That riddle is hypothetical. It is not our invoice, and it is not a source we are willing to name. It is where most acquisition reports still stop: volume, registrations, first-time deposits, cost per acquisition, then a colour.
Casino Capybara and Bukmacher na Topie are media properties. They earn through affiliate partnerships. They do not carry an operator's player ledger, and this article will not pretend they do. What operating them makes obvious is the other side of "buy iGaming traffic": the seller's report usually ends at the click, the registration, or the referred player. The buyer's economics start after that. The sequence below is the reading we would want on the buying side before anyone called a source profitable. Where a number would have to be invented, it is left out.
In short
- FTD CPA is the price of a first deposit. Two iGaming traffic providers at the same CPA can be different businesses by the second deposit.
- Second-deposit rate is the earliest quality signal a welcome offer does not fully explain. The player came back with money.
- Depositing retention at day 7 and day 30 separates a conversion from a customer.
- A blended affiliate average hides a broken campaign. Split by cohort, sub-ID, market, device and creative before you scale or cut.
- Fraud often arrives looking like a good week: very fast registration-to-FTD, repeated amounts, no play, no second deposit.
- CPA, revenue share and hybrid move the risk. None of them repairs a source you have not followed past the first deposit.
Read the whole funnel, not the first green cell
A useful acquisition funnel is click, registration, first deposit, second deposit, day 7, day 30, net gaming revenue. Each stage answers a different question. A green cell at the top does not answer the ones below it, which is how a weekly report stays cheerful while the book does not.
Click to registration tells you whether the traffic and the landing experience agree. A poor rate can be low-intent visits, a misleading ad, the wrong market, a form that fails on a phone, or a payments step the buyer never reached. It is not automatically the provider's fault. We have shipped pages, on our own properties, where the page was the problem and the traffic was fine. Registration to first deposit tells you how many of those accounts put money down. A source can look busy above that line and empty on it. Sometimes that is low intent. Sometimes the welcome offer did all the work the report is applauding.
The failure mode is stopping at whichever number is easiest to put in the slide.
Why FTD CPA lies about the provider
CPA is a real number. It is what you paid for a defined action, usually a first deposit that cleared a minimum. It says nothing about the player after that action.
Take the riddle as a shape, not as our figures. One source deposits once, takes the welcome offer, and never returns. The other deposits again, plays, and is still there a month later. The invoice can show the cheaper CPA on the first source. The business can still be the second. We will not invent a ratio for how often that split appears. Anyone who has reconciled an affiliate bill against what was left after bonuses has seen the shape, and a ratio without the book underneath it would be fiction.
This is why "buy iGaming traffic" is an incomplete brief. Clicks can be bought this afternoon. A customer cannot, until the cohort has been watched. If a provider's pitch ends at first-deposit volume, the pitch ended before the economics started. Buying a link for a ranking signal has the same rented shape on the search side; that decision is covered in iGaming niche edits. Neither rental becomes an asset because the dashboard was green on day one.
Second deposit, then day 7 and day 30
The first deposit is heavily shaped by the offer used to buy it. The second deposit is the player returning with money when the welcome banner is no longer the whole story. When sources are compared, second-deposit rate is the earliest split that should change a decision.
If one affiliate produces first-time depositors and almost nobody deposits again, the explanations worth opening are specific. Bonus-driven traffic. A market or device segment with no intent. Creative that promised a different product. Duplicate accounts. A payments or onboarding failure on the operator's side. The traffic is not always the culprit. The funnel still has to be opened before the source is blamed or scaled.
Then look at depositing retention, not login retention. Of the players acquired in that cohort, how many made another real-money deposit within 7 days, and how many were still depositing at 30 days? Logins flatter a source. Deposits are the commercial fact. There is no percentage here to aim at, because publishing one would mean inventing a benchmark. The comparison that matters is between sources inside the same market, the same product, and the same week.
