Financial diligence on a betting business is a solved problem. Revenue, hold, bonus cost, customer acquisition cost, cohort retention — the models are standard and the advisors are good at them. What is not solved is the question sitting underneath all of those numbers: is the product and trading capability producing them real, or is it borrowed?
That question is answerable, but not from the data room alone.
Margin level tells you almost nothing
Every information memorandum leads with gross margin. It is the wrong number to anchor on, because two operators with identical margins can be in completely different commercial health.
Margin can come from pricing skill — a trading operation that prices sharply, manages risk actively and holds a defensible edge. Or it can come from customer weakness — a book heavy in recreational accumulator play, where the margin is a function of what customers chose to bet rather than what the operator did.
The second kind evaporates. It evaporates when a sharper competitor enters the market, when a regulator caps stake sizes, or when the affiliate mix that delivered those customers changes. You are not buying a capability, you are buying a moment.
Separating the two takes market-level margin analysis over at least two seasons, split by product and by customer segment. It is not a difficult exercise. It is simply not one most financial advisors are equipped to run, so it usually does not get run.
Automation percentage is the number to ask for
If you have one question for a target's trading director, make it this: what percentage of your markets are priced and managed without human intervention, and what does that number look like by sport and by tier?
The answer predicts more about the asset than almost anything in the model:
- Scalability. A book at thirty percent automation cannot enter three new markets without tripling a cost line that does not scale.
- Margin durability. Manual trading is inconsistent trading. Inconsistency is where sharp money lives.
- Key-person risk. Low automation means the edge is in people's heads, and heads leave after a transaction.
- Integration cost. Two low-automation books do not combine into one efficient book. They combine into a larger inefficient one.
Targets are rarely asked this directly, and the number is frequently softer than the pitch implies.
Product maturity is not feature count
Feature comparison grids are the most common product diligence artefact and among the least useful. Every operator has cash-out. Every operator has a bet builder. The tick in the box says nothing about whether it works, how much of the catalogue it covers, or what it costs to run.
Better questions:
- Which markets does the bet builder actually cover, and what is the correlation model behind it?
- What is the settlement latency on in-play, and what is the void and resettlement rate?
- How much of the same-game parlay catalogue is priced by model versus by manual override?
- What percentage of live markets are suspended at any given moment during a typical match?
That last one is a fine proxy for overall trading quality, and it is measurable from the outside before you sign anything.
The technology answer changes with scale
Platform diligence usually produces a binary: proprietary is good, third-party is a dependency. Both readings are too simple.
A proprietary platform is an asset when it carries genuine differentiation — usually pricing and risk. It is a liability when it is a full-stack rebuild of commodity infrastructure that a supplier maintains better and cheaper. Plenty of operators have spent years rebuilding payment orchestration and account management, then run out of engineering capacity for the thing that actually differentiates them.
The useful diligence question is not what do they own but what have they chosen to own, and does that choice match where their margin comes from?
Value is concentrated in people nobody named
In most sportsbook acquisitions, a large share of the operating capability sits with twenty or thirty individuals: the traders who tune the models, the engineers who understand the risk engine, the product people who know why each override exists.
Almost none of them appear in the information memorandum. Retention arrangements are usually scoped around the executive team, which is the layer least involved in producing the margin.
A short structured exercise — mapping which capabilities depend on which individuals, and what a twelve-month exit would cost in each case — costs very little and repeatedly changes deal structure.
What a better process looks like
Sportsbook diligence works when it runs in parallel with financial diligence rather than after it, and when the people doing it have operated a book rather than modelled one. In practice:
- Margin quality analysis by market and segment, over multiple seasons.
- Automation and trading capability assessment, with the numbers verified rather than reported.
- Product depth testing against live traffic, not against a feature grid.
- Platform assessment weighted to where the margin actually comes from.
- Key-person dependency mapping, at the layer below the executive team.
None of this is expensive relative to transaction size. It routinely changes the price, and occasionally changes the answer.