Friday, 4:35pm.
This is apparently the natural habitat of the ambitious account change. The plan has been discussed all week, the final agreement arrives late, and suddenly everyone wants to leave the office with the new arrangement switched on.
I don't object to making a change on Friday by principle. I object to making a change with no record, no review point and nobody available if it behaves oddly.
There is a special kind of irritation in trying to reconstruct the old setup from memory. You are certain the setting was different. You are less certain how different. An unhelpfully confident person says, “Can't we just put it back?”
Put what back, exactly?
My small habit is to save a clear record before a substantial change. The relevant settings. The reason for the change. The part of the account affected. The page or offer involved. What would make us reconsider.
Not a ceremonial document with twelve headings. Enough to tell the next person what happened.
The way back needs thought. Some changes can be reversed neatly. Others involve data, a website release or a business arrangement that cannot simply be undone with one click. Pretending they are all easy rollbacks is another form of wishful thinking.
If we're changing a destination, I want the previous destination known and the new one checked. If we're regrouping something, I want the old boundaries recorded. If we're testing a message, I want to know what was shown before.
I also want a sensible review time. A quiet afternoon isn't always evidence that the change failed. A broken link is rather stronger evidence. The response should fit the problem.
The point is not to preserve the past forever. It is to make experimentation less dependent on somebody saying, “I'm pretty sure I remember.”
Confidence is a charming quality. A saved record is more useful at 5:12pm.
For a small account, this can be embarrassingly simple. A note, a captured configuration, a named person checking the result. You don't need a control room. You do need to avoid treating ordinary precautions as an insult to the cleverness of your idea.
I have made changes I was pleased with and changes I later reconsidered. In both cases, being able to explain the route helped.
There is no prize for sticking with a poor decision because the exit looked awkward. There is also no prize for reversing a sensible one after thirty nervous minutes.
Make the change understandable. Make the review possible. Know what a reversal would involve.
Then go home without leaving the next person a bloody puzzle.
A thought worth keeping?
