Shipping early and often doesn’t have to mean shipping a potentially un-good thing to everyone and poisoning the well
Disproof
Err, demonstrably not launching to everyone?
Consequences
You launch something untested and find out a whole bunch of ugly that you could have done with five hallway testers
It costs you an entire consumer cohort / demographic who’ve written you off as a bunch of incompetents (because you are)
Support gets overwhelmed with a spike across 100% of the user base, resolution times skyrocket and you look doubly bad
Hey, you might even get cancelled!
Causes
Pressure to move fast
Unwillingness to invest in tools/infrastructure to do sane things like a/b testing and feature flagging
Unwillingness to invest in tools/infrastructure to understand what’s working without picking up the phone and calling random customers
Planning by people who think you’ll magically get it right, so schedule a big public launch for the day the engineers submit their first pull request
When product leadership themselves are off fighting fires instead of managing and aligning expectations
Approaches
Usually a stakeholder education issue
If everyone knows the score and there’s still pressure to yeet it at everyone as fast as possible, just do the maths of failure again