Meta is updating its ad platform to make its AI smarter—educated on advertisers’ data—and launching a new way for marketers to tell if the ad machine is making the right moves for their campaigns. Today, Meta announced a string of upgrades to the ad system, including a way for advertisers to connect their Google Analytics data and other sources of “ground truth” into their ad accounts on Meta to see if the numbers add up.

Meta, which runs Facebook, Instagram and WhatsApp, has been talking more about how its ad manager, a self-serve platform used by millions of marketers, is becoming more automated lately, as it relies more on AI to manage campaigns. Today’s announcements showed that Meta wants to improve how its AI finds valuable customers and drives conversions, or sales. Also, it will give marketers a way to compare Meta’s data to their third-party analytics to understand better what is working.

“We’re kind of in the first step towards a long-term vision of being able to provide optimization that’s tuned for every advertiser,” said Fred Leach, Meta’s VP of product in an interview with Ad Age. “We’re at very basic step one, where we’re saying, ‘now the system has some idea of how Meta performs relative to everything else,’ and we’re updating it to take that into account, so that we can deliver more value.”

It’s been two years since Meta launched its AI-powered ad product called Advantage+, which was part of a growing lineup of automated ad campaigns from all the major digital ad platforms. Google has a similar product called Performance Max, or PMax, and TikTok has Shopping Ads. These are ad systems that use machines to set campaign goals, handle bidding in auctions, choose the best-performing copy, and find new customers. Meta and the others are putting generative AI into the ad creation, so more and more of the ad art is done by machines. Reliance on AI is supposed to make these ad platforms more efficient, calculating the best possible return on ad investment. But AI also is changing the nature of advertisers’ work, and there are concerns about whether people can trust AI.

Also read: Inside DDB’s AI marketing tools

Value proposition

To that end: “We are rolling out a few new ways for businesses to customize their campaigns by more precisely communicating what they value,” Meta said about its ad updates in a blog post today. One of the updates lets marketers educate the AI ad platform about what types of “conversion” goals they are going after. The marketer tells Meta’s ad manager which customers are the most valuable, and “they can define a higher bid for just these customers without having to create a separate campaign, enabling greater automation and performance,” Meta said.

“What we’re talking about today is a lot of different ways for advertisers to think about the value or to communicate to us how they value the outcomes that we're able to deliver,” Leach said. “And feeding that into our automation systems is going to just make it much more efficient and powerful.”

Advertisers are just starting to grapple with all the ways AI will reorient the digital marketing landscape. Meta has been at the forefront of this change, seeing automation as a salve for some lingering issues in the ad industry. For one, automation helped improve performance after it was degraded by data lockdowns online. With concerns about sharing consumer information online, and the slow decline of online trackers such as cookies, AI has been seen as a useful way to make smart decisions with limited data.

Meta’s executives, including CEO Mark Zuckerberg, recently talked about the automation revolution during a second-quarter earnings call with Wall Street analysts; Meta reported $38.3 billion in ad revenue for the period, a 22 percent increase year over year. Ad impressions and the cost for the impressions are rising, too, thanks to improved AI, according to Meta. Ad impressions across the family of Meta apps, and the cost per thousand impressions, were both up 10% year over year. The average cost per impression was down 16% year over year.

“AI is also going to significantly evolve our services for advertisers,” Zuckerberg said in the latest earnings call, adding “it used to be that advertisers came to us with a specific audience they wanted to reach, like a certain age group, geography, or interests.”

“But today advertisers still need to develop creative themselves. In the coming years, AI will be able to generate creative for advertisers as well and will also be able to personalize it as people see it,” Zuckerberg said. “Over the long term, advertisers will basically just be able to tell us a business objective and a budget, and we’re going to go do the rest for them. We’re going to get there incrementally over time, but I think this is going to be a very big deal.”

Ground truth

Today’s announcements were steps in that direction. Perhaps the most significant update was the integration of Meta’s ad manager with the third-party data, starting with Google Analytics and Northbeam. Meta also plans to include Adobe Analytics and Triple Lift as third-party data sources. These are analytics services that marketers already work with to get a fuller picture around all their media spend across platforms. By tying this data to their Meta ad manager, that could improve Meta’s reporting about which campaigns worked and how well they worked. The data also could help Meta’s ad system become smarter for future campaigns.

For advertisers, the analytics are key for improving measurement on Meta. For instance, Javy Coffee, which spends about $3 million a month on Meta ads, uses Northbeam as its analytics provider, and it is more accurate than the data from Meta’s ad platform, according to Matt Faber, head of growth at Javy Coffee, a direct-to-consumer brand. Reporting within Meta can be “quite inaccurate” when it comes to metrics around customer acquisition costs and ad measurements, Faber said. Now, that Northbeam’s data is plugged into the mix, it’s easier to adjust campaign setting on Meta to get better results, Faber said.

“Media buying is all decision-making,” Faber said. “So if you’re putting money behind, you know, things that are not actually working, then you're not going to have any success. So, when you look at Facebook's reporting and compared to Northbeam, it's a night and day difference.”

The analytic improvements are meant to enable to advertisers to use their preferred source of “ground truth,” Meta’s Leach said.