Facebook AN traffic fraud warning

Around October and November last year, I found that the MAI (installation) of FB could easily reach the AN position, and these positions had retention but no subsequent behavior, or no retention at all. In recent days, I have found that some series have abnormal retention when looking at retention. When running the AEO retention target (yes, I put the retention at R1 and directly invested in AEO), I found that the retention was directly 70%. “Fuck, how can my product deserve 70% retention?” I quickly checked, haha, it was done quite well, with retention, registration (very good), and previous behaviors, but there were basically no key monetization-related behaviors and in-depth events.

The data only started to become strange in the past week. It was in the normal range before. Sure enough… I checked and it was AN. There is no need to delve into the reason. It is definitely AN cheating traffic. When running MAI before, AN cheating appeared. I didn’t think it was strange that there was 0 retention or 0 follow-up. After all, there was no goal set. This time, cheating appeared in AEO. After reflection, I realized that running retention was obviously too “careless”. After all, if you encounter a publisher who brushes the volume (I encountered it), if the retention is good, it will be a good traffic that meets the standard in FB’s eyes. Hurry up and add it! The key is that I estimate that not only these AEO series will fail, but it seems that it will bring all my ads into the ditch.
Since the target is R1, the R1 of AN is particularly good, so when I study similar groups next time, more exposure will be made to certain publishers of AN, including my other MAI series and other AEO series. I estimate that all my old ad series will be affected to some extent. Solution: Turn off all R1 AEO series, strictly screen the MAI series, and kill them immediately if you see an increase in the proportion of AN. Key: Change other AEO targets immediately, switch all to new targets, and turn the entire advertising learning direction back. If there is no shortage of volume, start immediately and kill the old series that may have problems. Considering that retention is too easy to be brushed out by machines, it is recommended that you don’t play this kind of trick in the future. (Google users running R1 have not found the pit yet, but 99% will not escape, sooner or later)