What I learned from a bad week
A week where almost nothing went right, and what stayed with me when the week ended.
I had a week in March that I will remember for a long time. I shipped a bug that confused twenty real users. I had a difficult conversation that went worse than I deserved. I broke a tool I depend on, on Wednesday, and could not fix it until Friday. By Friday evening I was tired in a way that took a weekend to put down.
I am not going to dress this up. It was a bad week.
What I noticed at the time#
I had a strong urge to make decisions on Thursday afternoon that should not be made on Thursday afternoons. Quitting decisions. Restarting decisions. The kind of decisions that feel inevitable inside a bad week and that are obviously wrong from the other side of a Saturday.
I made none of them. The single biggest practical thing I did all week was to refuse to decide anything large between Wednesday and Sunday. The refusal cost me a couple of unrelated micro-decisions. It saved me from two macro-decisions that the following Monday revealed as panic.
What stayed#
The product is fine. The user with the difficult conversation is now something close to an advocate, because I followed up properly. The tool is fixed. The week is six weeks behind me and the weekend after it has left no mark.
The lesson I keep is small: a bad week is bad data. Decisions made inside one are usually wrong. The discipline of waiting through the weekend has paid for itself, in this case, by an order of magnitude.
I have had two more bad weeks since. The same rule held both times.