Case studies in iGaming marketing tend to be vague. "We exceeded KPIs by 30%" with no context about what the KPIs were, who was involved, or what actually drove performance. That's not useful for anyone trying to plan their own campaigns.
This is a real campaign. Brand and individual creator names are anonymized at client request, but every number is real, every decision point happened exactly as described, and the lessons are drawn from actual outcomes — not sanitized for marketing purposes.
Background and Brief
The client was a crypto-native casino that had been operating for 18 months. They had a functional affiliate program and a reasonable SEO base, but had never run an influencer marketing campaign. Their attribution infrastructure was standard — Income Access with S2S postback configured, unique codes per affiliate, 7-day attribution window.
They came to us with a specific ask: validate influencer marketing as an acquisition channel before committing to a larger budget. The test budget was €50,000 all-in, covering creator fees and our management fee. Campaign duration was six weeks. Primary markets were DACH (Germany, Austria, Switzerland) and the Nordic countries. Their product had strong live casino and crypto withdrawal features that we assessed as particularly marketable to the German-speaking audience.
The KPIs
The client set three KPIs that needed to be hit for the campaign to be deemed a success. Five hundred first-time deposits at a blended CPA under €80. A minimum 20% day-30 retention rate for that FTD cohort. And at least two creators who they'd want to work with again on an ongoing basis. That last KPI is often the most important in a test campaign — finding viable long-term partners is as valuable as hitting the acquisition numbers.
Streamer Selection Process
We started with a longlist of 20 streamers across Twitch and YouTube who met our baseline criteria: content in German, Swedish, Norwegian, or Danish; gambling-focused or high gambling content mix; audience geography aligning with target markets; no history of compliance problems or fraud flags in our network data.
From 20 to 8 involved three filter stages. The first was audience quality review: we pulled SullyGnome data on concurrent viewer trends and looked for any spike-and-cliff patterns suggesting bot inflation or purchased viewers. Three creators were eliminated here — two had suspicious viewership patterns, one had a concurrent viewer average that was dramatically inconsistent with their clip engagement rates.
The second filter was geographic validation. We requested audience screenshots showing top viewer countries from the remaining 17 creators. Five came back with audiences that were less than 40% in our target markets — not disqualifying by itself, but the combination of low geo match and competitive deal pricing made them lower priority. We deprioritized rather than eliminated three of these, keeping two with particularly strong slot content as backup options.
The third filter was a brief content review: had any of these creators previously promoted direct competitors in the past 90 days? Were their existing promotional segments done in a way that their audience appeared to trust? We assessed this qualitatively — chat reactions during sponsored segments, comment section sentiment on monetized content, and the creator's own approach to disclosure. Four more were eliminated or deprioritized at this stage.
The final eight included four mid-tier German-language streamers (2,000–8,000 concurrent viewers), two Swedish streamers in the 1,500–4,000 range, one Norwegian streamer with a tight, highly-engaged community, and one German YouTube creator who ran longer-form casino review content with strong search traffic. That last pick was deliberate — YouTube's longer attribution window and searchable content meant his videos would keep generating FTDs after the campaign technically ended.
Deal Structure
We used a hybrid flat plus rev-share model across all eight creators, but with different weightings based on our confidence in each creator's conversion potential.
For the four German Twitch streamers we had prior network data on, the deal was 60% flat fee paid upfront, 40% performance bonus structured as a flat payment per confirmed FTD above a threshold. No ongoing rev-share — the client didn't want perpetual liability from this test campaign. The per-FTD bonus was €15, paid monthly in arrears, for all FTDs generated within 30 days of each content appearance.
For the Nordic streamers and YouTube creator, who were newer to our network and carried more uncertainty, we weighted the deal more heavily toward performance: 40% flat fee, with a rev-share component of 15% NGR for 60 days post-campaign. This structure meant lower upfront cash commitment from the client on unproven channels, while giving the creators strong upside if their audiences converted well.
Total deal fees across all eight creators: €32,400 in committed flat fees and advance payments. The remaining €17,600 was reserved for performance payments and our management fee.
