StackAdapt is the leading AI advertising and orchestration platform marketers rely on. Built entirely in-house with an easy-to-use interface, StackAdapt unifies programmatic and owned channels—including CTV, DOOH, display, native, audio, email, and more—into one seamless experience.

Learn more.

Marcus Johnson: [00:00:00] Marketers don't need more dashboards, they need answers. StackAdapt is the AI advertising and orchestration platform built to deliver them. With IV Studio, their AI first advertising hub, uncover high-value audiences, spot performance gaps, and take action all in one place. Learn more at stackadapt.com/ivy-studio.

Hey, gang. It's Thursday, November 12. Welcome to Behind the Numbers, the e-market podcast made possible by StackAdapt. I'm Marcus, and today you'll be listening into a conversation between our principal AI analyst, Nate Elliott, regional VP of sales at StackAdapt, Katie Friday, and VP of monetization partnerships at OpenAI, Samantha Jacobson, as they discuss AI targeting capabilities.

Nate Elliott: I'm really glad to be here today with Katie Friday and Samantha Jacobson. Thanks for joining me.

Katy Friday: Thanks for having me, Nate. I'm excited to be here. Yeah, thank you so much.

Nate Elliott: I wanna start with, uh, perhaps the biggest [00:01:00] question around AI advertising and targeting right now. I mean, for years, for decades online, marketers have had access to various consumer behaviors or groups of consumer behaviors to understand, uh, which audiences are looking for what.

So whether that's using people's, uh, aggregated search data to figure out which groups of people are looking for certain products, whether it's looking at their browsing behavior to understand, uh, what they're browsing for. And of course, with more and more people using AI, that gives us a new way to categorize people and to understand which ads might make sense for which groups of users.

I'm really curious to hear from both of you. Uh, what is this new kind of data do to how marketers and advertisers understand and use this aggregated consumer behavior, and, and what does it do i- to the question of first-party data? Does it make it more important or less important?

Samantha Jacobson: Ultimately, marketers are using data in service of trying to [00:02:00] sell more of their product, right?

They're looking for particular outcomes from a campaign. You mentioned, Nate, some of the legacy ways of securing that data or trying to create proxies to drive purchase, but ultimately, marketers wanna make sure that they're actually bringing customers to their store or to their site to deliver on, you know, purchasing those items.

Uh, AI is another interface for consumers to research, to understand what different products are available, and to help make decisions. So I think of it as a new surface that marketers can tap into as a way to help both inform the consum- consumers' evaluation criteria, as well as help drive them to a particular action, whether that's using the organic side of AI, of using chatbots to help analyze and understand data that's available, or using advertising in those environments to help influence a consumer's choice when perhaps they're deciding between multiple brands.

Katy Friday: Yeah, and I think when I think about first-party data, I think it really enhances it, and it's really about the opportunity. So as consumers discover brands and new environments, [00:03:00] advertisers should put the data that they already have to work for them, so they can include those high-value customers or prospects that they know that they wanna reach, or they can exclude people who have already converted or past purchasers, or they can even prioritize and value audiences differently to pay more or less based on attributes.

So my point here is don't start from scratch. Use what you already know about your customers to make smarter decisions about where to reach them and where to invest to keep continuity in how you want to engage with them.

Samantha Jacobson: And one of the biggest challenges that marketers have had is figuring out is it correct data?

And so they know their first-party data is incredibly accurate because they're in control of collecting that, or they understand the origin of how that was collected, and so that gives them a really high-quality data set to start with, instead of trying to guess as to whether or not those signals are accurate, which I think is really valuable.

Nate Elliott: Yeah, it, it's a great point, and I've been amazed in my years in the industry at even with channels where the, [00:04:00] the data coming from the new consumer behavior's incredibly rich, I've been amazed at how much first-party data from the brand side has been able to unlock those opportunities and really make them work well.

I, I wanted to get your take, uh, uh, just dig a little bit further on this, though, because AI is fairly unique in that this is a channel people are using throughout their customer journey. That hasn't always been the case with some of the channels we've seen go before. Uh- Does the first-party data question change if someone's at the point of discovery versus research versus closer to that point of purchase?

Samantha Jacobson: I think the message that the marketer wants to give to a consumer will change depending on where they are in that purchase funnel. So if I'm going to chat and I'm saying, "I'm training for a 5K," chat now knows I'm gonna need help with my nutrition plan, what I should eat, that's a great marketing opportunity, making a race plan, figuring out apparel to wear.

