The structural problem
Picking a media-buying partner is a rare, high-stakes, low-information decision for most operators. The CMO does it once every two years. The data points come from sales decks, references the partner curated, and gut feel from a single discovery call. Failures are expensive — six months of stalled acquisition plus the opportunity cost of the cohorts not acquired during the rebuild.
The fix is to treat partner selection like any other procurement decision involving operational risk: explicit criteria, mandatory deliverables, structured pilot, written readout. Below is the framework we run when operators ask us how they should evaluate any partner — including us.
Stage 1 — Discovery call (60 minutes)
The discovery call is not a sales meeting. It is a structured Q&A on five dimensions. If the partner cannot answer specifically and quickly, that is itself the answer.
The 12 questions
- What share of your client base is licensed iGaming, vs general performance / e-com?
- How many BMs do you currently maintain in our primary market?
- What is your account-survival workflow for a Meta BM suspension at scale?
- What server-side tracking do you deploy, and at what point in the engagement?
- How do you handle creative compliance pre-pass for our license class?
- What is your monthly creative output per market in similar engagements?
- Who specifically would be hands-on on our account? Salaried or contracted?
- What is your engagement model — retainer, % of spend, CPA, hybrid?
- What does your reporting cadence look like — daily, weekly, monthly?
- Can we speak to two of your current operator clients in our market?
- How would you handle a regulatory review or licence-side query during the engagement?
- What is the exit clause if either side wants to terminate after the pilot?
The answers do not need to match what we'd say. They do need to be specific. "We'll figure it out as we go" is a red flag for any of these.
Stage 2 — Audit deliverable
Before any commitment, the partner should deliver a written audit of your existing setup. The audit is itself a vetting tool: how thoughtful, structured, and honest is it?
What a good audit covers, in 5–10 pages:
- Current spend efficiency by channel and market.
- Tracking stack assessment (server-side, postback coverage, BI reconciliation gap).
- Creative library review — variety, freshness, compliance pass rate.
- Account-survival baseline — how exposed are you to BM suspension.
- Three specific recommendations for the next 90 days, ranked by impact.
An audit that says "everything looks good, ready to scale" is almost always wrong. The partner who reflexively flatters is the one who will not push back when you make the wrong call later.
Stage 3 — The 21-day pilot structure
The pilot exists to remove the largest source of partner-fit risk: that you cannot tell from a discovery call whether the team can actually execute. Twenty-one days is enough to see real numbers without committing to a long contract.
What a pilot must include
- One market, one channel cluster. Not multi-market, not multi-channel. The pilot tests execution, not coverage.
- Capped media commitment. Typically €40–80K total media spend. Enough to draw conclusions, not so much that termination is painful.
- Server-side stack deployed by day 10. If they cannot stand up CAPI and postbacks in 10 days, they cannot do it in 6 months.
- Daily Telegram or chat updates, weekly written readouts. Reporting cadence is the canary for ops cadence.
- CPA-only or hybrid payout. If they refuse to put skin in the game on a pilot of this size, decline.
- Exit clean clause. No automatic continuation. Decision point at day 21.
Stage 4 — Written readout
At day 21 the partner should deliver a written readout, 4–8 pages, covering:
- What was deployed (channels, BMs, creative, tracking, audiences).
- What was learned (CTR, CPA, D7 retention, account survival).
- What worked (with hypotheses for why).
- What did not work (with hypotheses for why).
- The recommended next quarter — scale, kill, or restructure.
- The contracted scope, scope, and pricing for the next quarter if continuation is recommended.
The readout is also a tell on partnership style. Concise and specific = good. 50-page deck with stock photos = bad. We mean this exactly: the format itself encodes how the team thinks.
The seven red flags
From the operators we have inherited from previous partners, these are the signals that should have ended the conversation earlier:
- "We can guarantee X CPA" — nobody can guarantee CPA in iGaming pre-engagement.
- "We have a network of BMs" — vague, often means rented or grey-market.
- "We use AI to optimise everything" — no specifics, no stack diagram.
- "References after we sign" — references are the input, not the reward.
- "Retainer + % of spend" — incentive structure rewards spend, not performance.
- "Our pilot is 90 days minimum" — they cannot prove themselves in 21.
- "We don't share tracking architecture" — engineering opacity is a strong negative signal.
The reverse vetting
One thing operators rarely do but should: ask the partner who they have declined to work with in the last 12 months and why. The answer reveals their risk thresholds, their licence-class bar, and how seriously they take the B2B-only positioning. A partner who has never declined a prospect is selling against everything, not curating.
For our part: in 2025 we declined seven of fifty discovery calls because the licence class did not meet our threshold or the operator was already mid-investigation in their primary market. We will tell you that on the call if you ask.
Closing the loop
If you complete this framework end-to-end, you will spend 4–6 weeks on partner selection and one quarter on the pilot. That is two-and-a-half months of due diligence before scaling commitment. It feels long. It is roughly one-fifth the cost of picking wrong.
Our 21-day pilot is structured exactly this way — see the Process page for the day-by-day. The framework is the framework whether it is us or someone else.