Revenue share vs CPA, worked out with real numbers: break-even retention, twelve-month cohort stacking, cashflow limits and when a one-time CPA still wins.
Two affiliates send the same hundred first-time depositors to the same brand in January. One takes a flat $100 CPA. The other takes 50% of net revenue for as long as those players stay active. By December, the CPA affiliate has banked $10,000 from that cohort and closed the file. The revenue share affiliate has collected around $11,400 and is still getting paid every month.
That single comparison is the entire revenue share vs CPA argument, and almost every version of it you read online is wrong in one of two directions. One camp says revenue share always wins because it compounds. The other says take the cash because programs die. Both are answering a question that only has a numerical answer, and the numbers change per offer, per geo and per traffic source.
What follows is the arithmetic. Break-even retention, the twelve-month cohort curve, the capital-velocity problem that quietly decides this for most media buyers, how hybrids actually get structured, and the risks nobody prices in. By the end you should be able to look at any two-option payout sheet and know within five minutes which one you want.
A CPA is a purchase of a completed event. You are selling one qualified action — usually a first-time deposit above a threshold, sometimes a verified registration — at a fixed price, and the transaction ends there. Your risk is concentrated in whether the action qualifies at all.
A revenue share is a purchase of a claim on future cashflow. You own a percentage of whatever the advertiser books from that user, for as long as the relationship holds and the tracking holds. Your risk is spread across retention, the advertiser's margin, and the durability of the attribution.
Those are different asset classes. Treating them as two prices on the same shelf is the root mistake.
Before any modelling, read what the payout is a percentage of. Net revenue in iGaming usually means gross gaming revenue minus bonus cost, minus payment processing, sometimes minus gaming duty and platform fees. Two brands quoting 50% can be twenty percent apart in real terms once deductions land.
Four clauses do most of the damage:
None of this is exotic. It is just rarely on the first page of the offer sheet.
Strip the problem down. Let R be the monthly net revenue an active user generates, s your revenue share rate, and r the monthly retention — the probability a user active this month is still active next month.
If retention is roughly constant, expected lifetime in months is 1 / (1 − r). Total revenue share collected per user is therefore:
Lifetime revenue share = s × R / (1 − r)
Set that equal to the CPA on offer and solve for retention:
Break-even retention r* = 1 − (s × R / CPA)
Work it with real numbers. A user generates $40 a month in net revenue. Your share is 50%, so $20 a month. The alternative CPA is $100.
r* = 1 − (20 / 100) = 0.80
You need 80% monthly retention for the revenue share to match the CPA on lifetime value. At exactly 80%, average lifetime is five months and you collect exactly $100. At 85% retention, lifetime is 6.7 months and you collect $133. At 70%, lifetime is 3.3 months and you collect $67 — a third less than the CPA.
That threshold is higher than most affiliates expect, and it is the reason so many revenue share deals underperform in practice. Tier-3 geos with heavy bonus-hunting traffic routinely sit below 70% monthly retention on depositors. Tier-1 crypto exchange traffic can sit well above 90%. Same formula, opposite conclusion.
The formula above treats a dollar in month twelve as equal to a dollar today. If you buy media, it is not. Money you get back this month buys traffic that earns again next month.
Add a monthly discount rate d representing what you can earn redeploying capital. Present value of the revenue share stream becomes:
PV = s × R / (1 + d − r)
At 85% retention with no discounting, the stream is worth $20 / 0.15 = $133. Apply a 10% monthly hurdle — modest for a buyer running 2x ROAS on a short cycle — and it becomes $20 / (1.10 − 0.85) = $80. The same deal that beat a $100 CPA by a third now loses to it by a fifth.
This is not a rounding adjustment. For anyone recycling capital fast, the discount rate is the dominant term. It is also why the "revenue share always wins" advice is written mostly by content affiliates, whose marginal traffic cost is close to zero and whose correct discount rate is therefore close to zero too.
The strongest argument for revenue share is that cohorts stack. January's players are still paying when February's arrive, and so on. That is true, and it is worth seeing at full size.
