18 software release fails to avoid
I have seen all of these things happen multiple times over the last 25 years and they just make everyone very sad because they're so obviously dumb, and yet so easy to do. Please DO NOT do these things:
1. Deploying the wrong versions of software to the test environment
2. Testing against the wrong environment
3. Not recording which environment you are testing against
4. Not recording the date and time you did the testing, so nobody can find anything in the logs later
5. Not recording what versions you tested against, so nobody is sure what is broken and what is fixed
6. Not recording the actual output you saw, so the devs are mystified
7. Not working out what output you really need, so your tests are pointless
8. Not communicating the nature of a defect correctly to the developers, so they waste loads of time
9. Not tracking whether defects were actually fixed or not, so you either waste time or release defects
10. Not retesting something after a fix, so it is still broken
11. Not retesting other possibly related things after a fix, so you get regressions
12. Releasing the wrong version
13. Not having planned the test data you require, so your test system is full of random junk and you have no idea if the results are right or not
14. Not having any tooling to verify results, so you just give up
15. Not having any tooling for loading test data, so you don't bother
16. Not planning sufficient cycles, so you run out of time
17. Not correctly determining the severity of defects, so you fix the wrong things
18. Releasing software with high severity defects anyway and never really being able to explain why afterwards, so everyone hates you
