Secure your place at the Digiday Publishing Summit in Vail, March 23-25
If OpenAI’s latest moves indicate anything, LLMs are positioning themselves to be the go-to stop for searching and shopping. Between shopping integrations and ad product rollouts, AI chatbots could rattle retail media networks’ hold on sponsored and search ad dollars.
As user behavior shifts, so too could retail media’s value proposition.
Already, AI has upended traditional search as users increasingly turn to chatbots. Traditional search engine volume is expected to drop 25% this year as search marketing loses market share to AI chatbots, according to Gartner predictions.
‘It’s a threat’
It’s not far-fetched to expect a similar impact on retailers’ $69.33 billion U.S. ad business. If other AI platforms take OpenAI and Google’s lead, rolling out sponsored ads, they could build super media offerings that compete with retailers.
“I think it’s a threat, but I don’t think it’s a threat to extinction. It’s a threat to how retail media networks operate as they are. They’re going to have to evolve,” Rita Steinberg, vp of media at FUSE Create, told Digiday.
For example, say someone is shopping for cookware. Instead of searching on Walmart, Target or Amazon’s owned sites, they turn to ChatGPT to research and maybe purchase with ChatGPT’s Instant Checkout. A decline in retailer site traffic directly impacts the RMNs’ ability to monetize its site.
“That’s a huge business that gets disrupted directly in correlation to traffic to retail sites. The more time [shoppers are] spending elsewhere they’re spending less time [on retail sites],” said Mike Feldman, svp of commerce at Flywheel.
Retail media search spend in the U.S. was expected to reach nearly $38 billion in 2025, according to eMarketer. Search ads make up about 60% of retail media spend, Feldman said.
Publishers are already feeling the effects. LLMs have upended referral publisher web traffic traffic and reduced click-through rates. In response, some publishers have enacted bot-blocking measures or struck content licensing deals to stop the bleeding.
Heart and headache
Media networks have been advertisers’ heart and headache. Companies like Diageo have started to look at retail media as a full-fledged media channel. Procter & Gamble is also betting its strategy on retail media. While spend continues, advertisers lament RMNs’ walled gardens and incrementality issues.
Still, advertisers aren’t ready to count retail media out in favor of AI chatbots, and dollars certainly have yet to move. “Of all the brands that I’m working with, no one is taking money away from retail media to put into AI ads,” said Ross Walker, Acadia’s director of retail media.
Brands like Williams-Sonoma and The Knot have started testing ads in OpenAI’s ChatGPT. However, questions linger around things like high costs, brand safety and limited data transparency. Marketers are cautious until ad-supported AI chatbots, including ChatGPT, Microsoft Copilot, Google AI mode and others, prove out effectiveness. Unlike retail media networks, LLMs lack the final point of sale, per marketers.
To truly compete, AI chatbots would need APIs, measurement tools, and transactional data, which are controlled by the retailers. Notably, a few deals here have already been struck. Walmart has inked partnerships with Google Gemini and OpenAI’s ChatGPT for AI-powered shopping experiences. Meanwhile, Target has partnered with OpenAI for its own specialized shopping experience within the ChatGPT app.
“That’s going to be the struggle for these LLMs that sit outside of this world right now,” said Preston Larson, chief executive officer of Modifly and chief media officer of Court Avenue. He later added, “Walmart controls that ecosystem, and they’re the ones that can tell where their dollar is going and is it paying it back for brands. And that’s where I see the power.”
RMNs have already run out of on-site digital shelf space, moving onto off-site offerings in social and streaming to retain ad dollars. And several major retailers have introduced generative AI-powered chatbots themselves, like Amazon’s Rufus or Walmart’s Sparky. As retailers expand their ecosystems, there’s potentially more added value in RMN’s closed-loop attribution, end-to-end measurement capabilities, execs say.
“The retailers, because of having that information, actually have more leverage than people think. And they have also paved the path where these LLMs are going to have to try to catch up,” Larson said.
More in Marketing
In graphic detail: How Anthropic’s Pentagon refusal is paying off in downloads, brand trust and enterprise deals
OpenAI’s Pentagon deal seemed to spark uproar among its users, many of whom were against it. Anthropic’s refusal to agree to the terms was seen by users as the more trustworthy alternative.
‘Brand safety is moving from fear to curiosity’: Zefr’s Raddon on content-level accreditation – and what it exposes about the industry
The threat is no longer a discrete piece of bad content that a keyword list or a domain block can catch. Its volume.
Meta’s measurement and attribution updates are welcomed, but not ground-breaking, ad execs say
Meta is rolling out the updates starting this month, in a bid to bring it more in line with third-party platforms. While advertisers always welcome more transparency, the changes alone won’t necessarily lead to major budget increases.