Assumptions: you acquire 100 first-time depositors every month, each active user generates $40 of monthly net revenue, your share is 50% ($20 per active user per month), monthly retention is 85%, and the alternative is a flat $100 CPA paid on the same 100 depositors.
| Month | Active users | Revenue share that month | Cumulative revshare | Cumulative CPA |
|---|---|---|---|---|
| 1 | 100 | $2,000 | $2,000 | $10,000 |
| 2 | 185 | $3,700 | $5,700 | $20,000 |
| 3 | 257 | $5,145 | $10,845 | $30,000 |
| 4 | 319 | $6,373 | $17,218 | $40,000 |
| 5 | 371 | $7,417 | $24,635 | $50,000 |
| 6 | 415 | $8,305 | $32,940 | $60,000 |
| 7 | 453 | $9,059 | $41,999 | $70,000 |
| 8 | 485 | $9,700 | $51,699 | $80,000 |
| 9 | 512 | $10,245 | $61,944 | $90,000 |
| 10 | 535 | $10,708 | $72,652 | $100,000 |
| 11 | 555 | $11,102 | $83,754 | $110,000 |
| 12 | 572 | $11,437 | $95,191 | $120,000 |
Two things jump out.
Month nine is the first month where the revenue share run rate ($10,245) exceeds the CPA run rate ($10,000). That is the month the curve visibly bends. It is also the month most affiliates would already have quit the deal, because for eight months straight the CPA line was ahead.
And at month twelve, cumulative revenue share is still $25,000 behind cumulative CPA. Every article that tells you compounding makes revenue share obviously superior stops the story before this row.
A single cohort crosses at month nine: 100 players earning a decaying $20 each accumulate $10,245 by then, just past the $10,000 the CPA paid on day one. But your total is always dragged down by immature cohorts. In month twelve, seven of your twelve cohorts are still below their own break-even.
The active-user count settles too. At 85% retention with 100 new depositors a month, the steady state is 100 / 0.15 ≈ 667 active users, generating about $13,333 a month against a $10,000 CPA run rate. The gap closes at roughly $3,300 a month, so cumulative revenue share overtakes cumulative CPA somewhere around month twenty-two.
Twenty-two months of consistent volume before the model you were told is obviously better actually pays you more in total. After that it keeps pulling away, and the terminal value is enormous — but you have to survive to get there.
Here is the practical version. Suppose you acquire depositors at $50 in media cost.
On CPA at $100, with net-15 payment terms, you spend $10,000 and see $20,000 back inside about three to four weeks. Your capital roughly doubles per cycle. You can compound that.
On 50% revenue share at $20 per active user per month, that same $10,000 buys 200 depositors. Month one returns 200 × $20 = $4,000. Month two, 170 actives return $3,400. Month three, about 145 actives return $2,890. You cross your $10,000 acquisition cost midway through month three.
Same offer, same traffic, same economics — and one version lets you redeploy capital roughly monthly while the other locks it up for a quarter. If your bankroll is the binding constraint, the CPA is not the lazy choice. It is the correct one, and it stays correct until your working capital is large enough that you can afford to have three months of spend outstanding at all times.
The corollary: the moment your revenue share base throws off enough monthly cash to fund your media spend, the constraint disappears and the calculus flips. Many affiliates never notice the flip because they never re-run the numbers after the first year.
Hybrids exist precisely because the two failure modes are complementary. CPA fails on terminal value, revenue share fails on capital velocity. A hybrid pays a reduced upfront fee plus a reduced ongoing share.
Take the same user: $40 monthly net revenue, 85% monthly retention.
| Structure | Upfront | Ongoing | Lifetime total | Cash back by month 3 |
|---|---|---|---|---|
| Pure CPA | $100 | $0 | $100 | $100 |
| Pure revshare 50% | $0 | $20/mo | $133 | $51 |
| Hybrid $50 + 30% | $50 | $12/mo | $130 | $81 |
The hybrid gives up $3 of lifetime value against the pure revenue share and returns 60% more cash inside the first quarter. That is a good trade for almost anyone buying traffic, and it is why hybrid should be your default ask on any offer you intend to run for more than a quarter. The same logic plays out slightly differently on exchanges, where the fee model changes the shape of the residual — the commission models used by crypto exchange programs are worth comparing side by side before you negotiate.
Two practical notes on negotiating hybrids. First, advertisers usually have more flexibility on the revenue share leg than on the CPA leg, because the CPA is a hard cash outlay against their own payback window. Second, hybrid CPAs are often gated on volume — you get the upfront component only above a monthly threshold of qualified deposits. Confirm the threshold before you plan around the cash.
The mistake is choosing once, as an identity, instead of per campaign. Run through this each time:
As a rough shape: organic and SEO traffic in high-retention verticals leans revenue share; cold paid social in tier-3 leans CPA; email, Telegram and community traffic with a long relationship usually justifies hybrid. If you are monetising a chat audience, the retention profile is typically strong enough to argue for the residual leg — the mechanics of monetising Telegram traffic reward exactly that.
A revenue share is a bet on institutional continuity. Price it like one.
