended models get sum-of-the-parts valuations: value the payments and lending arms separately, because mixing credit risk into a clean model can drag the whole multiple down.
• Always ask first: does this company put its own capital at risk? That single question tells you which valuation language to speak.