GGR, NGR, bonus cost, lifetime value and payback, and how to tell a healthy operator from a busy one.
In this lesson:
- Move confidently between turnover, GGR, NGR and contribution, and explain what each one excludes
- Calculate and interpret the core player metrics used across the industry, including ARPU, CAC and LTV
- Explain cohort analysis and payback period, and why they reveal problems that headline revenue conceals
- Judge whether an operator's growth is being bought or earned
Why the vocabulary matters more than it should
A surprising proportion of disputes inside gambling companies are not disagreements about performance at all. They are two people quoting different metrics at each other without realising it. Marketing talks about revenue meaning gross, finance talks about revenue meaning net of bonus and duty, and the commercial team is quoting turnover because it is the biggest number available.
Getting the vocabulary precise is therefore not pedantry. It is the difference between an argument that can be resolved and one that cannot.
The revenue stack
Start at the top and work down. Each step removes something, and each step matters.
Turnover, also called handle or amount staked, is the total value of bets placed. In a casino this includes every restake of previously won funds, which is why casino turnover figures can look extraordinary relative to the money that actually entered the business. A player depositing £100 and playing through a 96% RTP slot may generate several thousand pounds of turnover before their balance is exhausted. Turnover is a measure of activity, not of value.
Gross gaming revenue is turnover minus winnings paid out. This is the first figure that describes money the operator has actually retained, and it is the standard basis on which gaming duty is calculated in many jurisdictions and the standard basis on which suppliers are paid revenue share.
Net gaming revenue deducts the cost of bonuses awarded to players and, depending on the definition in use, gaming duty and certain direct costs such as payment processing or platform fees. This is where comparison becomes hazardous, because there is no universally agreed definition of NGR. One company may deduct only bonus cost. Another may deduct bonus cost, duty, payment fees and platform revenue share. Both will call the result NGR. Anyone comparing two companies on NGR without first checking the basis is likely to reach a false conclusion.
Contribution goes further, deducting the directly variable costs of serving customers, including marketing attributable to them, payment costs, supplier revenue share and customer support cost. Contribution is the figure that tells you whether the customer relationship itself is profitable before any fixed costs of running the business.
Below contribution sit the fixed costs: technology, compliance, staff, offices, licence fees. What remains is operating profit, usually reported as EBITDA in this sector.
The single most useful habit to build is asking, whenever a revenue figure is quoted, which level of that stack it sits at and what has been deducted to arrive at it.
Hold, margin and their relationship
For the sportsbook, the relationship between turnover and GGR is expressed as hold, which is GGR as a percentage of turnover. Theoretical margin is what the pricing was designed to deliver; actual hold is what results produced. The gap between them is the vertical's volatility, discussed in the previous lesson.
For casino, the equivalent concept is the house edge, and because outcomes are governed by the operator's own certified mathematics, the actual result tracks the theoretical one closely at volume. Casino revenue divided by casino turnover across a large player base will sit very near the blended house edge of the games being played.
There is a subtlety worth noticing here. Because casino turnover includes restakes, the ratio of casino GGR to deposits is a far more meaningful commercial measure than the ratio of GGR to turnover. What the business actually converts is deposited money, not staked money, and a casino that extracts a high proportion of deposits is not necessarily healthier than one that extracts a lower proportion, because the latter may be producing longer, more satisfying sessions and better retention.
Player metrics and the tyranny of the average
The industry uses a standard set of per-player measures, and every one of them requires care.
Active players counts customers who placed at least one bet in a period. Monthly actives is the usual reporting basis. It is a volume measure and reveals nothing about value.
First time depositors, or FTDs, counts customers making an initial deposit. This is the standard unit of acquisition and the denominator in acquisition cost calculations.
Average revenue per user divides revenue by active players. This is where most analytical errors originate.
The problem is that revenue in gambling is distributed with extreme skew. A minority of players generates a majority of revenue, and in some operations the concentration is severe. In that distribution, the mean describes nobody. An ARPU of £80 might represent a base where most players generate £10 and a handful generate thousands, or a base where most players generate £70 and the top end is modest. Those are entirely different businesses with entirely different risk profiles, and ARPU cannot distinguish them.