Content Execution
Each creator received a content brief with three required elements: a minimum of two dedicated streaming sessions per week featuring the casino, a promo code integration at session start and within the first 20 minutes, and a genuine disclosure that it was a sponsorship. The brief was explicit that gameplay sessions should be genuine — no pretending to win more than they were, no fake excitement for bad sessions.
Bonus Hunt Series
Three of the German streamers ran a weekly "bonus hunt" format where they'd open 15–25 bonus buys and reveal them in a single extended session. This format works well for casino brands because it generates natural tension, shareable highlights, and strong community participation (viewers vote on game order, guess win amounts). We coordinated a branded series name — the client's name plus "Hunt" — which created a recurring branded content property rather than isolated sponsorship moments.
Live Sessions and Affiliate Link Placement
For all streamers, the affiliate link was placed as a panel below the stream (always visible), in the stream description, and announced verbally at natural points in content rather than on a fixed schedule. The YouTube creator embedded the link in the first 150 characters of every video description and ran a pinned comment. We tested a chatbot command trigger (!casino) across all Twitch streams — more on this in the performance data.
Week-by-Week Performance Data
The campaign ran weeks one through six with a clean start date. All eight creators launched within 48 hours of each other.
Week 1: 87 FTDs. Strong opening, primarily from the two largest German streamers. CPA was running at €74 blended. Chatbot command click-to-registration rate was tracking at 4.2%, which was above our benchmark of 3%. No fraud flags. The Norwegian streamer underperformed initial projections significantly — 6 FTDs from an audience that had looked highly qualified. We scheduled a check-in call to assess.
Week 2: 112 FTDs. The bonus hunt series hit its stride, with one session generating 31 FTDs in a single stream. The YouTube creator's first video launched mid-week and brought in 18 FTDs in 72 hours, tracking well. Norwegian streamer improved to 14 FTDs after adjusting their session timing to better match their audience's online hours — peak audience was 45 minutes earlier than they'd been streaming.
Week 3: 134 FTDs. Cumulative 333 FTDs at day 21, on track for target. One German streamer flagged by our velocity monitoring — 22 FTDs from a single IP subnet within 36 hours. Fraud review initiated. Held that streamer's performance payment pending investigation; the cohort was ultimately confirmed as likely shared network (student dormitory based on the ISP data) rather than organized fraud. Included after manual review with a note.
Week 4: 98 FTDs. Expected mid-campaign dip as initial audience curiosity normalizes. Swedish streamers showed stronger relative performance this week — both had run audience Q&A sessions about the casino's crypto withdrawal features, which drove a conversion spike from their existing audiences who were already crypto-native.
Week 5: 121 FTDs. The YouTube creator's two additional videos published during weeks 3–4 were now compounding with the first, producing 28 FTDs in week 5 from search-driven traffic to his older content. That long-tail effect was the best-performing outcome of the campaign on a pure cost-per-FTD basis.
Week 6: 79 FTDs. Final week, some creator fatigue visible in content quality for two of the German streamers who were clearly running out of genuine enthusiasm. CPA for that cohort deteriorated noticeably.
Final Results
627 confirmed FTDs at a €71 blended CPA. That's 25.4% above the 500 FTD target and €9 under the maximum CPA. Day-30 retention: 34% of the FTD cohort was still active at the 30-day mark — 14 percentage points above the client's 20% minimum threshold.
By creator, the distribution was not equal. The two top-performing German streamers together delivered 41% of total FTDs. The YouTube creator delivered 19% on a cost basis that was dramatically more efficient than the Twitch streamers. The two Swedish streamers delivered 22% combined. The Norwegian streamer delivered 11% — modest by volume but with the highest day-30 retention rate of any cohort (44%), suggesting highly qualified audience even at lower volumes.
The client identified three creators they wanted to continue working with for the next campaign: both top German streamers and the YouTube creator. A fourth, one of the Swedish streamers, was offered an ongoing rev-share arrangement without a flat fee. The Norwegian streamer and two remaining German streamers were thanked and not continued — not because of poor performance, but because the budget focus for the next phase was going to concentrate on the top performers.
What Worked and What Didn't
The bonus hunt series format was the single highest-converting content structure. Concentrated sessions with community participation drove more same-session conversions than standard background gameplay. The chatbot command generated meaningful additional click volume — roughly 12% of total clicks across Twitch streamers came through the chatbot trigger, which costs nothing extra to implement and should be standard practice on every campaign.
