The iGaming CPA Index: What New Brands Budget Above the Benchmark

Two operators’ own planning documents budget 2–3× the market rate for a new brand’s first 8–12 weeks. The penalty runs across the whole funnel, not just CPA.

Kazakhstan: +100–200% above benchmark. Russia: +150–250%. A second, independently built document’s own cost model runs a median 2.0× its cited benchmark across twelve markets. Every figure below is pre-launch planning data: a benchmark, a target, or a model the operator built before spending. None of it is confirmed post-launch spend.

Methodology

Two operators, pre-launch strategy documents, current (2025–2026):

• Operator A: multi-region crypto-only casino/slots launch. Scored twelve candidate markets on a weighted formula: 0.25× CAC + 0.25× LTV + 0.20× payback + 0.30× entry difficulty. A thirteenth candidate, Belarus, was audited but excluded before scoring.
• Operator B: CIS sportsbook + casino relaunch, RU/KZ core.

Not post-launch actuals: nothing above is confirmed spend. Not statistically representative: sixteen market-level observations (four phased, twelve modelled) from two documents, different methods and definitions, not a strict replication of one figure. Not independently audited: every figure traces to the operator’s own document, which itself cites third-party affiliate-network data (CPA.RIP, Affstar, and similar) of varying recency. Not internally reconciled: four markets carry two benchmark figures, two markets carry two internal targets, both shown, labeled by section, neither picked. Currency mixed (USD/EUR), not normalized.

Anonymized throughout: no brand names, license jurisdictions, or named individuals.

The Phased Plan

Operator B’s launch plan (§8.2) states a ramp-period CPA, a scale threshold, and an internal target for four markets, against the market rate it planned from:

Uzbekistan’s and Azerbaijan’s Phase 2 figures are labeled порог для масштабирования (“scale-up threshold”) in the source, not a stated target, unlike Kazakhstan’s and Russia’s.

Stop-loss: Kazakhstan halts scaling above $150 CPA for two consecutive weeks. Russia halts above $200, no time qualifier stated in the source.

The penalty isn’t confined to CPA:

Three mechanisms named in the source explain part of the gap, independent of either operator:

• Affiliates price in a margin for unknown conversion. A network can quote an established brand a rate because it has performance history to price against. It can’t do that for a brand it’s never run traffic to.
• Paid social needs a signal-training period. Meta specifically: 2–3 weeks before cost-per-outcome stabilizes, regardless of creative quality.
• The first 7–30 days of a CIS campaign are unprofitable by design. Payback is expected at D30–D90, not launch week.

Cost and Viability, Twelve Markets

Operator A’s dossier models its own expected CAC against the benchmark CPA it cites, plus LTV, LTV/CAC, payback, and an entry-difficulty score, for the same twelve markets. Sorted by LTV/CAC:

Median CAC: 2.0× benchmark.

Cheap CAC and good economics are different rankings. Kazakhstan is second-cheapest by CAC and sixth by LTV/CAC return. Argentina is third-cheapest by CAC and first by return. Norway has the highest LTV in the set and the second-worst payback.

Kazakhstan’s 6.2× compares $80–105 modelled CAC against Operator A’s own $10–20 benchmark. Against Operator B’s independently sourced $30–70 (§8.2) for the same market, the same modelled cost is ~1.5–2×, in line with the rest of the table. Likely a benchmark disagreement between the two operators, not a modelling anomaly.

Four caveats:

• CAC is defined as cost per FTD with a crypto deposit, narrower than a generic FTD. Part of the multiple reflects that qualifier, not a new-brand effect alone.
• Austria and Germany compare a EUR benchmark against a USD-modelled CAC, unconverted in the source.
• LTV is estimated from purchasing power, crypto adoption, and industry benchmarks ($150–800 over 6–12 months per §2.1), not directly measured.
• The entry-difficulty score is stated as weight-sensitive and meaningful at tier level, not by individual point (§1.3).

The Index: 15 Markets

Four markets (Kazakhstan, Russia, Uzbekistan, Azerbaijan) carry two market-benchmark figures: §2.2 (early audit) and §8.2 (phased plan), consistently higher. Kazakhstan and Russia carry two internal targets (§6.4, §8.2). Both shown, labeled, not reconciled. The source doesn’t say which is authoritative.

Belarus benchmarks cheapest in the table. The source reports conversion there as low, for that profile specifically, because of payment-rail blocks. A cheap CPA on a payment rail that doesn’t reliably convert isn’t a cheap market. Licensed operators already established in Belarus are a different case.

Spread: ~15–20× between cheapest and most expensive on the §2.2 figures. On the §8.2 figures for the same two cheap markets (Uzbekistan, Kazakhstan), the spread compresses to ~6–9×. The same document states both.

For Planning

• Budget the ramp period separately from steady state. The multiplier is market-specific, not a constant to copy between markets.
• Budget for the whole funnel degrading, not only CPA. Registration→FTD conversion, retention, churn, payment success, and LTV all ran below target in the one plan detailed enough to check.
• Set stop-loss thresholds before launch, with an explicit response attached. A ceiling with no predefined action isn’t a control.