The five deal structures
Almost every iGaming streaming deal is a variation on one of five structures. They differ less in headline cost than in who carries the risk.
| Structure | Who carries risk | Typical use |
|---|---|---|
| Flat fee — per stream or monthly | Buyer | Top-tier channels, launch campaigns, exclusivity |
| Affiliate revenue share | Streamer | Mid and long tail, ongoing relationships |
| CPA per depositing player | Streamer | Performance-led acquisition, mid tier |
| Hybrid — retainer plus CPA | Shared | The most common structure overall |
| Balance-funded play | Buyer | Bonus hunts, high-stakes content |
The fifth deserves attention because it is the least visible to the audience. When the operator funds the stake, the risk shown on screen is not the streamer's. Several jurisdictions treat undisclosed funded play as misleading advertising, and it is the structure most likely to become a compliance problem retrospectively.
What actually drives the price
Buyers routinely anchor on follower count. It is close to the least predictive input available.
Six inputs move the price, roughly in order of weight:
- Airtime-weighted average concurrency. What a sponsored hour actually reaches. Peak viewers is a negotiating prop.
- Format. An overlay logo, a mention cadence, a dedicated segment and full-stream exclusivity are four separate products. Segment-level sponsorship costs multiples of a logo placement for the same channel.
- Exclusivity. Whether competing operators appear on the same channel in the same window. This is the single largest multiplier at the top of the market.
- Casino share of airtime. A 100%-casino channel has a warm, self-selected audience already seeing competing offers. A 25%-casino variety channel has a colder, larger, less contested audience. Both are defensible buys at different prices.
- Language and market. The scarcity of channels in a language sets the floor. Arabic has 638 channels for 15.5M hours — a seller's market. Spanish has 9,245 channels for 28.7M hours — a buyer's market.
- Delivery reliability. A channel with a 1.5× peak-to-average ratio can be modelled. One at 8× cannot, and should be priced with that variance in mind.
Building an eCPM benchmark without a rate card
You do not need the seller's pricing to know whether a price is sane. Five steps:
- Establish delivered hours. Take the channel's hours watched over the last 90 days inside casino categories. That is average concurrency × airtime, and it is measurable from public data.
- Convert to impressions on the sponsored unit. For a full-stream sponsorship, impressions ≈ average concurrency × sponsored hours. For a segment, restrict to the segment's hours. Do not use the whole stream for a segment deal — that is the most common inflation in seller decks.
- Assemble a comparable set. Three to five channels in the same language, the same concurrency band (within roughly ±50%) and a similar casino-airtime share. Cross-language comparables are close to meaningless.
- Derive the eCPM range from what those comparables cost for the same format. Where you cannot see a comparable price, non-gambling sponsorship rates in the same market and concurrency band give a defensible floor.
- Apply a quality discount before comparing. A channel that fails the audience-quality checks does not deserve a same-band eCPM; it deserves a lower one, or no bid.
A worked example
Take a mid-tier Kick channel, using the shape of real numbers from our dataset:
- Average concurrency: 4,200
- Casino airtime, last 90 days: 210 hours
- Deal on the table: full-stream sponsorship, 40 hours over one month, flat fee
Estimated delivery: 4,200 × 40 = 168,000 watch-hours. Treated as impressions on a persistent overlay, that is 168,000 viewer-hours of continuous exposure — not 168,000 discrete impressions, and the distinction matters when comparing to display advertising.
At an assumed $8 CPM for a persistent-overlay hour in this band, the defensible range is roughly $1,300–1,500 for the month. If the ask is $6,000, either the format is being priced as exclusivity rather than a logo, or the price is not supported by delivery. Both are reasonable conversations to have — but you can only have them with the delivery number in hand.
The same arithmetic on a top-tier channel gives a very different answer, which is why top-tier deals cannot be benchmarked against mid-tier ones. Trainwreckstv averaged 29,916 concurrent viewers across 483 hours in H1 2026 — a single 4-hour stream delivers more watch-hours than the mid-tier example's entire month.
Contract terms that make delivery measurable
Most disputes in this vertical are not about whether the streamer cheated. They are about the fact that nobody defined what delivery meant. Five clauses fix that:
- Define the sponsored unit. Overlay for the full stream, or a segment of stated minimum length? Written down, with the minimum.
- Set a minimum-delivery floor in watch-hours or average concurrency, with a make-good if it is missed. This converts an unpredictable event channel into a priceable one.
- Require category and title accuracy. The stream must sit in the correct category. Miscategorised streams break third-party measurement, which is often the only measurement you have.
- Agree the measurement source in advance. Name the third-party provider in the contract. Arguing about whose numbers count after the campaign is unwinnable.
- Specify exclusivity and disclosure explicitly. Which competitors are excluded, in what window, and what disclosure language appears on screen and in the title. Disclosure is a regulatory requirement in several markets, not a courtesy.
Five expensive mistakes
- Paying on followers. The vertical's hours-per-follower ratio varies by more than an order of magnitude across the top twenty channels.
- Accepting peak viewers as the reach number. Peak-to-average ratios in our dataset run from 1.5× to over 20×. Peak describes one minute of a half-year.
- Buying English-language reach for a non-English market. More than half the vertical's hours are non-English, and audiences do not cross over.
- Skipping the audience-quality check to move fast. The checks cost hours. A bad top-tier deal costs five figures a month.
- Measuring the campaign after it ends. Without a pre-campaign baseline for the channel there is nothing to compare the result to, and no way to tell a good outcome from a normal week.
Related reading: spotting inflated viewership, platform gambling rules, and the top 20 casino streamers with the delivery metrics for each.
General commercial and measurement guidance based on 380.4M hours of observed viewership. Not legal, financial or investment advice.