Feature lists are unhelpful during procurement because every vendor ticks every box. What follows is deliberately short and specific — ten capabilities where the implementation detail determines whether the feature is real.
Each one is a control rather than a capability, which is the distinction that matters at scale.
1. Multi-provider payment routing
Not a list of supported providers. The ability to route each transaction by market, method, amount and current provider health, with cascading retry on recoverable failures.
Single-provider architecture puts a third party’s bad afternoon directly onto your revenue line.
2. Jurisdiction as a player attribute
Market rules — game availability, limit defaults, disclosures, age thresholds — applied per player at runtime, not per deployment or per brand.
Platforms requiring a separate instance per market make expansion arithmetic considerably worse than it looks.
3. Account-level responsible gaming enforcement
Deposit limits enforced across every payment route, not per method. Increases subject to a cooling-off period while decreases apply immediately. Self-exclusion propagating across all brands automatically, including to marketing suppression.
These specific details are the difference between a control and a checkbox.
4. Point-in-time audit reconstruction
The ability to answer what the system knew about a player on a given date — their limit, verification tier, applicable market rules and the promotional terms then in force.
Platforms storing only current configuration cannot answer the questions regulators actually ask.
5. Complete API coverage
Everything available through the administrative interface should be available programmatically. Where the API covers a subset, automation stops at the gap.
Webhooks with retry and delivery guarantees, rather than polling, and a versioning policy with a stated deprecation window.
6. Configurable withdrawal auto-approval
Rules the operator can define and adjust — verified account, known destination, amount consistent with history, no outstanding obligations — releasing the clear cases automatically.
Payout speed is bounded by manual review capacity in any platform without this.
7. Cohort-capable reporting
Retention and revenue by registration cohort, segmented by acquisition source, at matched ages.
Aggregate reporting cannot distinguish a growing business from a deteriorating product being outrun by acquisition spend.
8. Cross-brand identity resolution
For multi-brand operators, the ability to recognise the same person across brands despite different accounts and details.
Required for exclusion to work as intended, and for any honest view of player value.
9. Promotion simulation against historical data
Modelling what a promotion would have cost had it run against your actual transaction history, before it goes live.
Most expensive promotional mistakes are interaction effects that a specification review cannot catch and a simulation can.
10. Operational observability
Deposit success rate by provider and market, game provider availability, withdrawal queue depth and latency, all visible to the operator with configurable alerting.
Depending on a vendor to notice and report degradation means finding out after players do.
What is deliberately absent
Game count is not on this list. Beyond a threshold, engagement concentrates in a small proportion of titles, and library size is a weak predictor of performance compared with whether the studios your market actually plays are present.
Nor is any front-end feature. Presentation matters and is comparatively easy to change; the capabilities above are architectural and are not retrofitted cheaply.
How to test the list
Every item above can be verified in a sandbox rather than accepted on a slide.
Set a deposit limit and attempt to exceed it through a different payment method. Request a full player history export and time it. Configure a promotion and ask for the simulation output. Call an API endpoint for something you were shown in the admin panel.
Vendors positioning around comprehensive integrated stacks — PWP casino software platform among providers combining content, payments and analytics — should be able to demonstrate these directly rather than describing them. A capability that cannot be shown in a sandbox is a roadmap item wearing a feature’s name.
Ten items, all testable in an afternoon. That is a considerably better use of an evaluation than a hundred-row comparison matrix nobody will read again.
