Metric trees for fintech companies

Every industry claims its metrics are unique. Fintech has a stronger case than most. Three characteristics set fintech measurement apart from general SaaS or consumer businesses, and each one shapes how a metric tree should be structured.

First, fintech operates under regulatory scrutiny that directly affects product decisions. A payments company cannot optimise conversion without considering fraud rates and chargeback thresholds. A neobank cannot accelerate onboarding without meeting KYC verification requirements. A lending platform cannot grow loan volume without managing non-performing loan ratios within regulatory limits. Compliance is not a back-office concern in fintech. It is a product constraint that belongs in the metric tree alongside growth and engagement metrics.

Second, fintech unit economics are extreme. Customer acquisition costs in fintech average around £1,450, roughly twenty times higher than e-commerce and double that of B2B SaaS. This means that every customer who churns represents a significant sunk cost, and the path to profitability depends on retaining customers long enough to recoup that investment. The metric tree must make the relationship between acquisition cost, retention, and lifetime value explicit and traceable.

Third, fintech revenue models are often transaction-based rather than subscription-based. A payments processor earns a take rate on every transaction. A neobank earns interchange fees on card usage plus interest on deposits. A lending platform earns net interest margin on its loan book. These models mean that engagement frequency and transaction volume are direct revenue drivers, not just proxy metrics. The metric tree must reflect this tight coupling between usage and revenue.

Payments businesses live and die by Total Payment Volume (TPV) and take rate. TPV measures the total monetary value of transactions flowing through the platform. Take rate is the percentage of that volume the company retains as revenue. Together, they produce net revenue: the figure that actually matters for the business.

The metric tree for a payments company decomposes net revenue into these two branches, then breaks each further into the operational levers that drive them. TPV is a function of the number of active merchants, the number of transactions per merchant, and the average transaction value. Take rate is influenced by merchant mix (enterprise merchants negotiate lower rates), payment method mix (card-not-present transactions carry higher interchange), and geographic mix (cross-border transactions command premium pricing).

Below TPV, the tree splits into acquisition and retention branches. New merchant onboarding drives volume growth, but merchant churn can erode it just as quickly. For a payments company like Stripe or Adyen, a merchant that processes millions in volume churning to a competitor is a catastrophic event that no amount of new merchant acquisition can easily replace. The tree makes this asymmetry visible.

What makes the payments tree distinctive is the tension between volume growth and take rate compression. As a payments company scales and attracts larger merchants, its blended take rate typically falls because enterprise merchants demand lower pricing. The tree must track both dimensions simultaneously. Growing TPV by 40% while take rate compresses by 30% is not growth. It is margin erosion disguised as progress. The metric tree prevents this self-deception by keeping both branches visible at all times.

Payments companies must also track operational health metrics that sit alongside the revenue tree: authorisation rates (the percentage of attempted transactions that succeed), settlement times, and dispute rates. A declining authorisation rate directly reduces TPV, and a rising dispute rate can trigger card network penalties. These are not secondary metrics. They are structural constraints on the revenue tree.

Neobanks face a uniquely challenging path to profitability. Despite rapid growth (the sector is expanding at roughly 35% CAGR in North America), approximately 76% of neobanks remain unprofitable. The core problem is structural: neobanks acquire customers at high cost, offer free or low-cost accounts to drive adoption, and then must find ways to monetise those customers through card usage, premium subscriptions, lending products, or interest income on deposits.

The metric tree for a neobank must capture this entire journey from acquisition through monetisation. The North Star is typically revenue per customer or, for more mature neobanks, contribution margin per customer, because top-line user growth means nothing if each user costs more to serve than they generate in revenue.

The revenue branch splits into three streams that reflect how neobanks actually make money. Interchange revenue comes from card transactions: every time a customer taps their card, the neobank earns a small percentage. This decomposes into transaction frequency and average value, both of which are direct measures of engagement. Subscription revenue comes from premium tiers that offer features like higher interest rates, travel insurance, or fee-free international transfers. Interest income comes from lending out customer deposits, governed by the average deposit balance and the net interest margin the neobank achieves.

