This is a guest post I wrote for teambits.io.
The Problem
So we had a problem at work. Our test infrastructure was confusing, the tests took too long to run, and still only tested half the use cases. Someone needed to go in and rationalize the tests, so we could add more tests. So a meeting was called. Discussion points were sent.
The meeting was supposed to last an hour. Instead, it went on for three. Everyone had a favorite test they wanted to save. No one could agree on anything. The project manager took copious notes. All we could agree by the end of the 3 hour meeting was, that out of the 50 tests, only 3 could be possibly removed. I knew that was false, as more than half the tests were repeat of others. But it was impossible to reach a consensus.
The project manager made a big Excel file, and showed which test were a copy of the others. He sent it around. Two weeks later, no one had replied. Three months later, I went to see the project manager. “We are running out of time in the overnight tests. Did we reach a decision on which tests to throw away?”
He threw up his hands in frustration. “No one makes a decision. If I try to forge ahead, everyone starts throwing tantrums.”
That was a year ago. The manager has left. No decision has been made yet.
We were killed by the committee.