The Daily: How marketers are using GenAI, where we pay most attention to social media ads, and Chase's retail media network

On today's podcast episode, we discuss how marketers are using GenAI, how its affecting media plans, and whether GenAI’s usefulness for advertisers will be similar to the calculator’s effect on mathematicians. "In Other News," we talk about which platforms users are paying most attention to social media ads on and Chase bank using spending data to help advertisers target folks. Tune in to the discussion with our analysts Max Willens.

Subscribe to the “Behind the Numbers” podcast on Apple Podcasts, Spotify, Pandora, Stitcher, YouTube, Podbean or wherever you listen to podcasts. Follow us on Instagram

Episode Transcript:

Marcus Johnson:

This episode is made possible by Walmart Connect. Are you in the market to take your media strategy to the next level and make an impact with entirely too many customers? Walmart Connect capitalizes on the massive reach of America's number one retailer, Walmart. They can help you meaningfully connect with Walmart's over 130 million weekly online and in-store customers to find the right audience for your message. Using Walmart's proprietary customer purchase data, they can help you precisely target even niche audiences at scale. Take just a few seconds to head to walmartconnect.com to see how they can help you find the customers you want at the scale you need.

Max Willens:

If you have several generations of dos and don'ts in your brand guidelines or you employ multiple creative agencies, you're a lot less likely to let this technology just spit stuff out willy-nilly and deploy it.

Marcus Johnson:

Hey gang, it's Thursday, April 11th. Max and listeners, welcome to the Behind the Numbers Daily: an eMarketer podcast made possible by Walmart Connect. I'm Marcus. Today, I'm joined by one of our senior analysts who is based in Philadelphia. He covers everything digital advertising and media for us. It's Max Willens.

Max Willens:

Yo.

Marcus Johnson:

Hey, fella. Today we're talking about what happens as gen AI gets more and more acquainted with advertising, but we start, of course, with today's fact. That is, Max, that there's no such thing as a panther.

Max Willens:

I'm sorry, what?

Marcus Johnson:

Yeah, I know.

Max Willens:

Then what do you call this thing behind me? No.

Marcus Johnson:

I actually looked then.

Max Willens:

You're going to have to explain. You're going to have to explain this.

Marcus Johnson:

I didn't think there was a real one. Okay, just for the record, I thought maybe you had, what do you call it? A teddy, a stuffed animal. But no, so apparently there's no such thing as a panther. So there's a Sanskrit word applied to the tiger, which is called pandara, meaning whitish yellow. The Greeks borrowed it and adapted it to panthera, meaning all beasts. But scientifically speaking, all four of the larger species of big cats are panthers. So it says here, the lion is Panthera leo, the tiger Panthera tigris, the leopard Panthera pardis, and the jaguar Panthera onca. They're the only cats that can roar. The animals most people think of as panthers are in fact either black leopards in Africa or Asia, or black jaguars in South America.

Max Willens:

Whoa.

Marcus Johnson:

I blame The Jungle Book because what the hell is Bagheera?

Max Willens:

What indeed.

Marcus Johnson:

Just living a lie. Soul tone, anyway.

Max Willens:

That's crazy.

Marcus Johnson:

Yeah, I know. I still don't really believe it, even though it's true. Anyway, cheers. Real topic, gen AI and advertising. In today's episode, first in the league, we'll cover how gen AI is, and isn't changing the world of advertising. Then for another news, we'll discuss which social media platforms people pay most attention to the ads on and details on Chase Bank's new retail media network. We start, of course, with the lead. Max, we're talking about gen AI and how it's influencing, shaping, changing advertising. It may feel as though ChatGPT is everywhere and everyone is using it. However, 77% of Americans say they've never used it before, according to a February 2024 Pew Research survey. So most people, a lot of people still have never used it. However, we have some numbers on people who are using it. We estimate 30% of Americans will be classified as gen AI users today. That's folks who enter a prompt into a gen AI system at least once a month. So that's the general population. But Max, zooming in on market people, how many marketers are using the tech, do we know?

