Facebook ads stopped working in Glasgow and Dundee? 9 leads was a normal week

Sooner or later, every business owner who runs Facebook ads sends their marketer the same message: the ads have stopped working. We had one of those weeks on a glazing company's account running in Glasgow and Dundee, the kind that makes anyone fire off that message, and when we pulled the numbers the ads were doing exactly what they always do. What follows is the one reading that tells you whether a quiet week is a real problem or just a normal week lined up against the best week you ever had, which is what it nearly always turns out to be.

Most of what you will read about a bad week on Facebook sends you straight to the creative: change the picture, rewrite the words, throw more budget at it. The reading that actually matters happens before you touch any of it, and it takes about two minutes.

Why one week on its own tells you almost nothing

A single week is too small a thing to judge an ad account by. Leads never arrive evenly, some weeks are busy and some are slow, and the number on its own means nothing until you put it next to what the account normally does.

Here is the week that set the alarm off. The account brought in 9 leads at about £32 each, and the six weeks around it ran at a median of 9 leads at £28. So 9 leads was a completely ordinary week, sitting right on the account's own middle. These are the account's own weekly figures. The only reason the week stood out was the one before it: the best week the account had ever had, 19 leads at £14.57 each.

When you measure a normal week against the best week you have ever had, every normal week looks like a problem. Your baseline is the median, never the best week you ever ran.

What really changed that week in Glasgow and Dundee: the auction

There was one real change that week, and it had nothing to do with the ads themselves. The cost of the auction went up. Across both ad sets, the price of putting your ad in front of a thousand people, the CPM, rose 26% in a week.

That matters because it sits outside anyone's control. The auction gets more expensive when more businesses are bidding for the same screens, it happens to every advertiser at once, and no new picture or headline brings it back down. So a slightly dearer lead that week was the market moving, not the ad breaking. Knowing the difference is the whole job. A week where your leads cost a little more because the auction rose is noise. A run of several weeks where your leads thin out while the auction stays flat is a signal. The first you leave alone, the second you act on.

What to do with a week that looks bad

The instinct, when a week looks bad, is to do something: swap the creative, pause an ad, cut the budget, shuffle the money around. Most of the time that instinct is the very thing that breaks a campaign that was working fine.

Here is what we do instead, and it is three checks, not a redesign. We line the week up against the median of the last six to eight weeks, not against the best one. We look at whether the auction got more expensive across the whole account, which tells us whether the ground moved under us. And we leave the ads alone unless the trend over several weeks says otherwise, because three days is not a data set and a single week is barely more than one. A rule you set on a few days of numbers is a guess wearing a suit, and acting on it is how good ads get killed.

Changing the creative has its place, and showing a good ad to the wrong town is a real problem we have written about before. A quiet week is not that problem, and treating it like one is the fastest way to wreck something that was earning its keep.

How to check your own account in five minutes

You can do this on your own account today, and you do not need anything you do not already have.

• Pull your leads and your cost per lead for each of the last six to eight weeks, and find the middle one. That is your median, and it is the number every week should be judged against.

• Take the week you are worried about and set it next to that median, not next to your best week. If it is close to the median, nothing is wrong.

• Look at the cost of the auction, your CPM, over those same weeks. If it climbed the week your leads felt slow, the market got dearer and the ad did nothing wrong.

If whoever runs your ads cannot show you those three things without going away to look, they have not been looking either, and that is the real problem worth solving.

A slow week is the moment most money gets wasted on Facebook, not because the ads have stopped working, but because someone panicked and changed the thing that was working. If you are spending on Facebook ads and a quiet week has you ready to tear the whole campaign up, book a call and we will read the account with you, week by week, and tell you whether anything actually needs to change.