A lookalike audience is a platform-generated cohort of users statistically similar to a seed list of valuable iGaming players, used to scale prospecting at higher conversion rates.
Lookalike Audience
**TL;DR:** A lookalike audience is a platform-generated cohort of users statistically similar to a seed list of valuable iGaming players, used to scale prospecting at higher conversion rates.
What it means
Operators upload a seed of high-value players — typically NDCs with 3+ deposits or top 20% by NGR — to Meta, Google, or TikTok. The platform builds a 1%–10% lookalike across the geo, matching on signals like declared interests, on-platform behaviour, device, and graph proximity. The closer to 1%, the higher the match quality and the smaller the reach.
In regulated markets, the seed must be built carefully: only players who consented to marketing, with PII hashed (SHA-256 email + phone). Sweepstakes and social casinos rely on lookalikes more heavily because their audiences are easier to model than full real-money operators.
Formula / How it's measured
Not applicable — this is a targeting mechanism, not a metric. Mechanism: hashed PII match rates of 40–70% are typical; the platform expands the matched seed to similar users. Effectiveness is measured by FTD rate, CAC, and downstream LTV vs broad targeting.
Example: a Mexican sportsbook seeds 25,000 NDCs into a Meta 1% lookalike. Match rate 55% → 13,750 matched users → 1% lookalike of ~800k Mexican adults. CPA on lookalike: $48 vs broad targeting $72.
Why it matters for operators
Lookalikes are the easiest scalable bridge between retention data and acquisition. A 20–35% CPA reduction vs broad targeting is realistic. They also let CRM/BI teams influence media: the seed you ship determines who the platform finds.
Common benchmarks (2026)
- Seed size minimum (Meta): 1,000 matched users; ideal 10k–50k
- Match rate hashed email + phone: 45%–65%
- Typical CPA lift vs broad: 20%–40%
- Best-performing seeds: VIPs, 90-day high LTV cohort
- Refresh cadence: every 30–60 days
Common mistakes
- Seeding with FTDs only — captures one-time bonus abusers
- Stale seeds (>90 days) that no longer reflect current product
- Using lookalikes in jurisdictions with strict consent rules without legal review
See also