Advertiser longevity. Brands lose licences, exit geos, get acquired, or simply stop paying. A twenty-four-month residual assumes twenty-four months of solvency and goodwill. Discount unproven brands harder than you feel comfortable doing.
Attribution resets. The most common way affiliates lose a residual base is not fraud, it is migration. A program moves platforms, re-issues tracking links, or changes postback endpoints, and historical player-to-affiliate mapping arrives incomplete. Keep your own records: player IDs where permitted, deposit timestamps, sub-ID mappings, and monthly statement exports. If you are relying on someone else's database to prove two years of earnings, you have already accepted the risk. Solid server-to-server postback tracking and disciplined sub-ID hygiene are what make a residual claim defensible.
Terms drift. Percentages get renegotiated downward when volume falls. Deduction lists grow. Read the amendment emails.
Concentration. If 70% of your residual sits with one advertiser, you do not have a residual business, you have a counterparty exposure. Three or four programs across two verticals is a meaningfully different risk profile.
Your own continuity. Dormancy clauses mean a residual can require you to stay active. Factor that into any plan that treats revenue share as retirement income.
You do not need a data team. You need four figures per campaign, tracked honestly:
With those, the break-even formula takes thirty seconds. Everything else in this article is elaboration on that one calculation. If you want the wider set of metrics that feed it, the core affiliate KPIs are the place to start.
One caution on month-two retention: measure it on a cohort, not on a blended active-user count. Blended retention flatters you whenever volume is growing, which is exactly when you are most likely to be making the decision.
Comparing gross to net. A 50% share of net revenue after bonus and processing costs is not half of what you see on the brand's marketing page. Get one real monthly statement before modelling.
Assuming a flat retention curve. Real retention is steepest in month one and flattens after. Using a single r is fine for a first pass and slightly pessimistic for long-lived cohorts, but if you are committing serious budget, model month one separately.
Ignoring the deal you could have negotiated. Most published payout tables are opening positions. Volume, exclusivity on a geo, and a track record on a similar offer all move the number — and the revenue share leg moves more easily than the CPA leg.
Switching mid-cohort without checking the terms. Moving from CPA to revenue share on an offer rarely applies retroactively. You are starting the residual base from zero on new players only, which changes the crossover math completely.
Treating one bad month as signal. Revenue share income is genuinely lumpy in high-variance verticals. Judge it on a rolling quarter.
We run a 50% revenue share as standard across crypto, forex, iGaming, gaming, high-ticket and tech, and we build and maintain the tracking that makes a residual claim provable rather than hopeful. Sites and landing pages are built for partners at no cost, which matters here specifically because it removes the fixed production cost from your break-even calculation. Because we hold direct relationships with the platforms and advertisers, hybrid structures and payout bumps are a conversation rather than a form — and you see the retention data for a geo before you commit to a model, not after. You can see how we work with partners in more detail on the main site.
If you want a second opinion on a deal in front of you, message the Telegram bot with the offer, the geo and your traffic source, and we will run the break-even with you. The community chat is open too if you would rather ask other partners first.
No. Revenue share beats a one-time CPA only when player retention is high enough that the accumulated share exceeds the flat fee. With a 50 percent share on 40 dollars of monthly net revenue against a 100 dollar CPA, you need roughly 80 percent monthly retention to break even. Below that, the CPA pays more and pays sooner.
Divide your expected monthly revenue share payment by the CPA on offer, then subtract that from one. If the share pays 20 dollars a month and the CPA is 100 dollars, break-even monthly retention is 1 minus 0.2, or 80 percent. Above that retention rate the revenue share wins on lifetime value, though not immediately on cash.
A hybrid pays a reduced upfront CPA plus a reduced ongoing revenue share on the same player. A common shape is roughly half the pure CPA plus 25 to 30 percent revenue share instead of 50. It returns acquisition cost quickly while still building a residual base, which suits anyone buying paid traffic.
Advertiser longevity, attribution resets during platform migrations, negative carryover clauses in iGaming, and quiet changes to what counts as lifetime. A revenue share is a claim on a future you do not control, so weight it by how likely the program is to still exist and still track your players in two years.
Server-to-server postbacks are the only conversion signal that survives modern browsers. Here is how the plumbing works, how to test it, and how to find the conversions you are quietly losing.
Read postExchanges still pay the best lifetime economics online, but the traffic that earns them changed. What converts now, what quietly died, and how to price a program before you send traffic.
Read postThe gap between a 2% and an 8% click-to-registration rate is almost never the casino. It is the funnel. Here is the full chain, the arithmetic behind each step, and how to find the leak.
Read post