All of those are different options or different touchpoints with the consumer to help inform that. That's very different from when I would go to search and say, "I've already decided that I wanna buy a pair of [00:05:00] sneakers. Help me find the nearest Runner's Roost," or whatever my vendor of choice was. So marketers need to recognize that they have an opportunity to dramatically change the way a consumer is going through that journey, and we've seen that data.

70 to 80% of consumers coming from chat to brands in past studies have actually been new to brand. It's a great way to become discovered rather than having that existing consumer preference, so the way you market to a user at that part of their journey is different. But you also need to make sure that you're marketing to them when they're ready to have that decision point.

So if I'm using my first-party data to say, like for example, my husband only wears Sauconys, that is gonna be the thing that you wanna market him. And so instead of saying, "Hey, go try a pair of Nikes," that's not gonna be what he'll try, but, "Hey, we've got the latest," or, "We've got new colors," those are gonna be things that they can use to actually drive him into store to make a purchase.

Using that rich first-party data will help inform that, whereas I don't have that same brand affinity, and therefore I m- may be much more open to a particular type of shoe or a particular type of mileage. That's where I think first-party data can play an immense [00:06:00] role in delivering that performance.

Nate Elliott: And Brooks all the way, by the way, Samantha.

I, I'm, I'm sure the Sauconys work for your husband, but we're a Brooks family in my house. Um-

Katy Friday: I'm one that I dabble, so-

Nate Elliott: Oh, you, you switch around? Okay ...

Katy Friday: I, I am, I d- I do switch around. All right. And I have to say I have- Beautiful combo. There you go ... I- I have been influenced by, in my searches, asking those questions that Samantha just described, so it

Nate Elliott: can work.

And now I, now I really wanna see which targeting groups we're all in- ... based on that. Um, uh, uh, speaking of, of targeting though, I, I wanted to get into retargeting a little bit because one of, you know, one of... I talked about how first-party data and, and stuff that the, that the brands themselves can capture has been so important to unlocking the value of a lot of digital marketing over the past several years.

And retargeting, of course, is something that, uh, plays a big role in that, especially retargeting people who have been to the brand's app or been to the brand's website. But of course, there's retargeting, uh, data, retargeting capabilities built into the [00:07:00] channels themselves. I'm really curious for your take on, on the role that retargeting will play as AI becomes a bigger channel.

And, and what are the sort of special kinds of data that AI interactions or AI customer journeys might be able to offer marketers as they look for ways of getting back in touch with people who've expressed interest?

Katy Friday: I think that's a great question, and I think for years retargeting has meant, you know, looking at a pair of shoes, and you're seeing those same shoes follow you around the internet until the end of time.

And we should be asking, "What do they need next?" So there's so much more that we can be doing when a user sh- shows interest. So when someone clicks on an ad in that conversational AI environment, like ChatGPT, that engagement can help us inform the next interaction, whether that's on CTV or video or audio.

And it gives us the opportunity to think about, do they need more education about the [00:08:00] brand, or would another product or message be more relevant to Samantha, to go back to your husband, who just stays with his same shoes? Or maybe they've already purchased the product, and we should stop advertising to them altogether.

I think retargeting now should be about moving the consumer forward, not just repeating the same ad.

Samantha Jacobson: I agree, and I think It's also, I, I agree with your point, Katie, that the onus is on marketers to figure out what's gonna resonate most with consumers. We now are able to understand where consumers are in a journey in great, in much more s- close to real time than in legacy days of, you know, uh, circulars and sending direct mail to consumers.

And it, the onus is on marketers to update their understanding because of that. I mean, a great example I had last week was I bought a Lolo blanket. I don't know if you guys have tried them. They're fabulous. But now, uh, to your point, Katie, I'm being followed around about all these other blankets. I actually don't need 400 blankets in our house.

But what I would love is someone saying, "Hey, you really enjoyed this particular thing. Here's other [00:09:00] things you may wanna try. Here's slippers. Here's other items where people that liked this Lolo blanket also liked these things." And that would be a great marketing opportunity. So I think it needs to be not just around hitting the consumers with the same message, but how do you think about what's gonna enrich that e- consumer experience?

And that's actually where I think platforms like StackAdapt are incredibly helpful because marketers have so many channels that are available. Like, I live in the world of ChatGPT. I am thinking constantly about what is most relevant for consumers in our platform. What do we know about them? How do we make sure that we're serving them ads that enrich their experience and that's relevant and engaging, rather than just kind of inundating them with things that aren't useful?

