Start with the application, not the product.
Bring the same requirements to each provider. A useful proposal describes what recovers, where it recovers, who does the work, what gets tested, and what you pay under different scenarios.
Your provider conversation checklist
Check items as you cover them. Progress stays in this browser.
Ask for a complete cost scenario.
Request costs for normal operation, scheduled drills, a short failover, and several weeks in the recovery environment. Clarify capacity reservations, minimum commitments, licensing, storage growth, traffic charges, professional services, and exit costs.
Make testing part of the agreement.
Define what constitutes a successful test. Include application access, data validation, business-owner sign-off, recorded results, and an owner for fixing problems. Revisit the runbook when applications or dependencies change.
Take evidence into the decision.
Use recovery test results and a written responsibility matrix to compare the final candidates. Vendor marketing can explain capabilities; the proposed scope and your own tests establish whether the service meets your needs.
Further reading: Microsoft’s application recovery plan and testing guidance ↗.