Glossary

API Versioning

API versioning is the strategy for managing changes to a public API in ways that allow the SaaS product to evolve without breaking existing customer integrations. For Product Ops teams, API versioning policy has direct implications for customer trust and the support burden generated by deprecated endpoints.

?

What are the main API versioning strategies and their tradeoffs?

Three primary approaches exist. URI versioning (/v1/, /v2/) is the most common: clients explicitly specify the version in every request URL. Simple and highly visible, but routes the team must maintain in parallel grow with each major version. Header versioning (API-Version: 2024-01-01) keeps the URL clean and allows date-based versioning (Stripe's model), where each API version is named after a date and customers are upgraded to new defaults on their own account anniversary. Query parameter versioning (?version=2) is the simplest but often considered least clean. Most SaaS APIs use URI versioning for public APIs. Regardless of mechanism, the key commitments to customers are: how long will old versions be supported? (industry standard is 12–24 months after a new version is released); how are breaking changes defined and communicated? (a formal definition of "breaking change" prevents debates during deprecation); and what is the migration path? (provide clear migration guides with code examples).
?

How should SaaS companies handle API deprecation?

API deprecation is a customer trust event — done poorly, it generates significant negative sentiment even when the underlying product improvement is positive. Best practice for SaaS API deprecation: announce deprecation 12+ months in advance (minimally 6 months for low-usage endpoints); provide a concrete sunset date (ambiguous deprecation timelines create false security); send targeted notifications to every customer actively using the deprecated endpoint (identifiable from API logs — "Dear [Company], our records show you are calling /v1/users — this endpoint will sunset on [date]. Here's how to migrate:"); publish a detailed migration guide with before/after code examples; offer a compatibility mode or shimming period where the new version translates from the old schema; and create a deprecated endpoint usage dashboard visible to customers in their admin settings, showing their usage and the time remaining.
?

How do Product Ops and Support Ops mitigate support volume spikes during API deprecations?

API deprecations reliably generate support ticket surges — customers migrate at the last minute, integration issues surface, and some customers miss the deprecation entirely. Mitigation strategies: sunset in waves rather than all at once (disable for a randomly selected 5% of affected customers two weeks before the full sunset, so issues surface with a manageable volume before the full impact); staff up support for the deprecation window (schedule additional agent hours in the two weeks before and after the sunset date); pre-build a Zendesk macro with the migration guide and common error solutions, enabling rapid first-touch resolution by Tier 1 agents without engineering escalation; create a dedicated "API Migration" Zendesk view so these tickets are triaged by your most technically capable agents; and post a real-time status update on your developer status page for the sunset day, with monitoring for migration-related error spikes.

Knowledge Challenge

Mastered API Versioning? Now try to guess the related 5-letter word!

Type or use keyboard