TL;DR
- When the Halfsies ads switched to gamers, the cost per click went from $0.35 to $0.96.
- The 35 cents was flattered by accidental taps. The clean baseline was 64 to 74 cents, so the real rise is about a third.
- Views cost almost the same: $11.62 per 1,000 before, $12.14 after. Gamers tapped less often.
- One ad went the other way. The Sheriff ad clicked at 3.13% and 44 cents with gamers.
- No changes for three days after a targeting switch. The first check is 19 October.
Halfsies is my daily cutting game for iPhone. On 16 October its Meta ads stopped going to any adult in the United States and started going to gamers aged 18 to 50 in six countries. The next morning the cost per click had nearly tripled. The fourth Jeje Studios ad diary has the whole day. This is how I read that number before deciding to do nothing about it.
Check what the old number contains
The campaign average before the switch was 35 cents a click. That number has day one inside it, and day one was mostly junk: 133 of 164 clicks came from Meta's Audience Network, where ads sit inside other apps and a lot of taps are accidents. I wrote about turning that off. With Audience Network gone, the broad audience cost 64 to 74 cents a click.
| Audience | Cost per click |
|---|---|
| Whole campaign, 11 to 15 October, junk taps included | $0.35 |
| Any US adult, Audience Network off | $0.64 to $0.74 |
| Gamers, 16 October | $0.96 |
Against the clean baseline, gamers cost about a third more per click. That is a real increase, and much smaller than the headline.
Views cost the same; taps did not
The obvious explanation is that gamers are expensive to reach because every game studio bids on them. The numbers say otherwise. Before the switch, 1,000 ad views cost $11.62. On the first gamer day they cost $12.14. Showing the ad barely changed in price.
What changed is how often people tapped. The click rate went from 3.33% across the first five days to 1.26%. Part of that drop is the junk taps leaving the average, and part may be a crowd that already sees a lot of game ads. One day can't separate the two.
One ad liked the new crowd
Meta put most of the money on the infomercial ad: $11.38 for 10 clicks at $1.14 each. The Sheriff warning got $4.43 and also brought 10 clicks, at 44 cents each and a 3.13% click rate. On day one that same ad looked like the worst of the six, because its clicks were the accidental ones. Shown only to gamers on Facebook and Instagram, it is now the best.
So the average hides two very different ads. If I had judged the switch on the average alone, I would have missed the one ad that got better.
Why I'm not touching it yet
A targeting change resets Meta's learning, and the first day is noisy. My rule for this test is to wait three days after a targeting change before pausing any ad, so the clicks have time to settle and the first day doesn't make the call.
The first check is 19 October. Until then the six ads run as they are, and the third diary has the audience they run on. The game they point to is Halfsies, free on iPhone.