The correction is segmentation. Report revenue by value decile, look at the median as well as the mean, and track what proportion of revenue comes from the top few percent of players. That last figure is one of the most informative numbers in the entire business, both commercially and from a responsible gambling perspective, and the two considerations point in the same direction: heavy dependence on a very small group of very high-spending customers is a concentration risk in both senses.
Acquisition cost and lifetime value
The central commercial question for any operator is whether the customers it buys are worth more than it pays for them.
Customer acquisition cost divides total acquisition spend by the customers it produced. Doing this properly is harder than it sounds. Acquisition spend must include media, affiliate payments, the cost of welcome bonuses and free bets, creative production and the acquisition-attributable share of sponsorship. Excluding bonus cost from CAC, which happens more often than it should, produces a flattering figure that has no relationship to reality.
CAC also varies enormously by channel, by market and by moment. Acquiring a customer through organic search costs almost nothing incrementally. Acquiring one through television during a major tournament, in a newly opened market where every competitor is bidding for the same attention, can cost a very great deal. Blended CAC across all channels hides this variation, so channel-level CAC is what actually informs decisions.
Lifetime value estimates the total net contribution a customer will generate across the entire relationship. It depends on how much they deposit, how often, at what margin, for how long before they churn, and what it costs to serve and retain them.
The relationship between the two is the test of viability. If LTV exceeds CAC by a comfortable multiple, acquisition creates value and the correct response is to spend more. If it does not, the operator is buying revenue at a loss, and scaling that up accelerates the problem rather than solving it. This is not a hypothetical failure mode; it has been the direct cause of several high-profile retreats from expensively contested markets.
Two cautions apply to LTV. First, it is a forecast, and forecasts made about a young cohort rest on assumptions about behaviour that has not happened yet. Early LTV estimates are routinely optimistic. Second, LTV must be calculated on contribution rather than revenue, because a customer generating substantial revenue while consuming equivalent bonus and servicing cost is not generating value.
Payback period, which is where businesses actually fail
LTV to CAC tells you whether acquisition is profitable eventually. Payback period tells you whether you will survive until eventually arrives.
Payback measures how long a cohort takes to generate enough contribution to repay what was spent acquiring it. An operator with excellent long-term LTV but a payback period of two years is consuming cash continuously, because it pays for acquisition today and receives the return across many months. Growth in that situation makes the cash position worse, not better, which is the counterintuitive trap that catches fast-scaling operators.
This is why payback is watched closely by anyone funding a gambling business, and why the acceptable payback period is one of the genuine strategic constraints on how fast an operator can grow. It also explains why market entry decisions are so consequential. Entering a new market means a period of heavy acquisition spend against a customer base that has not yet matured, and the operator must be able to fund that gap for as long as it lasts.
A worked example
Abstract definitions become considerably clearer applied to a single month. The figures below are illustrative and chosen for arithmetic clarity rather than drawn from any real operator.
An operator reports £120 million of casino turnover and £80 million of sportsbook turnover for the month.
The casino runs at a blended house edge of 4%, producing casino GGR of £4.8 million. The sportsbook holds 7% of turnover, producing sportsbook GGR of £5.6 million. Total GGR is £10.4 million.
Bonuses awarded during the month, net of the wagering they generated, cost £1.6 million. Gaming duty at 21% of GGR is £2.18 million. NGR, on a definition deducting both, is £6.62 million.
From that, supplier revenue share on casino content takes £1.1 million, payment processing costs £0.4 million, affiliate revenue share takes £1.3 million, and customer support attributable to the period costs £0.3 million. Contribution is therefore approximately £3.5 million.
Fixed costs, covering technology, compliance, staff, premises and licence fees, run at £2.6 million for the month. Operating profit is around £0.9 million.
Several things are worth noticing in that chain. The headline £200 million of combined turnover produced under £1 million of profit, a ratio that surprises people encountering it for the first time. Gaming duty was more than double the eventual operating profit. Affiliate revenue share alone consumed more than a third of contribution. And a single percentage point of movement in sportsbook hold, well within normal monthly variation, would have changed GGR by £800,000, which is close to the entire month's profit.
That last point is the practical reason sportsbook-led operators are volatile and why casino stability is valued so highly. It is also why small improvements in payment acceptance, bonus efficiency or affiliate terms matter disproportionately: each one falls almost entirely to the bottom line.
