Give the greatest weight to useful coverage and the visitor’s ability to play it. The default weights deliberately exclude rights: the gates already handle them.
| Criterion | Weight | What good looks like | What bad looks like |
|---|---|---|---|
| A. Useful, eligible catalogue | 25 | Matches the buyer’s priority list; essential games are evidenced and tested; exclusions remain visible | Headline counts, duplicates and irrelevant titles obscure missing essentials |
| B. End-to-end launch and play reliability | 20 | Games become playable within agreed budgets across required environments; failures receive appropriate fallbacks | Successful page loads are reported as successful games; mobile or regional failures are hidden in averages |
| C. Integration and observability | 15 | Supported integration, reproducible deployment, meaningful events, secure credentials and predictable updates | Copied URLs, undocumented dependencies, exposed secrets and no reliable failure diagnosis |
| D. Metadata, freshness and version fidelity | 10 | Traceable fields, timely eligible releases and a clear relationship between demo configuration and review claims | Unknown configurations, stale dates and one unqualified RTP value copied everywhere |
| E. Privacy, security and presentation controls | 10 | Actual data flows and promotional behaviour match the approved design | Privacy text substitutes for network inspection; redirects or tracking surprise the publisher |
| F. Support and change management | 8 | Clear ownership, tested escalation and usable change notices | “Priority support” without defined service expectations or upstream escalation |
| G. Commercial transparency and total cost | 7 | Quote, usage definitions, internal work and exit costs fit a reproducible budget | Cheap access masks uncapped usage, manual work or costly changes |
| H. Continuity and exit readiness | 5 | Dependencies are mapped; disabling, exporting and replacing content has been rehearsed | Withdrawal would leave broken pages and an unknown migration task |
| Total | 100 |
Freeze weights before comparing offers. A multi-site group may reasonably increase continuity and integration weights; a release-led review publisher may emphasise coverage and freshness. Record the changes and keep the total at 100. Never reweight to rescue a preferred supplier after seeing the result.