Sportsbook margins are under constant pressure, but smarter use of data is changing the game. Gareth Crook, SVP of Sports at Pragmatic Play, Jeevan Jeyaratnam, Chief Betting Officer at Abelson Sports, Thomas Holland, Vice President of Product at Genius Sports Group, and Sabin Brook, Chief Revenue Officer at Bettormetrics, will explore how suppliers can help operators optimise margins without compromising the player experience.
SBC News: What are the most effective approaches right now for optimising sportsbook margins using data?
Gareth Crook: From a platform provider’s perspective, the most effective approach starts with access to high-integrity data. At Pragmatic Play Sports, we lead with an official data-first strategy to reduce latency, improve in-play uptime, and enable accurate pricing; especially across volatile sports and markets. This significantly limits the margin erosion that can come from latent feeds or data delays.
Beyond data integrity, we’ve built models that calculate expected margin using benchmarks like the closing price or next-available in-play price. These comparisons help identify patterns in user behaviour and highlight margin leakage at the user, market, or sport level. We then feed this into the trading layer, enabling granular configuration at scale. Some of the best operators I have worked with also implement liability-driven price adjustments, a model we support, where small odds shifts are triggered based on exposure and incoming bet quality. This mirrors broader industry trends, moving away from static pricing to a more fluid, responsive approach.
Jeevan Jeyaratnam: I’d say that more granular customer profiling, combating bonus abuse and multi-accounting are the most impactful uses of data for the operator, as those practices all damage margin. With increased confidence in profiling, and understanding the nuances within the user-base, you can, back-end allowing (a big caveat), start to make micro adjustments to clients’ maximum stakes dependent on vertical, sport or even bet type.
The increased interest in player markets and the different approaches to modelling and pricing these markets means that there can be fairly wide variation in prices for the same selection. It is, therefore, important, regardless of internal pricing confidence, to understand where the rest of the market positions itself. There’s little point being two points bigger than the rest of the market, even if you feel that price is correct. You’ll still lay the bet but at a greater perceived EV.
Thomas Holland: The most effective trading strategies hinge on deep data intelligence and precision. It’s not just about individual trades anymore, with millions of market-types and in-play bets to manage. At Genius, we’ve developed smart automated technology to drive fixture level pricing optimisation, based on each operator’s unique liability and bet data across related outcomes, for example match result and first goalscorer. We use advanced mathematical models and simulations to understand correlated probabilities and forecast millions of possible scenarios. This allows us to price sharper and take a truly profit-maximising approach, rather than simply reducing exposure.
When it comes to in-play, suspension settings are also critical. Our low-latency data and proprietary AI models consistently deliver maximum betting availability for our customers. For example, our football product records 99% uptime, by simulating the game with speed and accuracy to recalculate odds, with zero suspension time, during goals and VAR decisions in real-time. On popular leagues like the Premier League, this is resulting in more betting activity, more in-play turnover and higher retention rates.
Sabin Brook: The overall margin that operators can achieve is ultimately a composition of their entire trading strategy. The primary weapon in a trading director’s arsenal is the overround, which is effectively a measure of how much a market is expected to yield with balanced action on all selections of the book, and similarly, what the trading desk expects a market to yield in the long run even with imbalanced action. However, there is an assumption here that the prices they trade at are reflective of perfect probabilities of the underlying events occurring, which is ultimately an unknown quantity. With the breadth of offerings in modern sportsbooks, inevitably, things fall between the cracks, and poor pricing is prevalent. Thus, the trading desk resorts to alternative defensive strategies to minimise the amount of action taken on these poor prices.
