Introducing the Adweek Podcast Network. Access infinite inspiration in your pocket on everything from career advice and creativity to metaverse marketing and more. Browse all podcasts.
When The Hershey Company used AI to drive its media placements in a recent campaign, it not only sold more candy but also gave the confectionary brand an advantage in its negotiations with retailers.
Hershey’s partnered with Chalice Custom Algorithms for a Halloween campaign, a make-or-break period for candy brands, that incorporated its sales data. The brand served more ads in real time to markets where candy wasn’t selling and fewer ads to locations where Hershey’s was already flying off shelves.
The strategy sold candy so effectively that the brand’s sales team used it as leverage during negotiations with retailers, said Vinny Rinaldi, Hershey’s head of media and analytics at the Possible conference in Miami Beach.
“This is a little bit of a pivot away from ‘you have to buy retail media,'” Rinaldi said. “It’s a way to say ‘Hey, we’re actually supporting your stores.’”
Retail media has complicated the relationship between brands and the retailers that sell their goods. Even though interest in retail media is surging, consumer packaged goods (CPG) companies often feel pressure to buy ads from retailers to ensure their products show up on shelves.
Retailers will remain key partners in Hershey’s media strategy, which is powered by their sales data, Rinaldi said. Still, brands that build media plans that drive sales can show they don’t need to buy more retail media ads than necessary to drive good performance.
“You’ve got a weapon in the fight,” said Lou Paskalis, CEO of AJL Advisory and a longtime brand marketer, who moderated Hershey’s panel at Possible. A brand could buy retail media, “or [a brand] could actually use media dollars to drive product out of the store. Which do you want? It’s a different conversation. It’s about leverage.”
How custom algorithms work
Programmatic buyers can use off-the-shelf algorithms from demand-side platforms, but these are often optimized around media goals like video completion rates, rather than business goals like sales, Rinaldi previously told ADWEEK.
With custom algorithms, brands can build bespoke media plans for particular objectives using their own data sets.
Media agency Brainlabs used custom algorithms to optimize for an additional conversion points along the customer journey for a travel client and found a 120% increase in flight bookings, ADWEEK previously reported.
But custom algorithms can’t fix all the problems of programmatic buying, like how easy it can be to buy low-quality media on made-for-advertising sites, which are environments filled with low-quality content to lure readers in and buffet them with ads. Rinaldi said Hershey’s also employs strategies to ensure it’s only using custom algorithms to bid across premium inventory.
“We’ve got an incredibly tight site list and supply path strategy. We’re pretty much 100% MFA-free on The Trade Desk,” Rinaldi said. “Our site list makes sure all of those things are still applied regardless of changing the way that we buy inventory.”
Custom algorithms can help give marketers control over machines, and not the reverse, said Tylynn Pettrey, vp of data science at Chalice.
“It’s AI-driven by the marketers,” she said.
This article has been updated to clarify Lou Paskalis’ quote as: A brand could buy retail media, “or [a brand] could actually use media dollars to drive product out of the store.”
