Latest / Retail Media Breakfast Club / Trust in AI Shopping: Two New Studies
Transcript
- Trust in AI Shopping: Two New Studies === [00:00:00] Kiri Masters: I remember when Alex Bloomberg launched his podcast, StartUp, in 2014, and at that time, host-read ads on podcasts had been around for a couple of years, but Alex really [00:00:15] wanted to know exactly when an ad was coming on his show. He announced that he'd now be running ads in each episode, and he played a little music jingle that would come before each ad [00:00:30] so nobody could ever mistake a sponsor's pitch for Alex's organic endorsement. [00:00:36] At the time, it seemed to me like total overkill to go to such an effort to disclose an ad, [00:00:45] but it looks a little bit smarter now. That one move told listeners that ads pay for the show, that paid and unpaid content were separate, and that he had no plan to [00:01:00] dupe [00:01:00] anyone. And [00:01:02] that gave the audience something that they could check, and trust tends to follow when people can check. [00:01:10] Today, shoppers are applying the same test [00:01:15] to AI. In this episode, I'm gonna share some new research from Acosta Group and some highlights from my recent research in partnership with Bazaarvoice. The running theme here is trust and how, in some [00:01:30] cases, AI assistants are gaining trust while in others there's still a trust gap, And that gap is where there is opportunity for brands and retailers. [00:01:39] But before we jump in, I wanna let you know that I will be hosting a [00:01:45] webinar next week with Bazaarvoice CMO Doug Straaten to present some of the highlights from our recent research, which was called What Is a Brand Worth When AI Does the Shopping? We'll link up to it in the show notes for this [00:02:00] episode. [00:02:00] It's going to be short and sharp and actionable for brands of all sizes and categories, [00:02:06] So make sure you sign up to join us on October sixth. Now let's jump into the episode [00:02:12] [00:02:13] Kiri Masters: the first trend [00:02:15] across both of these research reports is that ads give shoppers the ick. So in Acosta's research, [00:02:23] and just for context, they have a very large shopper panel. They fielded this study in May [00:02:30] of this year amongst twelve hundred and seventy-one US shoppers. And they asked AI users how concerned they are that recommendations are influenced by ads or sponsorships. And On [00:02:45] a scale of one out of ten, ten being extremely concerned, the average was 6.6. [00:02:52] So quite concerned, I guess. And this foots with the findings from [00:03:00] my research with Bazaarvoice, which found that 54% of shoppers believe that a brand that they've seen advertised inside an AI tool is more likely to be recommended by that same tool, [00:03:15] even though AI companies say that paid and organic content is, are kept separate. [00:03:22] And that's really the crux of it, That the organic content and the paid content is separate, just like it [00:03:30] is with retail media, um, kind of. But people believe that they are linked and that there is some kind of thumb on the scale, even if that is not true And I don't think that the answer for AI [00:03:45] companies or retailers or brands for that matter is to stay out of chat assistance with ads. [00:03:50] I think the answer is really disclosure. Like I mentioned in that opening about the early days of podcast ads and the, this is an ad [00:04:00] music jingle was kind of invented as a signifier to say, "What you're about to hear is an ad. We want you to understand this is an ad." it doesn't really feel that way in AI chat right now, I think is what people are saying [00:04:15] in these survey responses. [00:04:16] So I think that the answer here is disclosure. Label the ad, explain how paid and organic content relate. Obviously, if you're a brand advertising in an AI chat assistant like ChatGPT, you can't... You don't get [00:04:30] to control how that disclosure is made. Maybe my call here is that going out to the retailers who are building AI chat assistants or upgrading them and introducing paid, um, paid placements inside of their retailer AI [00:04:45] assistants to really think about what that disclosure looks like. [00:04:47] And this was actually a comment in Acosta's [00:04:51] qualitative feedback from a Gen Z respondent that said [00:04:57] aI needs a disclaimer that they don't influence the [00:05:00] responses you get. That'd certainly be a step to building trust. And I don't think anyone can say it