Max Willens:

Well, there are lots of different numbers that have been supplied by third party researchers that speak to this, but a number that really stood out for me when I was writing a report on this subject came from Winterberry Group, which worked on a bunch of research on gen AI and the creative production process. They found that more than six in 10 respondents are using it in at least some part of the creative production process. I want to make sure that our audience understands that the creative production process has a lot of steps in it. This could be anything from using gen AI to pad out a brief really quickly to making a bunch of alternate versions of static images, to making prototypes that you show a client and to secure their approval. And so even though the top line number of over 60% is really powerful and really impressive, it's important to understand that a lot of that usage comes pretty early in the production process. It's hard to sometimes identify the role of gen AI in the finished product once we see it in our feeds or on television, right?

Marcus Johnson:

What's interesting and what supports that as well is the share of folks using it for work tasks has gone from 12% to 20% in the past year, according to some more Pew Research numbers, making it now the number one reason that people are using it. So they tracked using it for work, using it to learn new things, and using it for entertainment. Work was in third place of those three, and is now in first place going up eight points from 12 to 20 in the past year. People using it to learn something new only climbed three points and entertainment just two. And that also, they broke that out by age across those activities, and it's the 18 to 29 year olds who are actually driving up the most for everything, but in particular the most for work. So they seem to be driving that trend.

Max Willens:

Well, it makes sense, right? I mean, makes younger people in more junior roles, they get handed the most grunt work. You know what gen AI is great for? Grunt work. Gen AI is not something that you can set and forget and just take its results and input them into the fields, but if you need something to generate lots of texts that you can then edit down or tweak as needed and you need that to do it quickly, gen AI is great for that. I think it's early traction as a workplace technology in and around rudimentary tasks is, it's not surprising at all.

Marcus Johnson:

Yeah, that's a good point. You write, Max, in some of the research that you have on this, budgets are also growing, so more people using it. Budgets growing. Three in four CMOs and executives worldwide expect to have budget set aside for gen AI initiatives by midsummer according to Capgemini. It's been a year and a half max since ChatGPT hit the streets. What do we know about gen AI's leading use cases for marketers today?

Max Willens:

What's really stood out to me in the research that I did is that the uses are different from what people originally anticipated. Mediaocean did some research where they asked a bunch of marketers, and I think it was April of last year where they saw the most promise for gen AI and what they did, and then they asked them, I believe it was at the beginning of the fourth quarter how they are currently using it. There were some pretty striking differences.

To answer your question most directly, the things that they're using it for most commonly at the moment are for market research and data analysis. About four in 10 use it for data analysis, about a third use for market research. The third most common use case was copywriting, about a quarter say that they use it for that. But what's so interesting to return to my earlier point is that among the April respondents, nearly 60% of the respondents said that they thought copywriting would be where it had the most promise.

But I think, again, to return to the point that I made a moment ago, because of its tendency to hallucinate, because of a bunch of unresolved questions around copyright infringement, because of I think lingering and real questions about brand suitability and brand guidelines, there's just a lot of friction in place between any advertiser or agency that might want to just have ChatGPT write all the copy and what is actually happening. But for some sort of earlier in the production processes such as market research or data analysis, and there again, it's work that you have to check, they're finding that it's much more powerful and much more capable for that stuff.

Marcus Johnson:

Yeah. So what share of advertisers are taking the training wheels off and letting gen AI generate creative digital assets for them?

Max Willens:

I would say that this was something that I hoped to get at definitively when I started writing this report, and it's difficult to say definitively. And the reason that I say that is Google and Meta who are, as everyone listening knows, the two biggest advertising companies in the United States by considerable margin have gen AI oriented features inside their ad products. So PMax, PerformanceMAX, and ASC are both tools within Google and Meta respectively, and they offer gen AI flavored features within their ad components where you can spin up variations on creative or make ads out of whole cloth.