The cost branch is equally important. Neobanks that focus only on revenue per customer without tracking cost to serve will never find profitability. Support costs scale with customer problems, which often correlate with product complexity. Infrastructure costs should scale sub-linearly if the technology platform is well-architected. Compliance costs, including KYC verification, transaction monitoring, and regulatory reporting, are a uniquely heavy burden for neobanks and often represent a larger share of cost to serve than traditional banks experience, because neobanks handle the same regulatory requirements with smaller teams.

The separation between these revenue streams matters because they have different growth levers. Increasing interchange revenue requires driving card usage, which is a product and engagement challenge. Growing subscription revenue requires demonstrating premium value, which is a positioning and feature challenge. Expanding interest income requires attracting deposits, which is a trust and rate competitiveness challenge. A single "grow revenue" goal is meaningless without this decomposition.

Lending platforms, whether consumer lenders, buy-now-pay-later providers, or SME lending platforms, have the most financially complex metric trees in fintech. Their revenue is generated by the spread between what they earn on loans and what they pay for capital, but their risk is concentrated in credit quality. A lending metric tree must hold both dimensions together.

The core tension in lending is between volume growth and credit quality. Originating more loans increases revenue, but loosening credit standards to grow volume increases the non-performing loan (NPL) ratio, which can destroy profitability. The metric tree makes this trade-off explicit by placing loan origination volume and credit quality metrics on parallel branches under the same root.

The unique challenge for lending metric trees is that the same action can move metrics in opposite directions. Tightening credit criteria improves NPL ratio but reduces loan approval rate and origination volume. Offering lower interest rates attracts more borrowers but compresses NIM. Extending loan durations increases total interest income per loan but increases credit risk exposure.

A well-structured metric tree makes these trade-offs visible rather than hiding them in separate reports owned by separate teams. When the growth team sees that their origination targets sit alongside the risk team's NPL thresholds in the same tree, the conversation shifts from adversarial to collaborative. Both teams can see the constraints they operate within and find the strategies that improve one metric without destroying the other.

Most metric trees in non-regulated industries treat compliance as an external concern, something handled by the legal team outside the performance framework. In fintech, this approach is dangerous. Compliance metrics are performance metrics. A failed KYC process does not just create regulatory risk; it directly reduces conversion. A slow AML screening process does not just frustrate compliance officers; it delays merchant onboarding and reduces time-to-revenue. A rising fraud rate does not just trigger fines; it erodes customer trust and increases churn.

The most effective fintech metric trees integrate compliance metrics directly into the operational branches where they have impact, rather than isolating them in a separate compliance section that nobody outside the risk team monitors.

1. 1

   KYC completion rate

   The percentage of users who successfully complete identity verification. This metric sits on the onboarding branch of the tree, directly between sign-up and activation. A low KYC completion rate is simultaneously a compliance gap and a conversion bottleneck. Improving it requires collaboration between compliance (ensuring the process meets regulatory standards) and product (ensuring the user experience minimises drop-off).

2. 2

   Transaction monitoring false positive rate

   AML transaction monitoring systems flag suspicious activity for manual review. A high false positive rate overwhelms the compliance team, delays legitimate transactions, and increases operational cost. This metric belongs on the cost-to-serve branch and the customer experience branch, because every false positive that freezes a legitimate customer's account is a churn risk.

3. 3

   Suspicious Activity Report (SAR) filing rate

   The number of SARs filed relative to transaction volume. This is a regulatory requirement, but it also serves as a health indicator. A rate that is too low may signal inadequate monitoring. A rate that is dramatically higher than industry peers may signal an overly aggressive detection model that generates unnecessary work.

4. 4

   Fraud loss rate

   Fraud losses as a percentage of total transaction volume. This metric directly reduces net revenue and belongs on the revenue tree as a negative branch. It also connects to the trust branch: visible fraud events (such as unauthorised transactions on customer accounts) are one of the fastest drivers of churn in consumer fintech.