But that's just one channel for exposure. They're also watching television, searching online, and making sure that marketers can connect data across all of those different channels to deliver the best possible consumer experience is critically important.

Nate Elliott: Yeah, I love that point about understanding where people are in their journey because, you know, one of the things that I think about is if we're retargeting our audience just based on the fact that they've been to...

I, I don't know what a Lolo blanket is, but if they've been to the [00:10:00] Lolo blanket- Oh, you

Samantha Jacobson: need one ...

Nate Elliott: PDP. You definitely need one. Okay, I'm gonna look it up as soon as we finish taping. Um-

Samantha Jacobson: Wait, use a coupon code. There's a lot of coupon

Nate Elliott: codes out. Okay. I will- They're

Samantha Jacobson: critically important.

Nate Elliott: I will look for a coupon code.

No one needs 40 blankets. We already have 40 blankets, but maybe, maybe we'll get 41. Um- You know, if someone goes to a PDP, you don't necessarily know where they are in that journey. Maybe they heard someone mention in conversation that day, Lolo blanket, and they're just trying to figure out what the heck is this thing.

May- they've, they've been researching it for a couple of weeks, and, uh, and they're now at the point of purchase. And one of the things I like about using data from a channel that actually can help people throughout that journey is that now we have better insight into where in that journey they are.

It's not just, uh, someone hitting a page or not hitting a page. It's, it's some kind of understanding that someone is at the top of the funnel, in the middle of the funnel, or maybe getting closer to the bottom of the funnel, right? That brings me to this question of the consumer [00:11:00] journey more broadly. And of course, uh, as AI, uh, grows as a behavior, as the majority of US online users are now using AI on a regular basis, and shopping behavior in AI continues to increase, we know it's becoming a, a bigger part of the customer journey.

But we also know that there are still lots of other places people go to shop for things, to research products, to consider their options, to, to ultimately make decisions on what they're gonna buy. And I'm really curious to get your thoughts on what role AI and conversational AI in particular should play in that broader media mix.

I mean, how, how should marketers be thinking about where this tool and this channel fits in with search and retail sites and social sites, and all the other tools that people use to make these decisions?

Katy Friday: I think everyone is excited to test new channels when they come out, but they quickly wanna understand how they're performing along their ex- existing investments.

Is it [00:12:00] working? What should we think about differently? And I think that's a very fair question. Um, but now we have reports like cross-channel attribution, which can help advertisers understand how their media mix is collectively working and see how measurable touch points like ChatGPT, video, display, connected TV, and other channels are contributing to conversions.

And at StackAdapt, advertisers can look at cross-channel attribution directly within the platform, and I think the real value here is using those insights to make smarter campaign decisions while campaigns are running. So it's not just measuring what is working here, but it gives you the ability to actually do something with that information to optimize your campaign effectively.

Samantha Jacobson: And so much of marketing is driven by psychology, right? The fact that I think about a beer brand from Clydesdale horses or frogs tells you how much marketing can influence the way consumers think about brands. And maybe it's just too much Clay Christensen, but I think a lot about jobs [00:13:00] to be done around how do you think about creating a connection with a consumer to create that emotional connection versus when I'm in research mode and I want exactly what the data points are to compare and contrast two different products.

Those are gonna be best served through different mediums. For example, chat doesn't have video functionality in it today, so having a bunch of video ads to explain to consumers isn't gonna be the thing that's gonna resonate in that platform. But having a lot of specs to have your dot or, you know, your a- agent of choice help you compare different products that you may be deciding between is incredibly helpful.

And so that's where I think the job is on the marketing team to figure out what are we trying to accomplish with this particular campaign and what medium is gonna best lend itself, whether that's in a streaming environment, whether that's gonna be through a trusted endorsement from someone that you either believe in because they're an influencer that you think really highly of, or someone, you know, that's a confidant in your life, or if you're in more of the cerebral decision-making mode where you want all of the specs on the product and you want a recommendation based on what chat knows about you and what [00:14:00] you've liked in the past.

Those are all different touch points. A brand will need to show up in all of them, but the way in which they show up to connect with consumers is gonna look different depending on what channel they're in.

Nate Elliott: I mean, you both make fantastic points around, you know, Samantha, around orchestrating the different touch points and how brands show up, Katie, about trying to optimize for how and where you're appearing throughout that journey in real time.