I could not find any statistics that spoke to the prevalence of advertisers using those. But I think, again, to point to the Winterberry stats, more than six in 10 are using it for components, but it's hard to say what share are using it and basically taking what a gen AI spits out and using it as the finished product. I will say that there is a notable discrepancy between the approaches of small, independent, maybe D2C type brands and larger more heritage and tradition bound enterprise or legacy brands. If you have several generations of dos and don'ts in your brand guidelines, or you employ multiple creative agencies, you're a lot less likely to let this technology just spit stuff out willy-nilly and deploy it.

Marcus Johnson:

You write that many of the marketing and ad professionals that you interviewed for your report compared gen AI's usefulness for advertisers with the calculator's effect on mathematicians saying it's a simple tool that accomplishes tedious tasks quickly and frees up time for more exciting work. Do you agree with this analogy that a lot of folks seem to throw back at you?

Max Willens:

Yeah, I was surprised that there were so many people offering up similar comparisons when I spoke to them, but I think it makes a lot of sense. I mean, as I was saying earlier in this conversation, there are lots of parts of the advertising production and planning and execution process that are really tedious, where maybe you come up with a tagline or a graphic and that's going to be the centerpiece of the campaign, but then you've got to break it into 15 different mobile ad sizes, or you've got to translate it into Spanish, or you've got to change the locations alluded to in the ads that they're geographically suitable. All that stuff is not difficult, but it's tedious and it takes forever. But if you can just give a computer that task and say, "Do this for me and I'll check your work later," that frees up a lot of time.

And to me, that aligns very cleanly with, imagine looking at a piece of a worksheet that you were given in elementary school where you're asked to divide 83,476 by 7241.3. You can do that, you can definitely do it, but it's going to be a big pain in the butt where you're drawing lines and doing a bunch of arithmetic and filling out this space of white paper underneath the two numbers. Or you can pull out a calculator and in less than three seconds you have the answer. So I think it's a very elegant way to describe the effect that this is having on media.

Marcus Johnson:

Yeah, speaking the effect it's having on media. You write that gen AI technology is not yet affecting media plans. We're going to play slice of pie where you're going to give me up to three reasons why it's not yet affecting media plans, and attribute some of the shares to those reasons.

Max Willens:

I think there's probably three reasons that this is not happening, and in laying them out, I think you and the audience will see that it's still I think really quite possible that it will eventually start having this effect but I do think that it's a ways away. So right now, the promise of gen AI is such that we could maybe ultimately get to a place where the digital dream state of advertising is possible, where the right message reaches the right consumer on the right platform at the right time.

There's a fantasy world version of this where a gen AI system that is operated by a brand and has access to all the assets and knows exactly what kind of message it wants to send to people can be plugged into a gen AI product that is owned by a large advertising platform that has all the targeting capabilities and is capable of versioning and optimizing things in such a way that every ad that is sent to every single person that is targeted by those two parties is customized, either based on what that person has consumed before, where that person is located, what that person has purchased before, and in so doing potentially the effectiveness of the advertising goes way up. And that's nirvana for marketers. But right now, the cost associated with that, the technological capabilities associated with that and the kind of thinking around that are not there. So the much shorter answer to your question is I would say that unit economics prevalence and ROI have not yet been worked out. I would give all three of those equally sized lessons.

Marcus Johnson:

Okay. Zooming out, just to close out the lead, you had a really interesting number. Nearly 90% of CMOs and executives worldwide expected that gen AI would play at least some role in everything from asset creation to brand metric tracking within the next two to three years according to Capgemini. And then according to that same research, 40 to 45% of CMOs and executives said that they would use gen AI for over half of everything they did in two to three years. Customer service, SEO, and video generation campaign creation, writing emails, blogs, scripts, developing new products.

