A finance director looks at the acquisition report, sees the CPA figure, and asks the question every affiliate manager dreads: why does it cost this much to bring in one customer? Nine times out of ten, nobody in the room can answer with confidence, because the CPA was never broken down properly in the first place.

Cost per acquisition (CPA) is the single number that determines whether a fintech's affiliate programme is profitable or quietly draining the marketing budget. Get the formula wrong, benchmark against the wrong vertical, or ignore what's actually driving the number up, and you'll either overreact to a healthy CPA or underreact to a broken one. This article covers how to calculate CPA properly, what a reasonable range looks like across different financial products in Europe, and the practical fixes that actually move the number.

What Is CPA in Affiliate Marketing?

CPA is the total cost of acquiring one paying or fully qualified customer through a given channel or campaign, calculated by dividing total spend by the number of successful conversions.

That's the textbook version. In practice, CPA only means something once you've defined "conversion" precisely. For a digital bank, that might be a funded account. For a lending platform, it might be a signed loan agreement rather than a submitted application. For an investment platform, it could be a first deposit above a minimum threshold. Vague conversion definitions are the most common reason CPA figures get argued over in board meetings, because two people can be looking at the same campaign and calling it successful or unsuccessful depending on which event they're counting.

The CPA Formula, and Where Marketers Get It Wrong

The basic formula is straightforward:

CPA = Total Acquisition Spend ÷ Number of Conversions

The mistake happens in what gets included in "total spend." Affiliate commissions are only part of the cost. A properly calculated CPA should also account for platform or network fees, internal team time spent on publisher management, creative production, compliance review, and any bonus incentives paid to top-performing partners. Leave these out and you get a figure that looks better than reality, which is fine for a slide deck and terrible for a budget decision.

There's also a difference between blended CPA and channel-specific CPA. Blended CPA averages everything together across paid search, affiliates, social, and referral. It tells you very little about whether your affiliate programme specifically is efficient. If a lending brand's blended CPA looks acceptable but the affiliate channel alone is carrying a far higher number, that gets hidden until someone pulls the channel apart. Always calculate CPA per channel, and ideally per publisher tier, before drawing conclusions.

One thing worth flagging: don't confuse CPA with the commission you pay per action. A publisher might be paid under a CPA model (cost per action, where the "action" is a defined conversion event such as an approved application), but your internal CPA figure is broader than that single payout, because it includes everything else spent to generate the volume that publisher converts.

For a fuller walkthrough of the standard formula and how it applies across the wider affiliate marketing ecosystem, this guide on cost per acquisition in affiliate marketing covers the calculation in more depth.

What Counts as a High CPA? European Benchmarks by Vertical

There's no single "good" CPA across fintech. The acceptable range depends heavily on lifetime value, regulatory complexity, and how long the sales cycle is. A few patterns hold reasonably consistently across European markets:

Digital banking and payments. Customer acquisition tends to be cheaper here because the product is simpler to explain and the conversion event (account opening) requires less friction. Programmes in this space usually aim to keep CPA well below the customer's first-year revenue contribution, since margins per account are thinner than in lending or investing.

Consumer and P2P lending. CPA runs meaningfully higher because the conversion event, an approved and funded loan, sits further down the funnel and carries approval risk. A publisher can drive plenty of applications that never convert to funded loans, which is exactly why lending brands increasingly favour CPL or hybrid models over a flat CPA per application.

Investment and trading platforms. These programmes often accept a higher CPA because customer lifetime value, particularly for platforms with recurring trading activity or AUM-based fees, can be substantial. The trade-off is that investment products fall under MiFID II, so promotions must be fair, clear, and not misleading, and publisher content needs closer compliance review, which adds cost that doesn't always show up in the headline CPA.

Insurance (InsurTech). CPA sits somewhere in the middle, and tends to vary a great deal by product line. Comparison-driven products with short decision cycles acquire more cheaply than complex cover types that need advisory content.

The practical takeaway: don't benchmark your lending CPA against a payments brand's CPA and conclude something's broken. Benchmark against your own vertical, your own average customer value, and your own historical trend.

Why Your CPA Might Be Inflated

Before rebuilding a programme from scratch, it's worth diagnosing why the number is high in the first place. A few causes come up again and again in fintech affiliate audits.

Attribution gaps. If tracking is misconfigured, or if a portion of conversions are being attributed to the wrong channel, brands often overpay affiliates for traffic that would have converted anyway, while simultaneously miscalculating what the true channel CPA is. This is one of the most underrated causes of an apparently high CPA, because the spend is real but the credit is misallocated.

A publisher mix skewed toward volume over quality. Publishers optimising purely for click volume or lead volume, rather than qualified conversions, will inflate spend without moving the needle on funded accounts or approved loans. This is especially common with content or coupon sites that drive traffic but don't pre-qualify the visitor in any way.

Weak landing page or onboarding experience. Affiliate traffic can be perfectly well targeted and still convert poorly if the landing page, KYC flow, or application form creates unnecessary drop-off. In this case the "high CPA" problem isn't really an affiliate problem at all, it's a conversion rate problem that happens to show up in the affiliate report.

Flat-fee CPA structures on high-consideration products. Paying a flat fee per lead or application on a product like a loan or an investment account, without any tie to actual funded value, removes any incentive for the publisher to bring in higher-intent traffic. This is a structural issue, not a tactical one, and it usually needs a change to the commission model itself.

