Two weeks unsupervised

Bahrain is off today and tomorrow for its national day, and from about the seventeenth a good part of the region's marketing goes quiet until January. The offices close. The accounts they manage do not.

For years I believed the daily hand on the account was the job. Bids checked in the morning, budgets nudged at noon, creative rotated on a Thursday. An account left alone for two weeks was an account at risk, and I said so to a great many people.

Then I spent time on the platform side, where a great many small accounts are visible at once, and watched them go through more than one shutdown untouched. Nobody logging in, no pacing edits, no new creative, for a fortnight.

A share of them came out worse. That part I expected.

A share of them came out better. Steadier delivery, a lower cost per order, a learning phase that had finally been allowed to finish. Not a majority, but not a rounding error either, and enough that I had to sit with it.

The honest reading is uncomfortable. Some of what I had called management was noise. The daily edit was not protecting the account from the market; it was protecting the manager from the feeling of doing nothing, and the account was paying for the feeling.

Which makes the shutdown something else. It is the only controlled experiment most teams will ever run on themselves. Two weeks of no intervention, against fifty weeks of intervention, on the same account with the same budget. The result is waiting in the reporting when everyone comes back.

Nobody reads it. The first week of January is the week everyone changes things, because two weeks of not touching the account felt like negligence, and the fastest way to stop feeling negligent is to touch something.

The experiment runs every year. The result is thrown away every year, by the people it was about.