I use plenty of examples from my own life in the book. After a while it can give the impression that every problem belongs to me. This one does not. My wife and I are opposites here.
My rule for most things is: if it is enough to achieve what I want, it is fine. She wants the best.
Early in our relationship, when we were living in the United States, we added a couple of days in Miami to a conference trip. I still had some work to do during the day. Then dinner time arrived and we had to choose somewhere to eat.
Out came her computer, with 27 tabs open.
She would show me the first restaurant. Cuban music, this kind of food, these other things. I would say fine, let's go.
No, wait. There was another one.
She would show me the second. Again I would say fine, let's go. By the third or fourth, this was starting to irritate her. She had spent hours doing the research, and I arrived apparently willing to accept anything.
Which was more or less true. She was looking for the best restaurant. I was looking for the first restaurant where I would not be poisoned and could eat reasonably well while spending time with her.
To be fair, some of the places she found were exceptional.
But there was something useful about dinner: eventually it was nine in the evening and we had to choose somewhere, or not eat. The research had an external stopping point.
A lot of work has no equivalent of that hungry person waiting to go out. You can keep looking for the perfect approach indefinitely.
Some information only arrives after the test
Take a marketing campaign. You can spend a long time trying to decide which version is the right one before anything reaches a customer. Yet part of the information you need will only arrive when you test it.
If nothing goes out, you do not get that information. More time spent imagining the result does not replace it.
Ask which information you cannot obtain without an external test. Then arrange a test with as little risk as possible and use what you learn in the next version. There is no need to optimise in your head around information you do not yet have.
The other useful constraint is a reasonable deadline by which something must be finished. Give the deadline some force. Saying to yourself that you will do it tomorrow may change nothing; involving other people can create a reason to finish that is harder to keep moving.
In Miami, dinner did that job for us.