A games business with a casino's customers
The economics of a sweepstakes casino are a hybrid, and the mistake most analysts make is to model it as one or the other. It sells packages like a mobile game and pays prizes like a casino. Its revenue line is purchases, its cost line includes redemptions, and its margin is the difference, net of the cost of acquiring and keeping the people who buy.
This lesson builds the unit economics from the player up, then compares them with a licensed casino, then explains why the numbers that make the model attractive are also the numbers that make it a political target.
Who pays
The player base has the shape of every free-to-play product. A large majority never purchase; they play gold coins, collect the daily sweeps coin bonus and occasionally redeem a small prize. A minority purchase at all. Sweepstakes operators are private and do not publish the figure, but listed companies with social casino games do: DoubleDown Interactive's payer conversion rate, the share of monthly active users who made a purchase in the month, was 8.2 per cent in 2025, up from 6.7 per cent in 2024, and Playtika, a mobile games company whose social casino-themed titles made up 29.2 per cent of its 2025 revenue, reported an average daily payer conversion of 4.4 per cent across its games in 2025. A small fraction of purchasers account for most of the revenue.
The concentration is more extreme than in licensed casino because the free tier is so large. A licensed casino's active customers are, almost by definition, depositors; a sweepstakes casino's are mostly not. The metrics the segment borrows from mobile gaming reflect that: conversion rate, the share of players who ever purchase; average revenue per paying user, the monthly purchase value among those who do; and the share of revenue from the top one per cent, the same whale concentration that social casino companies disclose to investors: Playtika reports that a large percentage of its revenues comes from a small subset of its paying users.
The consequence is a business that is, in revenue terms, a VIP business wearing a mass-market costume. The public-facing product is free play for a large audience; the profit comes from a comparatively small group of heavy purchasers, and everything in the CRM stack exists to find them, convert them and keep them.
The purchase, decomposed
Take a $50 package: gold coins plus 50 sweeps coins as a bonus, a one-to-one ratio.
Revenue is $50.
Payment cost is higher than in licensed gambling, because the transaction is a card purchase of a digital good through processors who price the segment's risk into their fees. Operators do not publish the rate, and this worked example assumes 7 per cent; it runs higher where the operator has been pushed onto secondary processors after a primary one withdrew.
Redemption cost is the expected cash value of the 50 sweeps coins. They are played through the games at a configured return; what survives is redeemed. With a 95 per cent RTP and a one-times playthrough, the expected surviving value is around 47.5 coins, or $47.50, before further play. But players do not stop at one playthrough; they keep playing, and each further cycle retains another five per cent. The realised redemption rate, the share of bundled sweeps coin value that is eventually paid out as prizes, is the single most important number in the model, and operators manage it through the playthrough requirement, the redemption minimum, the RTP and the product mix. Realised rates in the range of 60 to 80 per cent of bundled value are the sort of figure that makes the model work; a rate near 95 per cent does not. At sector level, Eilers & Krejcik Gaming estimates cited by KPMG put 2024 sweepstakes casino gross revenue above $10.6 billion and net revenue above $3.4 billion. The primer does not define net revenue, but if it is read as purchases less prize redemptions, roughly two-thirds of purchase value went back out as prizes.
Gross margin is what remains: revenue, less payments, less realised redemptions. On the figures above, a $50 purchase with 7 per cent payment cost and a 70 per cent realised redemption rate leaves $50 minus $3.50 minus $35, which is $11.50, or 23 per cent of revenue.
That margin is on a different base from a licensed casino's, and the comparison needs care.
The comparison with licensed casino
A licensed casino reports gross gaming revenue: stakes less winnings. A sweepstakes casino reports purchases. To compare them, convert the sweepstakes margin to the equivalent of GGR: the $11.50 retained on a $50 package is the operator's "win", and the redemptions are the "payouts". On that basis the sweepstakes operator's net gaming revenue is its margin, and its cost structure below that line is the interesting part.
