Meta wants marketers to trust the machine.
Advantage+ Shopping Campaigns (ASC), Meta’s automated ad product, promises to take care of everything – budgeting, creative, targeting, optimization. Just hand over the keys.
But here’s a plot twist: Automation doesn’t always work better than manual campaign management.
This year, ad measurement startup Haus kicked off a large-scale study to compare the performance of Advantage+ with traditional manual campaign strategies – what Olivia Kory, chief strategy officer at Haus, jokingly referred to as “Boomer buying.”
“Is it worth continuing to resist or is it really time to just fully embrace these advertising tools?” said Kory, referring to ASC and Google’s PMax on stage at the Marketecture Live event in New York City earlier this week. “At the end of all this automation, are brands actually better off?”
Putting automation to the test
Both good questions.
To answer them, Haus ran a study examining 640 incrementality tests over an 18-month period across a wide spectrum of its client base in different verticals, from mid-market and DTC brands to large enterprises, including The Home Depot.
The average brand in the study spent just over $1 million per month on Meta, equating to roughly $14 million in annual spend.
Haus used geo-lift testing as its primary measurement approach, which assesses campaign effectiveness by turning off ads in specific regions and comparing aggregate sales data to see whether marketing actually moved the needle. The analysis considered both direct site sales and retail outcomes, such as on Amazon and offline.
So does Advantage+ work?
Yes, no, maybe, and it depends.
Plus (or minus?)
When you zoom out, Kory said, Meta’s impact is “undeniable.”
“Lift is mostly a function of spend,” she said, and the majority of marketers Haus works with spend a heck of a lot of their ad budgets on Meta, which accounts for 77 of the top highest-lift experiments ever run on Haus. On average, Haus sees Meta driving almost 20% lift to a brand’s primary KPI, including new customer revenue.
“Brands are highly dependent on Meta,” Kory said. “What this means is that if they turned off Meta, their business-level revenue would go down by 20% overnight.”
But digging in a little deeper, the data tells a more nuanced story about Advantage+, which, based on Haus’ examination, didn’t consistently outdo manual methods.
Haus found that Advantage+ only outperformed manual campaigns in 42% of tests and, when it did, the efficiency gains weren’t dramatic. Advantage+ delivered 12% lower incremental return on ad spend at 18% lower daily spend.
“This really surprised me – it was confusing for me,” Kory quipped. “I went through the seven stages as a marketer.”
Advantage+ also drove 17% less lift during the post-test observation window, which is the period of time after an incrementality test is over when Haus continued to monitor for any additional conversions or sales that could still plausibly be attributed to the campaign. Doing that helps capture any lagging effects, like purchases that occur days or even weeks after someone sees an ad.
‘Too good’?
So what the heck is going on? One of the main selling points of automation is that it performs really well.
One theory, Kory said, is that Meta’s algorithm may actually be “too good” at zeroing in on likely buyers.
“Is it so good at finding purchases, it’s actually targeting people who are already going to buy?” she said.
And there’s another nuance worth unpacking: Meta tends to deliver results much faster than other channels, often during a campaign itself. Its impact shows up almost immediately, Kory said, whereas platforms like YouTube often produce more gradual results over time.
In other words, Meta “hits hard quickly,” Kory said, and – rather surprisingly – actually underestimates its impact.
“We all think Meta tends to take too much credit,” she said. “But if you’re only looking at click-based attribution, they’re actually underreporting by about 15% on average.”
Beyond the soundbite
When Haus published its research earlier this year, a lot of people took it as an indictment of Advantage+ and evidence that manual campaigns perform better, full stop.
“Everybody started grabbing the headline of, like, ‘ASC doesn’t work. I knew it! Don’t use it,’” Kory said.
But that shouldn’t be the takeaway, she added.
Although manual beat automation 58% of the time, Advantage+ performed for 42% of brands in the Haus study. “So it’s close,” Kory said.
It’s also worth noting that Haus’ findings don’t account for recent personalization enhancements to Andromeda, Meta’s AI-powered recommendation model for ads, which is something Haus is considering studying down the line.
“It’s possible that, with the Andromeda updates, things have gotten much better,” Kory said.
The point is, brands should be experimenting with this stuff for themselves. Every brand is a little different depending on how long it’s been spending on Meta, the consideration cycle for its products and its price point.
“Please, test for your own business,” Kory said. “Meta consistently delivers, but brands really need to be testing this trade-off between automation and incrementality.”
