Marketing Briefing: How eBay is experimenting with generative AI for more ad personalization now

This Marketing Briefing covers the latest in marketing for Digiday+ members and is distributed over email every Tuesday at 10 a.m. ET. More from the series →

Much of the chatter about the impact of generative AI on advertising tends to focus on the potential future possibilities rather than current applications. Over the last six months, eBay’s in-house AI team has been figuring out how the company can leverage generative AI to help with ad personalization.

“A lot of where we’ve focused our AI and personalization activities in the past has been in predictive modeling, really understanding more of what you want, when you might want it, how you might want it,” said Adrian Fung, eBay’s global CMO. “But then there’s always limitations, right? We might even be able to get it to very granular data. But how do we actually produce content that is appealing to you based on what you want?”

Early experimentation with generative AI ad personalization for eBay has seen the company creating more banner ads, more personalized suggestions and subject lines for emails for its owned and earned channels as well as working with YouTube on different video edits for its paid ads. The challenge for eBay is not only producing content that will appeal to its customers on a personal level but also managing the roughly two billion items on its site with content written for those items by the item’s sellers and finding ways to make that more appealing.

“We can create multiple versions of that banner and start optimizing and testing which ones might work the best,” said Fung, adding that the hope is to have hundreds of banner ads that it can test and optimize to someone personally. “Right now, we are going through and getting from less than 10 at a time into dozens at a time. The aspiration is to get even higher than that.”

The company’s core in-house AI team is working with its product and marketing teams to define the capabilities and find ways to scale the use of AI for those teams. “We’re not at the stage where we’re comfortable letting the AI go out without human supervision,” said Fung. “What we’re building is platforms where the gen AI creates multiple variations quickly.”

What would’ve taken a person weeks to develop creative now takes only days for the teams to test and use, Fung added. Much of this current testing with generative AI personalization is focused on the company’s owned and earned channels. For example, rather than having marketing banners on the homepage scheduled based on topics that will become relevant like holiday or “broad generic topical stuff” the company is working to have banners that may be more “inviting” and “personalized,” explained Fung, based on someone’s interests.

This is an example of clients taking a “crawl, walk, run” approach to AI that other marketers will likely mimic, said Nick Miaritis, chief client officer at VaynerMedia.

“I’m seeing other Fortune 100 other companies having the same conversation, which is rather than continuing to craft subject lines for CRM to [a few] different audiences can we now craft thousands of them and then break the audience down into segments,” said Miaritis. “So that I think is a super simple phase that I would imagine over the course of the next 18 to 24 months, most organizations have that in a beta or pilot.”

While some marketers like Coca-Cola have been touting their use of generative AI, others like Under Armour have recently seen criticism of their use of generative AI. How generative AI is used, its purpose, how previous assets are used and what will come as well as the ethical and moral questions will continue to swirl as marketers figure out how to approach the technology.

3 Questions with Catherine Ayers, ASICS’ senior director of marketing

The marketing and advertising landscape has changed dramatically. How is ASICS adapting?

What we recognize is with all that ever-changing pieces to marketing, and obviously strategy, there’s a certain piece that we recognize that we need to continue and stay true — true to the ASICS brand and how we were founded. At the end of the day, that’s connecting to communities and making sure that we are uplifting and supporting those communities.

With that making sure that, while we’re uplifting those communities, that we’re connecting to the right partners so that we are really supporting individuals throughout. For us, on that formula [of] how we connect, we have ASICS brand ambassadors that [spread across a] plethora of different areas in parts of the U.S. to key media industry partners. Our athletes are a big voice for us, especially Olympic year.

For us, really taking it to that ground level community-centric formula is something that we will continue to focus on regardless of, to your point, some of these other things that are changing around us.

How does a community-driven formula help broaden awareness and create a brand position in the U.S.?

When we talk about those types of ambassadors, we also bring in our athletes [like Tenia Fisher and Asia Rawls] to be a part of that conversation. So we bring our, almost in a sense, our ASICS community all together.

A lot of the time, that is anchored in key moments. In these moments, how do we bring some key athletes with us? How do we make sure you know, we are supporting our ambassadors within their communities at those times and really connecting all those dots together, if you will?

What do the influencer marketing efforts look like?

We’ve stayed with micro [influencers]. There are some aspects of our business, specifically in our sport style side, that we partnered with partners like Stefon Diggs, which again, you could say macro [influencer]. But it’s hard … because it’s not like we really do big contracts with anyone. That’s not in our DNA. It’s really focused on again, who the person is, do they love the brand? Do they kind of embody that mind, sound, body essence? — Kimeko McCoy

By the numbers

Streaming continues to be the go-to way to watch television with 73% of adults turning to streaming services before cable and broadcast, according to research from Adtaxi, digital marketing agency. To keep up with changes in the way people watch television, advertisers are increasingly adding connected television to their media mixes. (A look at how the ad-supported streaming war will shake out here.) See more key findings from Adtaxi’s report below:

When choosing how to watch TV and video content, the majority (72%) watch through their TV set.

On average, streamers who respond to ads do so in two or more ways. Over half (51%) go directly to a company’s website, 40% look at the company’s social media, and 36% consider a company review.

When it comes to advertising on streaming services, a significant majority of respondents (two-thirds) have observed a rise in the volume of ads displayed while streaming. Of those respondents, half claim the increase in ads has a negative effect on their viewing experience. — Kimeko McCoy

Quote of the week

“Every campaign I got back from every retail media network was alway like ‘This is the best thing since sliced bread.’ Some were coming to us to say it was a control versus expose methodology. Others would tell us they were calculating based on a different control. There were very different levels of measurement. To this day, there’s no one set form or methodology to consider incrementality [across retail media networks.] A lot of what we had to do was vet them on how they measure incrementality.”

— Paras Shah, director of digital marketing at Georgia-Pacific, when asked about how the company has managed — and judged — the growth of retail media networks.