I thought that was shocking, to be honest, because a lot of that stuff may not be here quite yet, but two to three years is very soon indeed, basically around the time of what would that be is that we just had, I always get confused with the, I guess the next World Cup. Basically before the next men's or around the time of the next men's and Women's World Cup 2026 and 2027 respectively, which seems very close, very soon indeed to be doing half of everything or to have half of everything infused with gen AI. There's even more in Max's report, including what advertisers can do to stay ahead of gen AI. The full report is called Gen AI in Advertising - How Gen AI is Shaping Advertising and Brands' Approach to It. You can click the link in the show notes or head to emarketer.com for the whole thing.

All right, folks, time for the fourth quarter of the show today, in other news. On which platforms are users paying the most attention to social media ads? And Chase uses spending data to help advertisers target folks. Story one, on which platforms are users paying the most attention to social media ads? YouTube is outperforming all other social platforms in social media add attention according to our industry KPI data via Adelaide. Our briefings writer Daniel Konstantinovic, explains that Adelaide ranked several social platforms by their social attention metric indicating the likelihood of an ad placement to capture attention and generate impactful outcomes across social platforms. YouTube was way out in front. In a distant second was X, formerly Twitter, followed by Instagram, Pinterest, TikTok, and then Snapchat. Max, your biggest takeaway from this research was blank.

Max Willens:

That upper funnel metrics don't matter as much on social media as they might in other parts of media. You mentioned that YouTube comes out way ahead with respect to other social platforms and the platform that came in dead last and that research was Facebook. Facebook is going to generate 34 billion in US revenue this year, and YouTube will do a little over 8 billion. I'm someone who thinks attention metrics are super, super important, and super, super interesting because a lot of the other metrics that are used as proxy for attention are easy to be easily gained, are inefficient, are flattening to content in a way that I think is really bad for high quality content producers. But at the end of the day, brands and advertisers spend on Meta because Meta gets the results, and that's important to remember when you think about the broader context of the mission of something like the attention metric.

Marcus Johnson:

Yeah, interesting to see podcast ads holding their own outperforming all platforms except YouTube throughout 2023, which is a positive finding for the space as it deals with a fair amount of consolidation, but interesting to see them up there.

Story two, Chase Bank will start letting advertisers target its credit card customers with tailored offers based on spending data writes Katie Dayton of the Wall Street Journal. Chase Media Solutions, as it's called, will differ from most retail media networks in that the bank will only charge merchants when a sale is made. But Max, your biggest takeaway from this was blank.

Max Willens:

I guess the takeaway that I have is that we should get ready for a lot more chatter about solutions like this. So Chase sits on a pretty rich trove of customer data. I mean, I think we could paint a pretty detailed picture of what a person is like, or what a person is interested in based on their credit card transactions, but there are lots of other kinds of companies that sit on data that could be of interest to other advertisers that are either endemic or adjacent to their categories. And so because of how powerful and how quickly retail media has grown, I think you're going to see lots and lots of brands start to at least think about how they can leverage the purchase and location and other first party data that they have with their customers to try to spin up a media business at same time.

Marcus Johnson:

Yeah, they've got a customer base of 80 million.

Max Willens:

Good place to start.

Marcus Johnson:

Yeah, not bad. Everyone getting in on the action. It seems getting in on the action/huge and growing pot of money, we estimate the retail media in the US is $55 billion business today, accounting for one out of every seven ad dollars. One in every seven go to retail media in the US today. In two years time, will account for one in five. That's all we've got time for this episode. Thank you so much to my guest, thank you to Max.

Max Willens:

Always a pleasure, Marcus. Thank you.

Marcus Johnson:

Yes, sir. Thank you, thank you. Thank you to Victoria who edits the show, Stuart who runs the team, and Sophie does our social media. Thanks to everyone for listening in. We hope to see you tomorrow for the Behind the Numbers Weekly Listen: an eMarketer video podcast made possible by Walmart Connect.