The way sportsbooks work is that they assign every bettor to a cohort of similar players, based almost entirely on how profitable, or indeed unprofitable, that player is expected to be in the long run. The cohort that the player ends up in then inevitably affects their entire experience on that sportsbook, from how much they are permitted to stake on different markets to how long it takes for a bet to be accepted. That process is referred to as player management or player profiling. While this is the most effective strategy to reduce the amount of unprofitable business taken by a sportsbook, it ultimately results in less business being taken overall, as, more often than otherwise, bettors will be restricted in their wagering, which drives them to competitors. This isn’t a good thing for either the bettor or the sportsbook, and Bettormetrics believes there is a better way.
SBCN: How can sportsbooks personalise these margin strategies across different user segments without damaging player retention?
JJ: Attempts to profile large swathes of a client book into three or four categories can be harmful for retention. Granular data and the distinct edges it can identify mean that it should be increasingly possible to segment the user base into a detailed jigsaw. Depending on back-end flexibility, it should be possible to apply not only different maximum stakes, but also different margin application approaches based on client history.
TH: No sportsbook is the same. What’s alarming is that many operators don’t have true visibility of their liabilities on a fixture level. Our systems are custom-built for individual operators, allowing them to take full control or delegate to us depending on their preferences.
Firstly, our automated pricing optimisation tool, Edge, grows operator margins by driving odds movements based on their own, unique real-time bet and liability data. Edge works on a fixture and market level. In a world of growing fixture coverage and complex bet builder outcomes, it’s already having a major impact. Then from a player profiling standpoint, we proactively analyse user segments to avoid noticing sharp money when it’s too late.
But the real differentiator lies in the data. By automating decision-making based on real-time user data, our customers no longer need to spend time manually rejecting bets. Operators can confidently grow profits at scale and deliver a zero-friction betting experience at the same time, free from bet delays, limits and suspension. It’s true profit maximisation instead of the defensive risk mindset of old. We believe it’s the way forwards.
SB: The reality is that by the very nature of profiling, you are not going to be able to retain every bettor. Regardless of the strategies employed, discouraging unprofitable business relies on restricting selected bettors’ ability to wager on different parts of the sportsbook. The opposing argument is that losing those individuals is therefore not a major concern, as the operator is then certain that retaining their business would ultimately only damage their margins. The key, then, is to know that that certainty (in the player’s potential to damage margins) is warranted. At Bettormetrics we question whether sportsbooks actually have the tools at their disposal to get this right.
GC: Personalisation is only effective when it’s underpinned by solid user-level modelling. As a platform, we enable operators to estimate predicted P&L per user and apply tailored strategies, from stake limits and risk tagging to in-play delay configurations, without making the experience punitive. For example, operators can offer seamless experiences to recreational users while applying intelligent constraints to those who demonstrate sharp behaviour. Our tools allow in-play delay to be dynamically adjusted at player, competition, or market level, which gives a fine balance between protecting margin and maintaining engagement. We also support integrations for operators running customer segmentation, where tailored margin settings and bespoke user journeys can be deployed based on loyalty or volume; all without disrupting the front-end experience. Quiet, behind-the-scenes margin control keeps retention intact while supporting commercial goals.
SBCN: In volatile markets and across major events, how should data be used to protect margins in real time?
TH: When you think about the sheer volume of fixtures and market-types available, data and automation are crucial. Manual trading simply can’t keep up with the complexity and speed of this environment, which sees millions of data points sent across trading systems. For the first time, our automated Edge solution allows sportsbooks to instantly consolidate and respond to critical information, liabilities, supremacy issues, incoming bets and even betting activity in interlinked markets, all across one or two unified systems.
Each operator’s pricing is optimised in real-time, to maximise margins on a market-level, in a way that was never previously possible. Simultaneously, we can run real-time reports against user activity to optimise the end-user experience and ensure profit protection. Legacy systems often can’t handle modern features like same-game parlay liabilities, so this is a huge blind spot. Edge solves this problem. Without that visibility, margin optimisation is impossible.