better than that. Um, there is some really good verbatim, um, qualitative comments in the Acosta survey. [00:05:15] We'll link up to it in the show notes as well, but it really shows how i- in general, people don't- aren't against ads. [00:05:23] I think they understand that ads are a necessity to get free services, but [00:05:30] they are concerned about disclosure. All right, trend number two: consumers trust but verify. Generative AI is a daily habit now for 44% of AI [00:05:45] users. [00:05:45] So anyone that has said that they are using AI, that they have used AI, that cohort, which is the majority of people now who've at least tried it, are now using it daily, and that is up from 37% [00:06:00] last year. It's up seven points from last year. And Acosta [00:06:06] Found that consumers extend AI a fair amount of trust already. Twenty-eight percent rate [00:06:15] AI's information as more trustworthy than other sources, 60% say about the same, 12% say less trustworthy than other sources. But trust doesn't end the research when you are [00:06:30] considering something to buy. In the Bazaarvoice study, we found that shoppers who receive an AI recommendation go on to visit the brand's website 68% of the time. [00:06:42] They go on to check social media [00:06:45] 65% of the time, and go look on a retailer's website 58% of the time. And we also split out questions by whether a brand was familiar or [00:07:00] unfamiliar to that user, and to what extent does the behavior change. So if a recommended brand from AI is unfamiliar, the share of users who do no additional research drops from 13% [00:07:15] to 6%. [00:07:18] Meaning there is a big difference between a known brand and an unknown brand, and the amount of time that people spend verifying that information or even doing [00:07:30] additional verification at all Insight number three here is the aisle gets a second opinion. Now, there was this terrifying concept back in the mid-2010s called [00:07:45] showrooming, and it was terrifying for retailers because it meant that shoppers would walk into a store and pick up their phone and bring up Amazon to check the product details, check the reviews, [00:08:00] and check the prices of stuff in the store. [00:08:03] And this was a real threat to retailers, and I think it ultimately spurred on a lot of e-commerce development at the time. ~So they~ [00:08:11] ~Just cut that last so they, um, I'll just move on to the next point. ~And Acosta found that thirty-nine percent of AI [00:08:15] users now use AI tools while shopping in a store, and that increases to fifty-five percent of Gen Z using AI tools while shopping in the store. What are they doing with these AI [00:08:30] tools? [00:08:31] Mostly they're checking prices or promotions. That's sixty-one percent of the time. Comparing products, that's forty-six percent of, uh, the features that they use. Finding an item's location in the store and [00:08:45] reading reviews in the aisle, forty-one percent. So this-- my point here that I wanna make is that a lot of retailers who have primarily in-store sales are a little slower on AI, a g- [00:09:00] AI game plan because they think, "Well, we sell mostly in-store." People aren't using our app, people aren't shopping online for our category or our price point or whatever. But this is the [00:09:15] cautionary message here is that they might not be actually transacting online, but they're more than likely actually checking AI in store for additional information. [00:09:27] They might still buy it in the store [00:09:29] [00:09:30] But they're certainly supplementing their research with information from an AI chat assistant. [00:09:36] So these shoppers are already testing the retailer's shelf against a chatbot [00:09:45] your advertisers can't wait weeks for audiences. GrowthLoop's Composable Commerce Media solution helps media teams turn first-party data into [00:10:00] high-value audiences, launch campaigns faster, and prove what's working across every channel. Learn how retail media leaders at Costco, Fanatics, and Gopuff use GrowthLoop [00:10:15] to create highly segmented audiences and deliver stronger results for their brand partners. [00:10:22] Visit go.growthloop.com/breakfast. That is [00:10:30] go.growthloop.com/breakfast And trend number four, getting out of our comfort zone, but with some guardrails. So one of the biggest use cases and things that shoppers [00:10:45] appreciate about AI is its ability to surface new products or brands that they hadn't considered or hadn't come across before. [00:10:54] But