YouTube long-tail performance exceeded expectations. We've seen this pattern across multiple campaigns: the YouTube creator's FTDs continue arriving for 60–90 days after the campaign formally ends as their videos rank for search terms. On a cost-per-FTD basis that includes post-campaign traffic, the YouTube creator was by far the most efficient channel in this campaign.
What didn't work well: two of the German streamers lost genuine enthusiasm by week 5–6, and it showed in their content and their conversion rates. The flat-fee-heavy deal structure removed their financial incentive to maintain quality late in the campaign. A better structure for those specific creators would have been a pure performance arrangement or a very small flat fee with a larger per-FTD bonus — keeping their financial interest active throughout.
The Norwegian market underperformed initial geographic projections. The audience was high-quality when it converted, but the volume was lower than expected. That streamer's community turned out to be more cautious about new crypto casinos than the audience profile suggested — the brand's lack of licensing recognition in the Norwegian-speaking market was a trust barrier that took longer to overcome. Future campaigns in that market need to budget for a longer warm-up period.
Lessons for Other Brands
Run the YouTube slot at the same time as Twitch. Most iGaming test campaigns use Twitch as default and treat YouTube as secondary. The cost efficiency argument for running YouTube content alongside Twitch is strong, and the long-tail benefit means your campaign keeps generating return after your budget is spent.
Structure deal incentives to stay active through the end of the campaign. Flat fees paid upfront create a quality cliff in weeks 5–6. Hybrid structures with meaningful performance bonuses keep creator energy up throughout.
Always investigate volume anomalies in real time, not after payout. The week 3 velocity flag could have become a dispute if we'd waited until the post-campaign fraud review. Catching it during the campaign, investigating quickly, and communicating transparently with the client meant it was resolved without damaging the relationship or the budget.
Day-30 retention is a more reliable campaign quality signal than raw FTD count. The Norwegian streamer's cohort had the best retention of the campaign despite the lowest volume. If you only looked at FTD count, you'd dismiss that creator. If you look at player quality, they're one of the most interesting channels for future investment.
If you're planning a first or second influencer campaign and want to build a structure similar to what produced these results, let's talk through your market and KPIs. We run campaigns across European markets and can tell you quickly whether a test structure like this makes sense for your situation.
Frequently Asked Questions
Can a €50k budget realistically hit 500 FTDs in competitive markets?
Yes, but it depends heavily on creator selection and deal structure. DACH and Nordics are expensive markets for influencer fees, but they also produce higher-quality players than many alternatives. The key is not spreading the budget across too many creators — concentrating spend on 6–8 well-selected creators outperforms spreading €50k across 20+ smaller channels. Our target in this campaign was 500 FTDs from 8 creators, and we exceeded it. A 20-creator approach with the same budget would likely have generated worse results and more management overhead.
How were fraud risks managed during the campaign?
Real-time velocity monitoring on all affiliate links, device fingerprinting at registration through the platform's built-in tools, and a daily review of cohort data by our team. The one anomaly in week 3 was caught and resolved within 48 hours. Four accounts were ultimately removed from the final FTD count — a sub-1% fraud rate, which is within acceptable parameters for a Twitch-heavy campaign.
Would this same structure work for a fiat casino vs a crypto casino?
The content format and creator selection process are essentially identical. The material difference is in audience targeting — crypto casinos have a natural affinity with streamers whose audiences skew crypto-aware, which influenced our Norwegian and Swedish creator picks. A fiat casino might adjust the geographic and demographic weighting, but the overall structure — hybrid deals, bonus hunt content, YouTube alongside Twitch — transfers directly.
What would you do differently if you ran this campaign again?
Replace the flat-fee-heavy deal structure for the two German streamers who faded in the final weeks with a performance-weighted hybrid. Add a YouTube creator from the start rather than treating it as a secondary channel. Build a longer warm-up brief for the Norwegian market that addresses the trust question around crypto casino licensing earlier in the campaign. And schedule week 4 as a "content review week" — watching sessions live and giving real-time feedback to creators — rather than relying on post-session reports.