Compliance overhead treated as a hidden cost. Every piece of promotional content for regulated products needs review against the Unfair Commercial Practices Directive, which treats undisclosed affiliate relationships as misleading, plus product-specific rules such as MiFID II for investment marketing or the EU Consumer Credit Directive for lending. That review time is a real cost of acquisition that rarely gets added into the CPA calculation, which is one reason the "true" CPA on regulated products is often higher than the reported one.

How to Lower CPA Without Sacrificing Lead Quality

Cutting CPA the wrong way is easy: reduce commissions, cut underperforming publishers, and watch conversion quality collapse along with volume. A more durable approach focuses on structure rather than just spend.

Move from flat CPA to CPL or hybrid models where the product justifies it. For lending, insurance, and brokerage, a CPL model, paying per qualified lead rather than per completed application, shifts some of the conversion risk back onto the acquisition process rather than resting entirely on the publisher's traffic quality. For high value products such as P2P lending, investment platforms, and brokers, a hybrid model works well: a CPL paid upfront, plus a CPS earned on the lead's transaction volume during the first 90 to 180 days after registration, usually alongside a fixed fee for content production. This aligns publisher incentives with actual customer value rather than raw lead count.

Tier your publishers by conversion quality, not just volume. Ranking affiliates by approved-to-applied ratio, or funded-to-registered ratio, and directing budget toward the top tier, usually does more for CPA than any commission renegotiation. It's a slower fix than cutting rates, but it holds up over time.

Fix the conversion funnel before touching the affiliate programme. If drop-off is happening at KYC or document upload stages, no amount of publisher optimisation will fix the underlying CPA. It's worth running this diagnostic before assuming the problem sits with acquisition.

Test creative and messaging by publisher segment. A comparison site, a content publisher, and a cashback platform all need different messaging to convert at their best rate. Treating all affiliate traffic the same, with identical creative and identical landing pages, tends to underperform against a segmented approach.

Recheck attribution windows and deduplication rules regularly. Attribution logic that was set up two years ago rarely still matches how the customer journey actually looks today, particularly with the shift toward multi-device research before application.

When a High CPA Is Actually Fine

Not every high CPA needs fixing. If customer lifetime value is strong, retention is good, and the CPA sits comfortably below the payback period the finance team is comfortable with, a higher number than a competitor's isn't automatically a problem. The mistake is comparing CPA in isolation rather than against LTV, payback period, and retention. A brand with a higher CPA but stronger retention can easily out-earn a competitor with a lower CPA and high churn.

How Circlewise Approaches CPA Optimisation

Circlewise works with fintech and financial services brands across Europe to structure affiliate programmes around the metrics that actually matter, not just the ones that are easiest to report. That usually means auditing the true CPA including hidden costs, restructuring commission models toward CPL or hybrid where the product calls for it, and recruiting and tiering publishers based on conversion quality rather than raw traffic. For brands running lending, investment, or insurance affiliate programmes, getting this structure right early tends to matter more than any single rate negotiation later on.

Frequently Asked Questions

What is a good CPA for a fintech affiliate programme? There's no universal figure. A CPA is "good" when it sits comfortably below the customer's expected lifetime value within an acceptable payback period, which varies significantly between banking, lending, investment, and insurance products.

How is CPA different from CPL? CPA pays for a completed action such as a funded account or approved loan, while CPL pays for a qualified lead earlier in the funnel. CPL shifts less conversion risk onto the publisher, which is why it's common in lending, insurance, and brokerage.

Why is my affiliate CPA higher than my paid search CPA? This is common in regulated products where affiliate content needs compliance review and where publisher traffic quality varies more than a controlled paid search campaign. It doesn't necessarily mean the affiliate channel is underperforming once lifetime value is factored in.

Should I cut underperforming publishers to lower CPA? Sometimes, but only after checking whether the underperformance is a traffic quality issue or a funnel issue. Cutting a publisher whose traffic converts poorly because of a weak landing page fixes the wrong problem.

Does compliance review affect CPA? Yes. Reviewing affiliate content against rules such as the Unfair Commercial Practices Directive, MiFID II, or the EU Consumer Credit Directive adds real cost that's often left out of the reported CPA figure, particularly for investment and lending products.

What's the difference between blended CPA and channel CPA? Blended CPA averages spend and conversions across every acquisition channel, which can mask an inefficient affiliate channel. Channel-specific CPA isolates affiliate performance so it can be judged and optimised on its own.

Can a hybrid commission model actually lower CPA? It can, because it ties part of the payout to the lead's actual transaction volume in the months after registration rather than paying a flat amount regardless of quality. This tends to shift publisher behaviour toward higher-intent traffic over time.

Final Thoughts

A high CPA is a symptom, not a diagnosis. Before assuming the affiliate channel itself is the problem, check the formula includes every real cost, benchmark against your own vertical rather than a different one, and rule out attribution errors and funnel drop-off as the actual culprits. Once the number is genuinely understood, the fix is usually structural: better publisher tiering, a commission model that matches the product, and a conversion funnel that doesn't waste the traffic already being paid for. Get those three right, and CPA tends to sort itself out without a single rate cut.