Calculating true CAC across paid, organic, and partner channels in fintech

A neobank's channel review shows paid social at $28 per signup and a partner referral deal at $6. Paid social gets cut. Two quarters later funded accounts are down and nobody can reconstruct why. Two things were missing from both numbers: the partner was billing a revenue sharerevenue shareThe percentage of total industry sales your company captures in a given period. It measures competitive position relative to rivals in a defined market.View full definition → that sat in a cost-of-revenue line rather than in the marketing budget, and neither figure carried the identity checks, credit pulls and manual review that stand between a signup and a funded account. The cost side of a fintech ledger is spread across four or five budget owners, and CACCACCustomer Acquisition Cost (CAC) is the total sales and marketing spend divided by the number of new customers gained in a period. It measures how efficiently you grow.View full definition → is only as honest as your ability to pull them into one place.

Why fintech CAC is different from e-commerce CAC

A DTC brand's acquisition cost more or less ends at the click. A fintech's starts there. Between the ad and a revenue-generating customer sits a chain of costs that marketing does not control and often cannot see:

• identity and document verification vendor fees, roughly $1 to $5 per check, higher for document plus liveness
• sanctions, PEP and watchlist screening, priced per name and re-run periodically
• credit bureau pulls, soft for prequalification and hard at underwriting, on any lending product
• compliance and fraud analyst time spent in manual review queues
• fraud losses on accounts that clear onboarding and turn out to be synthetic or stolen identities
• cash incentives: the funding bonus, the referral payout, the boosted intro rate

Two of those deserve a flag. Customer cash incentives are normally booked as a reduction of revenue rather than as marketing expense, so a team reading the marketing P&L will not see them at all. Fraud charge-offs land in credit losses. Both are acquisition costs in economic terms, and neither appears where marketers look.

Count only media and salaries and you get funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition → CAC. Count everything required to produce a funded, compliant customer and you get true CAC. The gap runs 30% to 60% in regulated lending and neobanking (directional; a16z's 16 fintech metrics is a reasonable frame, and a16z invests in the companies it writes about).

Building the blended, fully loaded CAC formula

Standard CAC:

CAC = Total Acquisition Spend / New Customers Acquired

True CAC for regulated fintech:

True CAC = (Marketing Spend + KYC/Verification Vendor Costs
            + Underwriting Costs + Allocated Compliance Labor
            + Partner Referral Fees + Customer Incentives)
            / Funded Customers (not signups)

The denominator is funded customers. The signup-to-funded losses themselves are the funnel lesson's territory; here they matter only as arithmetic. A lender with 100,000 app starts, 40,000 completed applications, 22,000 clearing verification and 9,000 funded has a spend base divided by 9,000, not by 100,000.

Worked example

One month at a US lending app:

• Paid marketing: $450,000
• Identity vendor fees: $2 per applicant × 40,000 completed applications = $80,000
• Underwriting and credit pull costs: $3 × 22,000 applicants reaching that stage = $66,000
• Compliance review labor (allocated): $40,000
• Partner referral fees: $30,000

Total fully loaded cost: $666,000. Funded customers: 9,000.

True CAC = $666,000 / 9,000 = $74

Naive CAC, marketing spend over signups: $450,000 / 100,000 = $4.50. A 16x understatement. The naive number is the one that reaches the board deck; the loaded one decides whether the business works.

Allocating shared compliance costs across channels

The hard part is not the total, it is splitting shared infrastructure (verification platform licences, fraud model maintenance, compliance headcount) across channels without inventing precision. Three workable methods:

1. Pro-rata by funded volume: each channel absorbs compliance cost in proportion to its share of funded customers. Simple, and the default.
2. Pro-rata by risk-flagged rate: channels producing more manual reviews or fraud flags absorb more. More accurate, more work, and it needs your review queue tagged by source at intake, which most are not.
3. Marginal cost only: assign the costs that scale per application, leave fixed compliance headcount as company overhead. Cleanest for comparing channels, and it understates company-wide unit economics.

For "should we cut paid social", method 3 is usually right. For unit economics reported to investors, method 1 or 2.

A simple allocation snippet

# Simplified true CAC by channel
channels = {
    "paid_social": {"spend": 180000, "signups": 40000, "funded": 3200},
    "partner_referral": {"spend": 30000, "signups": 15000, "funded": 2100},
    "organic": {"spend": 5000, "signups": 20000, "funded": 1800},
}

shared_kyc_cost = 80000
shared_underwriting_cost = 66000
shared_compliance_labor = 40000
total_shared = shared_kyc_cost + shared_underwriting_cost + shared_compliance_labor
total_funded = sum(c["funded"] for c in channels.values())

for name, c in channels.items():
    allocated_shared = total_shared * (c["funded"] / total_funded)
    true_cac = (c["spend"] + allocated_shared) / c["funded"]
    print(f"{name}: true CAC = ${true_cac:.2f}")

Run it with risk-flagged rates instead of funded volume and the ranking usually moves. That difference is the number worth arguing about.