No gaming tax. A licensed online casino pays state tax on GGR at rates set state by state and, in some states, game by game: 15 per cent in West Virginia, and in Pennsylvania 14 per cent on online table games and 52 per cent on online slots. A sweepstakes operator pays no GGR tax, on the argument that it is not a gambling operator, but is subject to federal and state corporate income tax and, in a growing number of states, may have to collect sales and use tax on its gold coin sales. On $11.50 of retained value, the licensed operator in a high-tax state keeps around $5.50; the sweepstakes operator keeps most of the $11.50.
No licence cost. No application fees, no per-state compliance functions, no regulatory reporting, no certification of games and systems to a state standard.
Much higher marketing. The segment's growth was bought with national advertising, influencer campaigns and affiliate commissions, largely outside the state advertising rules that restrict where and how licensed operators can advertise; an Eilers & Krejcik analysis for the industry counted $44 million in affiliation deals in New York alone in 2024. Customer acquisition cost as a share of revenue is the line that most distinguishes the operators that are profitable from the ones that are growing.
Higher payment cost, higher fraud, higher churn. The trade-offs of a grey market: worse processing terms, promotional abuse from players who farm free coins, and a customer base that leaves when a state closes or a brand's reputation turns.
Netted out, a well-run sweepstakes brand in an unlicensed state earns a margin on retained value that a licensed operator cannot match, because in the high-tax states the largest cost in licensed online casino is the tax, and the sweepstakes brand does not pay it. That is the economic fact that drove the segment's growth and the political fact that is ending it: the margin is, in the states' reading, tax that is not being collected.
The numbers a state sees
The same Eilers & Krejcik estimates put sweepstakes casino gross revenue above $10.6 billion in 2024, which KPMG says followed four years of 60 to 70 per cent compound annual growth from 2020, against about $7.1 billion in 2024 for the global social casino market, where many analysts project flat results for 2025 to 2027: sweepstakes is now the larger and faster-growing part. A state with, say, a five per cent population share and a licensed online casino tax of 20 per cent can compute what it is not receiving, and the licensed operators and tribal casinos in that state have made the computation for it. A ban is not only a consumer-protection measure; it is a fiscal one, and the bills have the licensed industry's support. California's AB 831 was sponsored by the Yuhaaviatam of San Manuel Nation and supported by other tribes, the American Gaming Association, the Sports Betting Alliance and Light & Wonder, and opposed by the sweepstakes trade groups and VGW.
The industry's counter-offer, a licensed sweepstakes regime with a tax, is an attempt to convert that fiscal argument into a reason to regulate rather than prohibit. In New York the Social Gaming Leadership Alliance proposed operator registration fees and a tax on player purchases in place of a ban, and in Florida a 6 per cent tax on player purchases. Lesson seven considers whether any state will take it.
Metrics for running one
An operator or an investor assessing a sweepstakes brand should have the following on one page.
- Registered players, monthly active players and monthly purchasers, and the conversion rate between them.
- Average revenue per paying user and its distribution, with the top one per cent's share.
- Sweeps coin ratio on purchases, blended across promotions.
- Realised redemption rate, bundled value eventually paid out, by cohort.
- Free-route share, the proportion of sweeps coins in circulation that came from postal entry and daily bonuses rather than purchases. This is an economic number and a legal one; lesson three explains why.
- Payment cost and decline rate, by processor.
- Customer acquisition cost by channel, and payback period.
- Excluded-state exposure: the share of revenue from states with active bans, enforcement or litigation, which is the number that decides what the business is worth.
The last of these has moved from a footnote to the headline. A brand earning a third of its revenue from states that have banned the model, or are about to, is worth up to a third less than its run-rate suggests, and a buyer will price it that way. The exposure is concentrated: analysts at the Global Gaming Expo in October 2025 estimated California at about 20 per cent of sweepstakes casino revenue, and the Eilers & Krejcik study put New York at $762 million of 2024 sales, and both states have since enacted bans.
What to take from this lesson
A sweepstakes casino sells packages and pays prizes; its margin is the package price less payments less realised redemptions, and the realised redemption rate is the number that runs the business. Its structural advantage over a licensed casino is the absence of gaming tax and licence cost; its structural disadvantages are payment cost, marketing cost and legal exposure. The margin that makes the model attractive is, from a state's point of view, uncollected tax, and the excluded-state share of revenue is now the number that determines what an operator is worth.