Key takeaways
The click is bought once. Everything after it is lost for free.
Most of the drop happens before the form, not inside it.
KYC is not the problem; an unexplained wait is.
Fix the step with the most traffic behind it first, not the one with the worst rate.
The shape of the loss
Take a thousand visitors who have already arrived — the click is paid for, the landing page has loaded, the acquisition budget is spent.
A typical flow, before any conversion work, looks roughly like this:
| Step | Of 1,000 |
|---|---|
| Landed | 1,000 |
| Reached the form | 310 |
| Started signing up | 140 |
| Completed verification | 62 |
| First deposit | 38 |
Thirty-eight. The other 962 cost exactly the same to acquire.
Two things in that table are worth sitting with. The first is where the biggest absolute loss happens: not in the form, but before it. Nearly seven in ten visitors never reach a field at all. The second is that every subsequent step compounds — a ten per cent improvement at the top carries all the way down, while the same improvement at the bottom moves a much smaller number.
This is why we treat the page, the form and the deposit step as part of the acquisition channel rather than as somebody else’s backlog. The media budget and the funnel are the same purchase. One of them just has an invoice attached.
The three biggest leaks
Field order
Forms are usually built in the order the database wants the data, which is almost never the order a person is willing to give it.
The cost is front-loaded: an early field that feels invasive — full legal name, date of birth, address — ends the session before any commitment has been made. The same field asked after a person has already entered two easy ones is answered at a much higher rate, because they have started and would now be abandoning something.
Splitting registration across steps is the single most reliable change we make. It does not reduce the amount asked. It changes when it is asked, and the measured effect is consistently in the double digits.
How verification is framed
KYC is not optional and it is not the problem. The problem is that most flows present it as a wall, with no explanation of what is happening or how long it takes.
A player who is told “we verify your details to meet licensing requirements — this usually takes two minutes” behaves differently from one who sees a spinner. The requirement has not changed. The uncertainty has. Unexplained waiting is where sessions die, and it is cheap to fix because it is copy and state handling rather than a rebuild.
Payment priority by market
The deposit step is where a flow that has done everything else right still loses people, usually for one mundane reason: the method they use is third in a list of nine.
Payment preference is strongly regional and changes faster than most operators update their ordering. Putting the two methods that actually dominate a market at the top — and testing that order per market rather than globally — is a change of a few hours that moves the last and most valuable step in the funnel.
What to test first
Not in order of how bad each step looks. In order of expected return: the conversion rate multiplied by the traffic behind it.
Working from the flows we see most often, that order is usually:
- Split registration across steps. Largest and most consistent effect, and it applies to every market at once.
- Reorder payment methods by market. Small effort, lands on the step closest to revenue.
- Explain the verification wait in place. Copy and state handling, no rebuild, and it recovers sessions already deep in the funnel.
- Show bonus terms before the field, not after. Fewer registrations, more deposits — and a measurably better player afterwards, because nobody arrived on a misunderstanding.
- One offer per campaign page instead of a grid. Matters most where paid traffic lands on a page built for browsing.
- A fixed action on mobile instead of a button in the footer. Smallest of the six, and it costs an afternoon.
Two rules make the list worth running at all.
Ship with measurement attached. If a change cannot be measured through to first deposit, it does not go live. Attribution that stops at the click tells you which creative was popular, not which change made money.
One change at a time, long enough to be sure. Betting traffic is seasonal and lumpy. A test read across a weekend with a big fixture in it is not a test, it is a fixture.
The full list, with the typical range each change moves, is on the webdesign, landing pages & CRO page.
