Test Launch With a New iGaming Agency: How to Run a Pilot Without Disputes

A test launch is the first piece of work a brand and an agency do together. How it is set up decides not only the result, but whether either side wants to keep going. The brand is afraid of burning budget on traffic that never deposits. The agency is afraid of spending a month on the campaign and not getting paid because of criteria it only heard about after the fact.

Both fears have the same fix: agree on everything that matters before launch, and put it in writing. Here is what that covers.

The goal: what exactly are you testing

“Let's see how it goes” is not a goal. A test should check one or two hypotheses that you can confirm or reject by the end of the pilot. For example:

• can the agency deliver FTDs in our GEO at or below our target CPA;
• is player quality from this source comparable to our current channels;
• does the agency fit us on process: speed, transparency, how it handles feedback and change requests.

The last point is often forgotten, but it matters as much as the numbers. A pilot shows you how a partner behaves when things don't go to plan. You can get some of those answers before the test even starts, using our 12 questions to ask a Facebook agency.

The agency, in turn, should ask a question back: what happens after a successful test? If the brand can't name even a ballpark volume for the success scenario, the pilot risks becoming a one-off gig. Talking about the upside before launch is a normal part of the negotiation, not pressure.

One more rule: change only one thing per test. A new agency, a new GEO and a new landing page at the same time means three unknowns — and at the end nobody will know what worked and what let you down.

Budget, cap and timeline

Budget

The test budget has to cover as many FTDs as you need to draw conclusions. A test that is too small produces random numbers that neither the brand nor the agency can base a decision on.

Cap

A cap limits volume: how many FTDs the brand is willing to accept and pay for per day, per week or over the whole test. It protects the brand's budget and gives the agency a clear target at the same time: valid traffic within the cap gets paid. Agree in advance what happens if the agency sends more — is the overflow paid, and on what terms.

Timeline

Build three things into the timeline: setup (paperwork, platform authorization where required, creative approval), the campaign learning phase, and the quality evaluation window. If day-7 or day-30 redeposits matter to you, you can't make a call on quality until that window has closed for most of the players.

KPIs, baseline and success criteria

A baseline is your point of comparison. Most often it is your current numbers for comparable traffic in the same GEO: reg-to-FTD conversion, average first deposit, share of players who redeposit. If you don't have comparable data, agree on target ranges explicitly rather than going “by feel”.

Then lock down the definitions. What counts as an FTD — any deposit, or only one above a minimum amount? Whose data is the source of truth: the brand's back office, the tracker or the ad account? How are discrepancies resolved? Without this, at the end of the test both sides will be arguing about the numbers instead of discussing the result.

Define success in advance — and in writing

Before launch, describe three outcomes and the conditions for each: scale, iterate, stop. That removes the main source of conflict — criteria invented once the numbers are already known.

“Let's put it on record: if CPA lands within target and redeposits are no worse than our baseline, we raise the cap. If CPA is higher but quality is good, we run one more cycle with adjustments. If quality falls below the tolerance, we stop — no hard feelings on either side.”

Communication and reporting during the test

A pilot is an intense period, and silence during it is expensive. Agree on a few simple rules:

• One channel, clear owners. A shared chat and one decision-maker on each side.
• Cadence. A short daily status in the first days, then a weekly report on an agreed template. We cover which metrics to include in our piece on agency reporting.
• Feedback from the brand. The agency optimizes campaigns on deposit data. If the brand shares it once every two weeks, the agency is flying blind. Postbacks or a regular data export are the brand's responsibility during the test.
• Pause rules. Who can stop a campaign, and under what conditions: signs of fraud, a compliance issue, the cap being exceeded. How to agree on disputed traffic upfront is covered in our article on traffic quality and fraud.
• Mid-test changes. Switching the offer, bonus or landing page halfway through breaks the comparison. If it can't be avoided, log the date of the change and evaluate the periods separately.

The agency has its own part to play here: flag problems early — rejected creatives, conversion drops, suspicious patterns in the traffic. Bad news delivered straight away costs less than the same news in the final report.

The post-test decision: scale, iterate or stop

When the test ends, hold a dedicated review call instead of settling it in a chat thread. Reconcile the numbers against the agreed sources, compare them with the criteria and make one of three decisions.

Scale

The criteria are met — raise the cap or the budget gradually. Quality can shift as volume grows, so keep monitoring cohorts after the test too.

Iterate

Some of the criteria are met. Name exactly what changes — creatives, audiences, landing page, offer terms — and run another short cycle with the same logic: goal, timeline, criteria.

Stop

The criteria aren't met. If they were written down in advance, stopping is a calm process: the brand pays for valid volume as agreed, the agency hands over the final report, and both sides revoke access. A negative result is still a result — and it is cheaper than six months with the wrong partner.

Pre-launch checklist

1. One or two test hypotheses are written down.
2. The brand's license covers the test GEO; platform rules and local advertising requirements are accounted for, targeting is 18+ only, and for Meta the ad account is authorized.
3. Budget, cap and overflow terms are agreed.
4. The timeline allows for setup, campaign learning and the quality evaluation window.
5. Definitions are written down: registration, FTD, valid player.
6. Every metric has a source of truth and a reconciliation process.
7. The baseline or target ranges are fixed.
8. Three outcomes — scale, iterate, stop — are described, with the conditions for each.
9. Postbacks or a regular export of deposit data are set up.
10. Owners, a communication channel and a reporting cadence are assigned.
11. Pause rules and the process for reviewing disputed traffic are spelled out.
12. It is clear who owns the accounts, creatives and data after the test.

Pilot terms are easiest to agree during negotiations, before either side has expectations about the numbers. If the brand and the agency meet through Avero, agreeing this framework is part of the negotiations we run. To discuss your task, get in touch via our contacts page.