The three channel types do not have comparable cost structures

Paid platforms. Google and Meta price by auction, and finance is consistently among the most expensive verticals on Google Ads, with high-intent loan and card terms trading far above the all-industry average click price. Cost per funded account therefore rises with your own scale: you bid against yourself as you push volume. Meta's channel-level numbers also carry a measurement caveat. After Apple's App Tracking Transparency rollout in 2021, Meta publicly put the 2022 revenue impact at roughly $10 billion, and app-install attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → degraded across the market. Reported per-channel CAC drifted away from blended CAC (all acquisition cost over all funded accounts), and blended became the control number that platform dashboards cannot inflate.

Organic. Not free, and its cost curve is inverted. Content, SEOSEOSearch Engine Optimization: the practice of improving your pages' natural (unpaid) rankings in search engine results pages to attract more organic traffic.View full definition →, app store optimisation and the salaries behind them are largely fixed, so marginal CAC on the next funded account is near zero while average CAC in year one can exceed paid. Load those salaries in, or organic will look miraculous and get starved anyway.

Marketplaces and revenue-share partners. MoneySuperMarket in the UK and Credit Karma in the US (Intuit-owned since 2020) sell exactly the placement under discussion: they are paid when a user clicks through, applies, or is approved. That structure makes cost per funded account low-variance and success-contingent, which finance teams like. It also selects hard for rate shoppers, who arrive with a comparison table open and leave when someone else prices better, so the cost per *retained* funded customer is materially worse than the cost per funded customer.

Revenue-share deals break the formula outright, because there is no fixed acquisition cost to divide. You can either capitalise the expected discounted stream of partner payments as CAC, which can double the number and makes the channel look expensive against paid, or treat the share as a permanent margin haircut and keep CAC low. Pick one and apply it everywhere. Mixing treatments across channels is how a partner deal gets renewed on a CAC of $6 while quietly consuming 20% of that cohort's revenue for its lifetime. Whether the resulting figure is healthy for your sub-category is the benchmarking lesson's question; what matters here is that every channel's number was assembled the same way before you ask it.

Common distortions to watch for

Channel mislabeling. Partner and organic get credited for customers moved by paid brand advertising. GeoGeoThe practice of making your brand and content visible and citable inside AI-generated answers from tools like ChatGPT, Gemini and Perplexity.View full definition → holdouts settle this better than any attribution model: dark a market for four weeks and read the change in total funded accounts, not in the dashboard. Teams that cut brand search often watch "organic" volume fall with it. See the Marketing Science Institute for methodology basics.

Denominator games. Under pressure, teams start reporting CAC per "activated" or "verified" customer instead of per funded one. Ask which denominator, every time.

Fraud cost hiding. Synthetic identity losses in the US are estimated in the billions of dollars a year, and they are booked as credit losses. If one source drives a disproportionate share of them, that cost belongs in its CAC.

Key Takeaways

• Fully loaded CAC includes verification and screening fees, underwriting, allocated compliance labour, partner fees and customer cash incentives, over *funded* customers. The gap to naive CAC is commonly 30% to 60%, enough to reverse a channel decision.
• Two large costs sit outside the marketing P&L by accounting convention: customer incentives (contra-revenue) and fraud charge-offs (credit losses). Go and get them.
• Use marginal-cost allocation for channel comparisons and pro-rata allocation for company-wide reporting, and say which one a given number uses.
• Paid CAC rises with your own scale in an auction; organic CAC is mostly fixed cost amortised over volume; marketplace CAC is success-contingent but selects for shoppers who churn. Revenue-share partners have no single CAC until you choose a treatment.
• Since ATT, blended CAC (total acquisition cost over total funded accounts) is the only figure platform-reported conversions cannot flatter. Pair it with the LTVLTVLifetime Value: the total revenue (or profit) a customer generates throughout their entire relationship with your business.View full definition → the value-side lesson builds, on the same funded-customer definition.

🎬 [VIDEO: "Fintech Unit Economics Explained" - youtube.com/results?search_query=fintech+unit+economics+CAC+LTV - search for current operator-led breakdowns of CAC, LTV, and payback period calculations in fintech business models]