By Marty Cagan · svpg.com · LinkedIn
Cagan argues that shipping updates to an established user community requires care, because users dislike change and today's connected internet lets them punish bad releases quickly. He calls the disciplined approach to rolling out changes "gentle deployment" and notes that many PRD changes are needlessly incompatible or gratuitous. Since users adopt updates on different schedules, products, especially web services, often must run old and new versions side by side for a period. He lays out techniques including advance communication, heavier QA to avoid rollbacks, and risk-containing approaches such as opt-in parallel versions, regional rollouts, and incremental releases. The core message is to be highly sensitive to the disruption each change causes.
01Key takeaways
- Assume users dislike change and actively minimize the disruption each update causes.
- Question whether incompatible or gratuitous changes are truly needed before shipping them.
- For web services, plan a transition window where old and new versions coexist for users who can't switch immediately.
- Invest extra QA when reliability, scale or performance is uncertain, since rollbacks amplify community anger.
- Use opt-in, regional or incremental rollouts to contain risk for significant changes.
- Give clear advance notice before retiring the old version, and expect pushback from engineering and operations.
02Key sections
- Why updates need care
- Users resist change, and in the connected web era a poor release can trigger strong community backlash. Large consumer services must weigh community impact in every update.
- Unnecessary disruption
- Many proposed changes are needlessly incompatible or gratuitous, so teams should look for ways to reduce the disruption before shipping.
- Users move at different speeds
- Not everyone can adopt a new version at once, so web services often need a window where both old and new versions remain accessible.
- Techniques for gentle rollout
- Advance communication, stronger QA to avoid rollbacks, and risk-containing deployment methods help limit disruption.
- Parallel, regional and incremental deployment
- Opt-in parallel versions, phased default switching with notice, regional launches, and bite-size incremental changes each let teams contain risk, though supporting parallel versions is costly.
03From the post
“A partnership dedicated to teaching best practices to product teams and product leaders”
04Frameworks mentioned
Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.