I Killed Nine Ads Last Month. That’s the Good News.
Nine dead ads on one account were nine answers. Zero dead ads on another taught me nothing. The difference was where the budget sat, and whether I knew which setup I had picked.
Last month I killed nine ads on one account. Not in a panic. No disaster, no lesson learned the hard way. Just a Tuesday, coffee going cold, ads manager open, working down the list. Nine of them were costing more than I was willing to pay for a signup, so I switched them off and got on with my morning. Nine ads, dead. Sounds like nine mistakes. It isn’t. It’s nine answers.
Nine dead ads is not nine failures
We were testing concepts. Different angles, different promises, different ways in. Nine of them did not land. So now I know nine things that do not work on that audience, and I know it for certain, because I paid to find out. That is not waste. That is the price of the answer.
And here is the part people get backwards: the platforms are built to sort. Hand one ad to the algorithm and you have given it nothing to do. Hand it eight and it will go and find your two. Variety is the raw material. The nine I killed were not a mess I made, they were the pile the winners came out of.
Then there’s the other account
Same fortnight. Different account. I killed nothing. Not “a couple”. Zero. I did not switch off a single ad, and it did better. Eight ads, all running out of one shared budget. The platform put most of the money into the two that were working and quietly starved the rest. One of those ads got nine cents. Nine. It never got the chance to prove anything, one way or the other.
So that account was cheaper and it needed nothing from me. Sounds like the obvious winner. Except I still have no idea what those six ads could have done. They did not lose. They were never really tested. They got starved before they had a chance, and I saved my money and learned nothing.
The difference was not the ads
It was where the budget sat. One pot, and the algorithm decides. Fast, cheap, and it does not ask. It will also bury something good before you ever see it work. Many pots, and every ad has its own money to spend. Nothing dies until you kill it, so a loser will happily bleed for weeks if you are not looking. But you find out. You pay for the lesson and you own it.
Neither one is right. I run both, on purpose. The mistake is not picking wrong, it’s not knowing which one you picked.
It is more complicated than I have just made it sound. If you do this for a living you know the words already: ABO, where each ad set holds its own money, and CBO, where the campaign holds it and the platform spreads it around. And here is the bit that catches everyone, including me. Ads inside an ad set always compete, whichever setup you run. The platform crowns a favourite in each set and starves the others. So counting pots tells you who decides between your ad sets. Inside them, the algorithm is always holding the knife. That nine-cent ad was not starved by the campaign budget. It was starved by the ad sitting next to it.
Your one job this week
Open your ads manager and find out whether your budget sits on the campaign or on the ad set. Then count your pots. That number tells you who is killing your losing ads. If it’s more than one, the answer is you. And if you are not looking every week, the honest answer is nobody.
The week in brief
- Google is reining in campaigns that beat their targets: from 17 August, a Target CPA or ROAS campaign that has quietly been outperforming the number you set gets pulled back toward it. A tool for adjusting those targets appeared in accounts on 6 July. If a campaign is doing better than you told it to, go and tell it the truth before August does.
- Google moved some ads from pay-per-click to pay-per-view: since 15 July, Demand Gen campaigns on Discover with view-through conversions on are billed per thousand impressions, not per click. Same campaign, different meter. Check your pacing if that’s you.
- Your AI ads are about to be labelled as AI ads: Google is rolling out a “How this ad was made” panel. Automatic if you used Google’s own tools, self-declared if you didn’t, which tells you what the label is worth.
- Apple Maps ads are coming, and plumbers are banned: Apple’s Maps ads launch in the US and Canada this summer, one ad per search, and the policy locks out home services: plumbers, electricians, locksmiths, roofers, HVAC. If that’s your trade, a local channel is arriving without you.