Thirty Days Is the Wrong Window

Thirty days is usually too short a window to judge marketing, and it’s the window almost everybody uses.

Why the month became the default

Part of it is inherited rhythm. The ad platforms report monthly, the invoices go out monthly, and the month becomes the unit without anyone choosing it. It’s a convenient unit. It just doesn’t match how customers decide or how a business’s year is shaped.

What a month can’t tell you

If your average customer takes three weeks to make up their mind, and your busy season looks nothing like your slow one, then a month of data is telling you about timing at least as much as it’s telling you about performance.

That cuts both ways. A channel can get cut in week three that would have paid for itself by week ten. Another can look brilliant for no better reason than it happened to run during the stretch when the phone always rings anyway. In both cases the decision gets made on the calendar, not on the channel.

Seasonality works the same way. If your busiest months fall in the summer, a spring ad test and a July ad test aren’t comparable, even when the ad is identical.

A longer memory

What fixes this is a longer memory. Once you have last year’s numbers broken out by month, you can tell the difference between a channel that’s working and a season that’s working. At that point a thirty-day read becomes useful, because you’re comparing it against something instead of reading it cold.

Building the baseline

It isn’t hard to build. Pull last year’s leads and booked jobs by month, put the spend by channel next to them, and keep it in one sheet. If lead sources weren’t written down along the way, the baseline will be partial, and that’s fine. A partial baseline still shows the shape of the year, and it gets better every month you keep recording. It also helps to add a column for anything unusual that year, such as a price change, a bad-weather stretch, or a week the crew was out, so the baseline carries its own footnotes.

Most of the small businesses I talk to don’t have one, and the reason is rarely difficulty. Nobody ever sat down and did it.

I don’t have a universal answer for how long a window should be. It depends on how long your customers take to decide, and that varies a lot between a same-week emergency repair and a project people plan for months.

How long does your average customer take to go from first contact to a booked job, and does your reporting window reflect that?