I’ve read a lot of “affiliate fraud” guides that rehash the usual suspects—bots, cookie stuffing, brand bidding. Useful, sure. But the most expensive hit I’ve seen didn’t look like any of those. It was glossy, “legit,” and wrapped in the halo of influencer marketing. In Brazil’s rush, we okayed a €300 CPA with a €10 baseline. An affiliate quietly paid €100 per FTD to influencer reps. The traffic looked clean—until LTV collapsed. We wired ~€500k in CPAs and got ~€50k back from those players. Classic incentivized FTDs disguised as creator love.
Below is the full breakdown: what happened, how to spot it fast, and how to harden your program without killing growth.
Why classic guides miss the real threat
Most playbooks treat affiliate fraud as a technical hygiene problem—block bad devices, spot bot clusters, patrol brand terms. All fair. But modern fraudsters aren’t just hiding in logs; they’re hiding in relationships. Agencies and influencer reps can convert on paper while quietly subsidizing the baseline behind the scenes. The promo looks organic: a creator “recommends” your casino or sportsbook; attribution flows through neat sub-IDs; deposits land; your dashboards sing.
Meanwhile, the intent is rotten. Players weren’t persuaded by your brand—they were paid to cross the baseline. That intent gap is the difference between a cohort that sticks and one that churns before D7. Any time someone demands no baseline or soft KPIs, my guard goes up. Quality traffic doesn’t fear cumulative baselines.
This is why classic guides miss it: they treat “influencer” as branding, not acquisition. But in an iGaming CPA deal, creators are acquisition channels with cashable KPIs. If someone can arbitrage your CPA with a secret bounty (e.g., €100 per FTD), your unit economics are their playground.
The setup: A €300 CPA deal meets a €10 baseline
Context matters. Brazil was red-hot; competition pushed CPAs up. An affiliate with prior success asked for €300 CPA on casino with a €10 first deposit baseline. On paper, it wasn’t insane. In practice, it was a trust test we failed. We leaned on historical performance and relaxed the usual friction: modest caps, a short hold, and optimistic assumptions about LTV for that geo.
The mechanics were straightforward: the operator pays €300 once a player makes a €10 deposit. For good affiliates, baselines aren’t scary—they align incentives around real value. But baselines can be abused when third parties cover the deposit. If the end user doesn’t feel any cost, your “conversion” is just a payout trigger.
Trust, history, and high CPAs are a combustible mix. An affiliate with prestige can argue for faster approvals, looser holds, and higher caps—exactly what lets a subsidy scheme scale before anyone blinks.
The playbook: incentivized FTDs disguised as “creator love”
Here’s what actually happened. The affiliate leaned into influencer marketing and, through reps, offered ~€100 per FTD off-platform. Creators or their agents handled the messaging: “Sign up here; I’ll cover your first deposit.” The content looked like genuine endorsements—no blatant “cashback” banners, nothing that screams incentive. To users, it felt like a perk from the creator, not the operator.
Because the incentive sits one layer away (the rep, not the operator), it evades your usual checks. Attribution is clean. Fraud tools see normal devices, normal IPs, reasonable velocity. Sub-IDs map to individual influencers, which actually increases perceived legitimacy. Meanwhile, the money loop is brutal: €300 in, €10 out, and most players never become net-positive.
Since then, we interrogate creator economics: “How exactly do you pitch the offer?” “Show sample scripts.” “Who funds promos?” If an agency dodges those questions—or insists they “can’t share” scripts—that’s a smell. We cap, hold, and claw back when influencer-sourced sub-IDs tank retention.
The bill: €500k out, ~€50k back—when LTV collapses
The KPI autopsy
- FTD rate: spikes beautifully.
- Valid FTD %: looks decent at first because the baseline is met; quality signals lag.
- ARPU/LTV by cohort: craters. Players make the minimum deposit and vanish.
- Retention (D7/D30/D60): near-zero; cohorts behave like one-time triggers.
- NGR per cohort: negative after CPA.
- Payback period: effectively infinite.
- Blend CPA vs. MER/ROAS: wildly misaligned.
The operational pain is timing: you’ll likely be paying CPAs weekly while the meaningful LTV picture takes months to materialize. By the time finance screams, you’ve already shipped hundreds of thousands. That’s why holds and cumulative baselines matter. If you’re going to run high CPAs, either release on lagging quality KPIs (e.g., D7 retention threshold) or enforce contractual clawbacks—for real.
Spot it early: a 12-point detection checklist
This is the field checklist we wish we’d had on day one. It includes the baseline clustering signal and the “baseline + a little” camouflage you mentioned.
Signals you can monitor today
In Brazil’s rush, we saw exactly this: 100+ FTDs landing right on the €10 baseline, with a secondary bump at €11–€12. Once we split by sub-ID, the pattern was unmistakable.
Deal hygiene: rewrite your affiliate T&Cs before you scale
The non-negotiables
- Explicit incentive bans: no third-party payments, cashbacks, deposit coverage, or “gifts” tied to registration/FTD.
- Disclosure & proof: affiliates must disclose creator relationships and provide sample scripts/placements on request.
- Granular tracking: mandatory sub-IDs per creator/campaign; no pooled traffic.
- Pre-approval: creators and agencies require written approval before going live.
- Caps & holds: default caps per creator; holds released on quality, not just FTD count.
- Clawbacks: retroactive clawbacks for misrepresented or incentivized traffic; define evidence standards.
- Audit rights: right to audit creator payouts and communications tied to your brand.
- Termination triggers: clear thresholds (e.g., D30 ARPU/CPA ratio) that allow immediate pause.
Since that Brazil episode, I only green-light creator campaigns where economics make sense without hidden subsidies.
Playbook to recover: audits, clawbacks, and how to brief legal
Eight steps that actually work
- Freeze and preserve: pause payouts; snapshot click→signup→FTD; save creatives, scripts, sub-ID logs.
- Cohort proof pack: simple charts: FTD spike vs. D7/D30 retention and ARPU collapse by sub-ID.
- Creator audit: request placement proofs, scripts, and any “rebate/coverage” messages. Lack of cooperation is evidence.
- Contract leverage: invoke incentive bans, misrepresentation clauses, and quality thresholds.
- Negotiate clawbacks: start with the most egregious sub-IDs; tie refund % to KPI gaps (e.g., D30 ARPU/CPA ratio).
- Network route: if a network is involved, package the evidence clearly—make agreement the easy path.
- Refactor deal: if you keep them (rare), move to rev-share or hybrid with strict baselines and long holds.
- Post-mortem: codify signals, update T&Cs, and align finance on invoice timing to suit holds.
What evidence persuades partners? Clean, visual before/after comparisons vs. your channel median; creator-level anomalies; and any screenshots or messages implying deposit coverage.
FAQ: your toughest questions answered
Is incentivized traffic always bad?
For iGaming CPAs—yes. If someone covers the deposit, your “conversion” is a payout trigger, not intent.
What’s a safe baseline?
Depends on brand and geo, but the principle is cumulative baselines + quality holds. If a partner resists both, slow down.
How soon can I tell if a creator is bad?
Within 72 hours you’ll see ARPU compression and a spike in baseline-only deposits. By D7, retention gaps are undeniable.
Should I ban influencers altogether?
No. Treat them like PPC: structured tracking, caps, holds, audits, proof of placements, and economics that work without rebates.
When do clawbacks stick?
When your T&Cs are explicit, your sub-ID data is clean, and you can show creator scripts or messages implying deposit coverage.