How AI chat search could disrupt online advertising

Almost every service we use on the internet is basically a platform for advertising, especially search engines.

Advertisers pay to get their sites at the top of search results, have their businesses show up on digital maps or populate their products at the top of shopping carousel pages.

The search engine companies are not only paid, but get data about what users want, which they can then turn around and use to sell more advertising.

But how does all this work if, as chat-based artificial intelligence permeates web search, the results become less like a big list and more like a one-on-one conversation? That’s where it looks like Microsoft and Google are headed with their Bing and Bard chatbots.

Marketplace’s Meghan McCarty Carino spoke with Garrett Johnson, assistant professor of marketing at Boston University, about how this new approach could really shake up the online ad space.

The following is an edited transcript of their conversation.

Garrett Johnson: The chat interface is kind of like a Goldilocks approach. So, like, the traditional search results page with four sponsored and 10 organic links, I think, is going to become a thing of the past. But then the other extreme would be like a voice assistant, which is typically providing a single answer. So a chat-based search is more amenable to multiple options than voice search because of the relative speed of reading versus speaking. And we all know that chat-based search also has a fibbing problem. So search engines, users and advertisers want this information, but also like footnotes and links to make it easier for users to be able to check the source themselves.

Meghan McCarty Carino: When we think about traditional search versus the chatbot approach, it seems like you’re kind of going from multiple results to maybe one result or fewer results. How could that affect advertising?

Johnson: That’s one of the key things that could be disruptive here, is that it creates a “winner take most” dynamic. So, you know, you get fewer search results with chat search. And so that typically is going to benefit strong brands No. 1, and No. 2 is going to benefit the leaders in SEO or search engine optimization, those that [adapt] to the parts of SEO that are relevant to chat.

Garrett Johnson (Courtesy Boston University Questrom School of Business)

McCarty Carino: What are some maybe creative ways that you can imagine advertising getting worked into these systems that we haven’t seen yet?

Johnson: One thing that’s kind of conspicuously missing right now is shopping ads with images. So you know, text ads is a really clunky way to search if you’re shopping for a dress. I think in the future when we see chat-based search be better integrated with images and shopping ads, then chat-based search becomes a powerful tool for refining shopping searches. The second major thing that I see further down the road is thinking that chat-based search will eventually want to hand off to an advertiser chat AI. And this is analogous to voice apps with voice assistants. So you can order a pizza, you might see the chat-based search, like, hand off to Domino’s, whose chatbot would take you through an order. It’ll be interesting to see whether there is actually a hand-off to another website, or whether this actually stays on the search engine website. And then maybe instead, the search engine is trying to sell a branded experience like having the Geico gecko talk you through an insurance quote or having Domino’s logo and display ads for pizza while you order pizza. So that’s, I think, gonna be really interesting to see when that hand-off takes place. Like, does it stay on the Bing website? Or does it move to Domino’s?

McCarty Carino: Yeah, it seems like kind of a finer line to walk with chatbot advertising compared to traditional search that, you know, when you’re interacting with something almost like one on one, like a human, that it’s more jarring to have advertising jump into that than, say, when you’re scanning through a list of results.

Johnson: Yeah, and it’ll be interesting to see how that’s done in a way that feels a bit natural. We think [of] like word-of-mouth advertising. You know, if you asked me for a restaurant that I would recommend, I could think of a few different names. How would I frame that for somebody that’s actually paying me to say that? I think there’s ways to do it that make it transparent, but I agree it’s a little bit awkward.

McCarty Carino: How else could this model challenge advertising?

Johnson: I think another problem for advertisers is to be mindful of brand safety. Because, you know, in the past these AI chatbots kind of devolved into hateful messages. I think today they have more guardrails to limit this, however imperfect. But I think a big challenge for advertisers is going to be that they’re just going to have to leap in with both feet for competitive reasons. I mean, just like the search engines themselves, I think these chat-based searches are, are going to be a bit of learning in real time, and the search engines don’t want to be left behind and so they’re going to try to release these products into the market. For advertisers, they need to maintain the relevancy and stave off a potential land grab here by search engines. Advertisers don’t want consumers to be booking hotels and ordering pizza and buying car insurance without leaving the search engine because that would give way too much control to the search engines.

The symbiotic relationship between search engines and advertisers has been especially important for Google, but recently its parent company, Alphabet, reported its first drop in ad revenue since the pandemic began.

A big part of that is the recent economic slowdown, but also there’s pressure from new AI-powered competition.

A piece in The Conversation points out that reliance on advertising puts Google in a tougher position than its main competitor in the chatbot wars — Microsoft — which earns much of its revenue from hardware and software.

And if you want to get a sense of how advertising is being used already in Bing’s chatbot search engine, we tested it out with The Wall Street Journal’s Joanna Stern.

And I know what you’re thinking: Isn’t it kind of the pot calling the kettle black to be critiquing awkward transitions to advertising on a podcast, of all things?

Well, we wouldn’t know anything about that here.

link