Everyone in marketing can see AI making ads easier to buy and easier to make. The awkward detail in plain sight is the quality of the information advertisers send back to the ad platforms after those ads run. If every brand can describe what it sells and let the same machine produce its campaigns, the advantage moves into the feedback that machine receives.
On Meta’s earnings call for the first quarter of 2025, chief executive Mark Zuckerberg set out that future bluntly. Any business would tell Meta the result it wants, such as a sale or a new customer, and how much it is willing to pay for it. The platform would handle the rest.
That sounds convenient until a category gets crowded. In supplements, many companies chase similar buyers with similar promises, similar budgets and the same ad products, so each asks the same system for scarce attention. Better buttons make the contest cleaner, faster and more crowded.
Inside that crowd, ad platforms aim at whatever they are told to value. If the value an advertiser reports back is shallow, delayed, inflated or disconnected from profit, the machine gets better at finding the wrong thing. If that value is closer to what the business wants, the platform has a better target.
Google’s help pages show the logic. One automated bidding setting spends a campaign budget to bring in as much value as possible from sales and other actions the advertiser counts. When the advertiser sets a target return on ad spend, Google says it must send the value of each individual transaction.
Meta is moving the same way. On the same call, it described a feature called Incremental Attribution, which aims at sales Meta believes would not have happened without an ad. Knowing which ad gets credit for a sale no longer satisfies advertisers, who want to know whether the ad caused it.
Answering that question takes an ongoing cycle, and running it becomes the marketer’s job. The marketer decides what outcome matters, tests which ads caused it and sends the platforms better value figures, then starts again. The aim is to teach the buying systems what profit looks like for that one business.