NGR is where a traffic source becomes a business
Eventually the source has to be read in money kept, not money that passed through. Deposits are not that number. The page we would want in front of a buyer is short:
| What you measure | What it is actually telling you |
|---|---|
| FTD CPA | The price of a first deposit |
| Second-deposit rate | Whether the player came back |
| D7 and D30 depositing | Whether they are still a customer |
| NGR after bonuses and commission | Whether the source leaves any money |
| Chargebacks and duplicate instruments | Whether the growth is real |
A source with fat deposits and a bonus bill that consumes them is a weak source. A source with impressive gross gaming revenue and a chargeback problem is a weak source. A source with a dull CPA and sustainable net gaming revenue is the one to keep, and it will lose a report that only ranks on first deposits. Gross deposits, withdrawals, bonuses, payment costs, tax where it applies, and the affiliate commission all sit between the CPA and what is left. Skip any of them and the CPA is doing work it cannot do.
We are not quoting an NGR figure from a client or from our own properties. Casino Capybara and Bukmacher na Topie do not have that ledger. The point of the metric is that the buyer does, and that FTD CPA is upstream of it.
Cohorts, not the blended average
Blended affiliate lifetime numbers hide a change in the traffic. A partner can have two acceptable years and then switch on a campaign that floods the book with registrations that never deposit twice. The lifetime average still looks fine. The current cohort is the business you are about to scale.
Break it by affiliate, campaign, sub-ID, market, registration week, device, product, landing page, and creative. The question is not whether the partner is "good". The question is which slice of that partner, this month, is still a customer. That is a cut you can actually make. A single blended CPA for "iGaming traffic" is how the bad slice gets the same increase as the good one.
Fraud that presents as growth
Bad traffic does not always look bad on day one. Sometimes it is the best day on the dashboard, which is why a spike in volume is a reason to look, not a reason to raise the cap. The patterns that should stop the celebration are ordinary once you join the data: registration-to-FTD wildly faster than the rest of the book, repeated deposit amounts, identical devices, tight IP ranges, bonus claiming with almost no play, duplicate payment instruments, no second deposit, and a spike that arrived the day a new sub-ID went live.
A source can hit its acquisition target and still be worth less than nothing after the bonus and the chargeback. Affiliate reports, fraud checks, payments, and CRM have to be readable against the same player. If they live in exports nobody joins, the pattern gets to look like growth for a whole invoice cycle.
This is a list of checks, not a scorecard of traffic platforms. We do not rank providers, and we are not going to name one we "caught". Naming one without a book we can show would be a story, not evidence.
CPA, revenue share, and hybrid
The commercial model decides who holds the risk after the click. It does not measure quality by itself.
CPA
You pay on the acquisition event. The affiliate's risk mostly ends when the player qualifies. Everything after that first deposit is yours to notice, because nobody else is sharing the downside. Post-FTD monitoring is the whole job under this model, not a refinement.
Revenue share
The affiliate stays exposed to the player over time. That lines the incentives up, until the players are valuable, at which point the share can be the most expensive contract on the sheet. Alignment and cheapness are different words.
Hybrid
Some cash at acquisition, some share of later value. It works when both halves are visible. It hides a bad source if the report only watches the CPA half and treats the share as a footnote.
There is no correct model in the abstract. Market, margin, and how well you can see past day one decide it. A model you cannot reconcile to net gaming revenue is a model you are hoping about.
When to cut a source
A red CPA is not a verdict, and the lowest CPA on the sheet is not a scale decision. The sequence is short on purpose.
- Name the weak slice. Affiliate, campaign, market, device, or creative. "The partner" is too large to act on.
- Name where it breaks. Registration, first deposit, second deposit, or what is left after bonuses and commission.
- Rule out your own side. Payments, onboarding, product, CRM, a landing page that started failing the same week.
- Check fraud and bonus behaviour before the label becomes "low quality traffic".
- Compare with a peer cohort from the same week, not with the partner's lifetime average.
Then scale, hold, renegotiate, restrict, or cut. The decision sits on what the players did after they paid, not on how many of them paid once.
If you want that read done against a book you already have, it sits with iGaming consulting: acquisition economics and commercial terms, written down. It is a method for the traffic you are already buying. It is not a promise of more of it.