Data quality and the reporting layer
A caution before leaving the subject. Everything described in this lesson depends on the underlying data being correct, and in practice it frequently is not.
Common problems recur across operators. Attribution is genuinely difficult, because a customer may see a television advertisement, click an affiliate link, search for the brand and register through a paid search result, and every channel will claim them. Bonus cost accounting varies, and treating awarded value rather than net cost distorts every downstream metric. Currency handling across multi-market operations introduces reconciliation gaps. Player identity across brands and devices is often imperfect, so the same person may appear as several customers. And definitions drift between teams, so marketing's active player count and finance's active player count are calculated differently and never reconciled.
The practical discipline is to establish a single agreed definition for each metric, documented and applied consistently, and to be sceptical of any figure whose lineage cannot be traced. In a business where decisions about millions of pounds of acquisition spend rest on cohort contribution estimates, the quality of the measurement is not a technical detail. It is the decision.
Cohort analysis and what it exposes
The single most useful analytical technique in this industry is cohort analysis: grouping customers by the period in which they were acquired and tracking each group separately over time.
Its value lies in what aggregate reporting conceals. An operator's total revenue can rise steadily while the quality of the customers it acquires deteriorates every month, because new customers keep arriving in sufficient volume to mask the decline. Aggregate revenue looks like success. Cohort analysis reveals that each successive cohort deposits less, retains worse and repays its acquisition cost more slowly than the one before.
That pattern almost always has the same underlying cause. The operator has exhausted its most efficient acquisition channels and is progressively reaching further into less responsive audiences, paying more for customers who are worth less. It is a well-documented sequence, it precedes most acquisition-driven failures, and it is invisible in headline numbers.
Cohort analysis also answers questions that matter operationally. Does a particular welcome offer produce customers who stay, or customers who take the bonus and leave? Do customers acquired through affiliates retain better or worse than those acquired through paid media? Did the product change made in March improve retention for customers acquired after it? Each of these is answerable only by comparing groups acquired under different conditions.
Retention, churn and reactivation
Because acquisition is expensive, retention is where margin is defended.
Churn is the rate at which active customers become inactive. Defining it requires a judgement about how long a customer must be silent before they count as churned, and that definition should reflect actual behaviour: a monthly sports bettor who only plays during the football season is not churned in June.
Reactivation covers efforts to bring lapsed customers back, and it is generally far cheaper than fresh acquisition because the customer already has an account, has been verified, and has a known history. A well-run reactivation programme is one of the highest-return activities available to a CRM team.
Retention by cohort age is the underlying pattern to understand. Attrition is steepest immediately after acquisition, then flattens. The customers who survive the first weeks are disproportionately likely to survive much longer, which is why the early experience receives so much attention, and why friction in registration, verification or first deposit is treated as an emergency when it appears.
Reading an operator honestly
Bringing this together, a small set of questions will tell you more about a gambling business than any headline figure.
What proportion of revenue comes from regulated markets, and is that share rising or falling? What proportion comes from the top few percent of customers? What is the trend in cohort quality over the last twelve months? How long is the payback period, and is it lengthening? What is bonus cost as a proportion of GGR, and is it rising to sustain the same volume? And is growth coming from more customers, from more revenue per customer, or from entering new markets, because those three sources have very different durability.
An operator growing through improving revenue per retained customer in regulated markets is in a fundamentally stronger position than one growing through escalating acquisition spend in contested ones, even if both report identical revenue growth. Learning to see that difference is the practical purpose of everything in this lesson.
Key takeaways
- Turnover measures activity, GGR measures gross revenue, NGR measures what the business actually works with. Most disagreements about performance turn out to be disagreements about which figure is being quoted.
- NGR has no universal definition. Two operators can report the same underlying business with materially different NGR figures depending on what they deduct.
- Averages are close to useless in this industry because revenue is concentrated in a small minority of players. Segment before you conclude anything.
- The LTV to CAC relationship is the central test of whether a growth strategy is viable, and payback period is the test of whether it is survivable.
- Cohort analysis exposes deteriorating acquisition quality that headline revenue growth actively hides, which is why it is the first thing an experienced analyst asks for.