shoppers still tend to put more weight on the brands that they already [00:11:00] recognize. The Bazaar Voice research found that when AI offers several options to a customer, 65% click the brand that they already know and trust first. [00:11:13] We call this the first-click [00:11:15] advantage in that report But then looking into this research from Acosta, They found that respondents whose purchases had been influenced by AI, which was thirty-nine percent of the cohort, [00:11:30] forty percent of those respondents tried a brand that they had never bought before, and twenty-five percent switched from their usual brand. [00:11:41] So this usage of AI is really [00:11:45] harboring in a whole new era of consumers willing and able and interested in [00:11:53] switching brands. And that obviously represents a lot of [00:12:00] opportunity and risk for brands today, which we'll get into in just a second. But just wrapping up this comfort zone, piece. In this work that I did with Bazaarvoice, we had a hunch when we were [00:12:15] designing the study that this openness to familiar brands or unfamiliar brands rather, would be v- different by category, depending on whether it was a high-risk purchase or a highly considered purchase [00:12:30] versus something a little bit more pedestrian. [00:12:33] And that did in fact bear out in the data. So thirty-eight percent of shoppers are comfortable buying an unfamiliar brand that AI recommends in a low-risk category [00:12:41] Versus slightly less would buy [00:12:45] AI-recommended brands in all circumstances, and 17% would never do this. So category by category, we found a bit of a difference in both willingness to switch to an unfamiliar brand and the discount that would be [00:13:00] required to get someone to switch from a known brand to a dupe. [00:13:06] So we get into that category by category information in the report if you wanna dig a little further. So what next? What does this mean? So [00:13:15] if you're a challenger brand, this, as I mentioned, is a major opportunity to level the playing field and compete more on product attributes than with these really large incumbent [00:13:30] brands who have, you know, built decades and decades of brand equity and have a strong advantage on the shelf with brand and packaging. [00:13:40] Now is your chance. But you have to double down on proof. So [00:13:45] 76% of shoppers want reviews with photos or video before trusting an unfamiliar AI-recommended brand, and 75% want written reviews at the very least. Now, if you're an incumbent [00:14:00] brand and you're trying to fight off these challengers who and, and/or, um, might be offering a big discount because they are a dupe of your brand, then, you know, you're safe for now. [00:14:14] You [00:14:15] have that first click lead, but that is a lead that is on loan. Shoppers say that they stay with a brand because the product performs. They believe that it performs. That's 45% of the reason why people [00:14:30] stick with a known brand. They think that it works better. Only 8% say that it's because of habit, which I think is maybe over-attributing habit, potentially. [00:14:42] But, um, you're safe for now, [00:14:45] but it is certainly on loan, and you also need to be doubling down on reviews, UGC, and other trust elements from real human beings. And finally, if you are the retailer, you are playing the [00:15:00] role of this validation arbiter [00:15:03] 58% of shoppers check a retailer's website after an AI recommendation, and many are specifically seeking out rich product content and [00:15:15] real human reviews. Acosta finds that in-store AI users are checking prices and reviews in your aisles. Where they buy comes down to product availability and [00:15:30] pricing in their preferred channel or retailer. [00:15:33] And this is interesting, and I think heartening for retailers who are concerned about a race to the bottom. Only twelve percent follow AI to the lowest price. Most people wanna continue buying in [00:15:45] their continued-- in their preferred channel or retailer, rather than actually switching to a new retailer with a lower price. [00:15:52] So that's great news there. Well, that's it for today. Don't miss my webinar next week with Bazaarvoice CMO Doug Straighton [00:16:00] on October sixth. We're gonna present some highlights from that research. It's short, it's punchy, it's for brands of all sizes and categories. [00:16:09] We'll link up to the registration in the show notes. Thanks for listening, and I'll catch you tomorrow [00:16:14]