SB: While there is more data available, it is speed and accuracy which is highly prized, particularly in the in-play category. Nonetheless, operators have access to more data than perhaps they are aware of, and being able to react to market forces in an efficient way is what ultimately differentiates your super sharp books – a very short list – from others. Sportsbooks cannot react to market forces if they lack the necessary observability of the market. While everyone understands the importance of dynamic liability monitoring, strong price referencing and the subsequent ability to react to these forces, in many instances, human interaction is still required to enact meaningful adjustments to their prices.
GC: Volatility brings both opportunity and risk; the platform’s role is to provide the infrastructure for real-time control. We prioritise official live data to ensure operators are not vulnerable to latency exploitation; particularly during fast-cycle markets like tennis, darts, or snooker. Our system supports automated, liability-aware pricing and real-time alerting on sharp activity or unexpected bet velocity spikes. This lets our managed services teams respond proactively, whether through pricing logic, suspensions, or adjusted delays, without waiting for human review. We also enable syndicate detection and clustering logic to flag correlated bets early. The combination of data-driven automation and human trader insight is key; our platform gives operators the tools to act faster, with better information, and fewer blind spots.
JJ: Volatile event data suggests in-play markets and one of the key data points is around customer advantage. Are there a group of customers who are accessing faster pictures or information and beating the sportsbook that way? Or are the in-play margins healthy and perhaps a reduced bet delay window and/or more market uptime would be a preferable next step. Perhaps, analysing the data granularity, there is a mixture of both in the sample. Can you increase a bet delay for one segment while reducing it for others?
Major events with high levels of liquidity and volatile events both feature strong reasoning for a less conservative approach to trading. Reducing bet limits and increasing uptime are ways to attract more volume and provided there is trust in the overall pricing approach the more volume the better, especially if you are able to weed out any courtsiders.
SBCN: What role will automation and AI play in next-gen margin optimisation? Is human decision-making still critical?
SB: The unique strength of human involvement on the trading desk lies in the ability to make informed decisions despite missing data, or to rapidly incorporate additional supporting information in real time. As AI and automation continue to evolve, they will undoubtedly play a larger role in trading operations. In time, we may reach a point where humans act primarily as stewards of AI-driven systems. However, to enable that future, the data powering these systems must be more robust, granular, and accessible than ever before.
Bettormetrics provides a next-generation performance and observability platform that ingests, visualises and aggregates big data into actionable insights for sportsbook operators.
GC: Automation already powers much of modern sportsbook trading; the future lies in refining and scaling it. On our platform, AI models surface expected margin across tens of thousands of markets monthly, enabling traders to focus on strategic interventions rather than repetitive micro-decisions. That said, automation is only as effective as the logic behind it. Human input is still critical especially when dealing with betting integrity alerts, injury or team news sensitivity, or risk appetite adjustments that require judgment. The best outcomes come from collaboration between smart tech and experienced trading teams. From our perspective, it’s not about replacing traders, it’s about elevating them. Operators using platforms that prioritise automation, modelling, and segmentation will continue to outperform in margin optimisation while maintaining scale and service quality.
JJ: Automotive responses to real-time cues are already changing the way sports are traded. Using computer vision in real-time allows deep analysis beyond the capability of human detection. Being able to identify and predict likely behaviours in real-time will enable sharper pricing and improved margins. Where AI solutions are, perhaps, still lacking, is in nuance, the human element as to why a price should be artificially biased based on external factors that models find difficult to interpret. This is where human oversight is essential. At Abelson Sports we believe that we have an optimal mix of automation and human input.
TH: Automation and AI are the future of sportsbook margin optimisation. AI will help at every stage, from writing code and recognising betting patterns to generating accurate models for event outcomes. Machine learning continuously improves these models, helping us make more informed decisions faster. However, human oversight remains vital, just in a different capacity. We’re moving from humans being involved in the process to them monitoring that the systems are functioning correctly. It’s about scale: AI replicates what a human trader would do, but at speeds and volumes that are simply not feasible manually. In short, AI enables better margin control, faster reaction times, and less operational drag, while still benefiting from strategic human guidance.