5. 5

   Time to regulatory approval

   For fintech companies entering new markets or launching new products, the time required to obtain regulatory licences or approvals is a critical operational metric. It determines speed to market and belongs on the growth branch alongside product development timelines.

The fintech industry is moving from a growth-at-all-costs era to one where unit economics determine which companies survive. With 76% of neobanks still unprofitable and investors demanding clearer paths to sustainable margins, the metric tree must make the economics of each customer visible and traceable.

Unit economics in fintech revolve around a simple question: does each customer generate more value than they cost to acquire and serve? The answer lives in the relationship between three metrics: Customer Acquisition Cost (CAC), Lifetime Value (LTV), and the payback period that connects them.

The fintech LTV calculation differs from standard SaaS in two important ways. First, ARPU is often a composite of multiple revenue streams (transaction fees, subscriptions, and interest income) rather than a single subscription price. This means that increasing LTV requires understanding which revenue stream has the most room to grow for each customer segment. Second, fintech CAC includes regulatory onboarding costs (KYC verification, credit checks, compliance screening) that do not exist in most other industries. These costs are non-negotiable and must be factored into the payback calculation.

A healthy fintech LTV:CAC ratio is typically 3:1 or above, with a CAC payback period under 12 months. But these benchmarks vary significantly by sub-sector. Lending platforms can sustain higher CAC because their LTV per customer is higher (each loan generates substantial interest income). Payments companies need lower CAC because per-customer revenue depends on transaction volume, which varies enormously. Neobanks often have the hardest path because they acquire customers with free products and must migrate them to revenue-generating behaviours over time.

The metric tree makes these dynamics actionable. When the LTV:CAC ratio falls below target, the tree shows whether the problem is on the LTV side (low ARPU, high churn, poor margins) or the CAC side (expensive channels, low conversion, high onboarding costs). This specificity is what turns a concerning ratio into a solvable problem.

Building a metric tree for a fintech company follows the same principles as any metric tree, but with specific considerations that reflect the industry's unique constraints. Here is how to approach it.

1. 1

   Start with your North Star, not your reporting requirements

   Regulators require you to report dozens of metrics. Investors ask for another set. Internal teams track their own dashboards. The metric tree is not a dumping ground for all of these. Start with the single metric that best captures the value your business creates. For a payments company, this might be net revenue. For a neobank, contribution margin per customer. For a lending platform, risk-adjusted net interest income. Everything else in the tree should decompose from or connect to this root.

2. 2

   Decompose revenue by how you actually earn it

   Fintech revenue models are diverse. Use the decomposition that matches your business. Payments: TPV multiplied by take rate. Neobanks: interchange plus subscriptions plus interest income. Lending: loan volume multiplied by NIM minus provisions. The structure of your revenue tree should mirror the actual mechanics of how money flows through your business.

3. 3

   Place compliance metrics where they constrain growth

   Do not create a separate compliance section. Instead, place KYC completion rate on the onboarding branch, fraud loss rate on the revenue branch as a negative input, transaction monitoring costs on the cost-to-serve branch, and regulatory approval timelines on the market expansion branch. This ensures compliance is treated as an operational reality, not an afterthought.

4. 4

   Make unit economics visible at every level

   Every branch of the tree should connect to a unit economics view. What does it cost to acquire a merchant? What revenue does each merchant generate? What is the cost to serve per transaction? When unit economics are embedded throughout the tree rather than calculated separately, every team can see how their work affects the path to profitability.

5. 5

   Assign ownership that reflects your organisational reality

   Fintech organisations often have shared ownership challenges. The fraud rate is influenced by product (fraud detection features), engineering (model accuracy), operations (manual review), and compliance (policy thresholds). The metric tree should have a single owner for each node, even when multiple teams contribute. This prevents the diffusion of responsibility that allows metrics to deteriorate without anyone noticing.