The numbers that describe a sportsbook
The first lesson explained that a bookmaker's profit comes from margin rather than prediction. This lesson explains how that shows up in the numbers a business reports, which are the numbers you will read in results announcements, analyst notes and this site's coverage.
Handle, stakes or turnover is the total amount bet. It is the headline figure regulators and states publish, and it is the least informative about profit, because most of it is paid back to customers.
Gross gaming revenue (GGR), or gross win, is handle minus winnings paid out: what the book kept. It is the base most betting taxes are charged on: Britain's General Betting Duty is 15% of a bookmaker's profits, calculated as stakes received less winnings paid out, with a new 25% rate for remote bets due from 1 April 2027 (remote bets on UK horse racing stay at 15%), and New Jersey has taxed online sports wagering gross revenue at 19.75% since 1 July 2025.
Hold, or margin achieved, is GGR divided by handle, expressed as a percentage. If customers bet 100 million and the book kept 6 million, hold was 6%. Hold is the theoretical overround adjusted for results: in a month where favourites win, hold falls; where longshots win, it rises. Over a long period hold converges toward the book's structural margin, and a sportsbook's hold rate is one of the most-watched figures in the industry. In 2025 US commercial sportsbooks kept $16.96 billion of the $166.94 billion staked, a hold of about 10%, while Nevada's sportsbooks held 7.45% across the year. Flutter reported a structural revenue margin on sportsbook, which it presents separately from the effects of sports results and player incentives, of 14.2% at FanDuel in the US and 16.6% in its international business in 2025, and attributed the increase to its pricing and to customers betting more parlays.
Net gaming revenue (NGR) is GGR minus bonuses and free bets given to customers. Large listed operators such as DraftKings and Flutter report revenue on this basis: DraftKings records customer incentives as a reduction to revenue and books gaming taxes, payment processing and platform costs as cost of revenue, and its sportsbook revenue in 2025 was 7.1% of $53.6 billion of handle. Some definitions also deduct gaming tax and platform fees, so check which one a company or contract uses. This is what is left to run the business.
Actives are customers who bet in the period; average revenue per user is NGR divided by actives; customer acquisition cost is marketing spend divided by new customers. Together with churn, these tell you whether a book is growing profitably.
Where the margin goes
From NGR, a sportsbook pays gaming tax (where not already deducted), data and content costs (the feeds and the pricing), marketing (the largest discretionary cost: DraftKings spent $1.38 billion on sales and marketing in 2025, about 23% of revenue, down from about 33% in 2023), payment processing, compliance and technology, and people. What remains is EBITDA. In 2025 Flutter's International segment, which covers Britain and Ireland, Italy, Australia and other markets outside the US, reported an adjusted EBITDA margin of 23.4% of revenue, against 13.2% in its US segment, up from 5.3% in 2023, helped by revenue growth and by marketing falling as a share of revenue; both figures include online casino. In newly opened markets an operator can lose money for years while it buys customers: DraftKings reported net income of $3.7 million in 2025 after a history of losses, including a net loss of $507.3 million in 2024.
This is why results announcements are read the way they are: a book that grows handle but sees hold fall is being unlucky or pricing too keenly; one that grows NGR while cutting marketing is harvesting a mature base; one with rising acquisition cost and flat actives is fighting for share in a saturated market.
Why bettors lose
Most customers lose over time, and it is worth being precise about why, because the industry's social licence depends on being honest about it.
The margin is the first reason: at 5% overround a customer who bets randomly loses just under 5% of stakes over time (1 minus 1/1.05, about 4.8%), and at 20% on accumulators, far more. The favourite-longshot bias is the second: bettors overbet longshots and underbet favourites, so long odds tend to offer the worst value. Parlays are the third: the product with the biggest payouts has the biggest compound margin. And behaviour is the fourth: chasing losses, betting on what you support rather than what is priced wrong, and betting more when winning.
A minority of customers are ahead, and they are the ones books are most likely to restrict. When the Gambling Commission collected data in early 2025 from large online betting operators, covering almost 15 million British accounts active in the previous calendar year, 25.42% were in lifetime profit and 72.54% in loss; 4.31% of accounts had been restricted, and 46.78% of restricted accounts were in profit. The rest are recreational, and the industry's position is that they are paying for entertainment, which is true for most and untrue for some. That is where responsible gambling regulation enters, and why licensed sportsbooks monitor for the customers for whom it has stopped being entertainment: in Britain, remote licensees must monitor customer activity to identify harm from the point an account is opened.
Products and profit
Not all products earn the same. In rough order of margin: pre-match singles on major markets earn the least (competitive, price-shopped); in-play usually earns more (wider margins, less price-shopping); props and multi-outcome markets usually earn more still; accumulators and same-game parlays earn the most. Nevada's figures show the gap: in 2025 its sportsbooks held 5.88% on basketball and 7.98% on football, but 32.59% on parlay cards. A book's product mix, and its ability to steer customers toward higher-margin products through promotion and design, is a large part of its economics, and it is also where regulators look hardest at whether promotion has become manipulation.
The competitive landscape
Sports betting operators fall into rough groups: the large international groups with many brands across many markets; national champions that dominate one country; the US-focused operators that grew with state-by-state legalisation; betting exchanges; and the B2B suppliers that provide platforms, pricing and data to all of them. The listed companies among these are tracked on this site's iGT 25 index, and their share prices are one of the clearest reads on how the business is going.
The industry's structural trends: consolidation, because scale spreads the fixed costs of technology, compliance and marketing; the shift of turnover to in-play and to parlays; the rise of the United States, where commercial sports betting revenue grew 22.8% to $16.96 billion in 2025, as one of the largest markets; regulation tightening, especially on advertising and affordability (since August 2024 British remote operators have had to run financial vulnerability checks, now triggered at net deposits of £150 over 30 days); and prediction markets emerging as a competitor with a different legal basis: event contracts traded on exchanges regulated by the federal Commodity Futures Trading Commission rather than by state gaming regulators, a market DraftKings itself entered in December 2025.
Careers
Sports betting employs people in trading (pricing, risk, in-play), quantitative modelling, product management, data engineering, marketing and CRM, affiliate management, customer service, compliance (AML, responsible gambling, marketing review), payments and fraud, and the B2B side (data, platforms, integrity). Trading and quant roles reward numeracy and composure; product and marketing reward understanding the customer; compliance rewards rigour. A grounding in how odds work, what the products are, how a book is run, how bets settle and what the market is saying, which is what this course has tried to give, is the common base for all of them.
What to carry forward
Convert any odds to a probability and sum a market to find its margin. Know the core markets and why bet builders are different. Understand that a book manages prices, limits and liability rather than predicting results. Know the settlement rules that generate disputes. Read a price move as information, and judge a bet against the closing line. And read a results announcement knowing that handle is not revenue, hold is luck plus margin, and NGR is what the business lives on.