[Peter Matteson]

Originally Posted by Jon Stratis

When we have an issue in production, my manager doesn't just want to know what happened. He wants reliable, reproducible testing AND a fix proving we identified the correct answer before he'll disseminate information to other parts of the organization. Saying you found the issue and coming back a week later and saying "just kidding" is about the worst thing an operations team can do.

We are the same way where I work. We review with a team of engineering fellows along the way to make sure we haven't gotten off track at each step during the DIVE process. (Define, Investigate, Verify, Ensure)

If you during your investigation that you don't verify the cause of the problem you can't ensure it won't reoccurr. I would want to make sure my data supported the conclusions before they were published, and preferably have a working solution.