But it does require a lot of effort, and it's-- I mean, let, let's be honest, a lot of CMOs and marketing departments are still trying to figure out the basics of GEO and AI advertising. What would you guys say to a marketing department who i- they want to do that kind of orchestration behind the scenes, but there's, there's-- it's just, it's a new muscle or a new skill that they're still working on building?

How would you encourage them or help them figure out how to do all of that organization?

Samantha Jacobson: My first piece of advice would be just start, right? I think the more you can start to test and get comfortable to figure out what learnings you can capture, [00:15:00] what's similar from past behavior versus what's different, and set up a lot of experimentations.

You know, A/B testing different creative formats, different messaging. Something that's been really unique in AI versus other channels is having a very tailored message resonates a lot with a consumer. So in the past, if you were REI and advertising for outdoor goods, you would typically advertise your banner in hopes of bringing people to your site or into your store.

Now, if I'm planning a camping trip on chat with my kids, actually surfacing a sleeping bag based on where we are and what the weather condition is and the age of my child is gonna be much more relevant than just giving me the REI banner. In the past, some of those may have seemed creepy. Now it's incredibly helpful.

And so I think understanding what works and what connects with consumers to drive that outcome versus what doesn't work is gonna be the most important thing. But getting started and starting in small steps and leaning on trusted partners and advisors to help guide you around what they've seen work elsewhere, how they think about industry best practices, is gonna be a way to get your foot in the door, and then you can [00:16:00] slowly get more comfortable and scale.

Nate Elliott: Right. So, uh, let me just back it out as a, as a last big picture question to this notion of audience data and audience strategy, because one of the things that's clear in everything that's happened in the last year or two in this industry, in the conversation we've had today, is that we now have, uh, you know, a yet another tool for helping us understand our audiences and which groups of people are engaging in which behaviors, who we might wanna reach, who we might not wanna reach.

And, and as we learn more about these groups of people, it, it does raise the question of who's most important for us to be talking to, who maybe is a less valuable audience for any given brand to be reaching out to. And I, I just wanted to, to get your point of view on this. As, as marketers find this additional piece of data that helps them understand their audiences, you know, what should they be thinking about, and what's the most important thing for them to get right as they continue to [00:17:00] better understand these groups of people?

Katy Friday: One thing that I, I would say is more data doesn't mean better marketing. It's really what advertisers choose to do with that data to evolve their audience strategy. You know, I have two kids, they're six and seven, and brands that I have purchased with from since they were little are at times still treating me like I have babies.

I still get a diaper ad from time to time. So while they may not know that my kids are six and seven, they have years of my customer history, but the messaging hasn't evolved to with what my needs may be now, or even anticipating what my needs may be in the future. So to me, a better audience strategy means understanding when to engage, when to change the message, and maybe even when to stop hitting that customer.

So it's about improving that customer experience, eliminating wasted spend, and, and start reinvesting those dollars into reaching new customers, is how I view that.

Samantha Jacobson: And in the past, I think we sometimes looked [00:18:00] at audiences as the end-all, be-all. If I can just show these people my message, then I'll be successful.

And I think who you are is one element of it, but there's so many other dimensions that influence that consumer's experience that need to be taken into consideration. In our world, for example, we start with what is gonna be most relevant. So even if a user is included in an audience segment, if the a- if the advertisement for that audience segment is not highly relevant and something that we think the user will find valuable, we won't show the ad.

We'd rather not show the ad than show a user something that we think isn't gonna resonate with them or be of value. And I think that speaks to the broader context in which you're trying to reach consumers. Brands love that, right? You of course wanna reach consumers when they're most engaged, when they're most likely to enjoy your product or be open to experimenting with it, but you need to make sure that you take into consideration all of the other external factors, whether that's context, whether that's environment, whether that's format, in addition to audience, rather than seeing audience as the end-all, be-all and the only tactic you need to worry about.

Nate Elliott: And do [00:19:00] you see brands, given this additional set of data, making big changes to the audiences they're trying to reach, or has it been more of a fine-tuning exercise to this point?

Samantha Jacobson: Personally, I think AI is encouraging marketers to rethink their entire strategy. I think organic traffic is causing marketers to reevaluate how they share information, how they talk about what their products can provide, how they compare products that are available.

Connecting with consumers in those different ecosystems is encouraging them to revisit what's working. I think it's really healthy. I think it's great to revisit things that we've done in the past and figure out what should I keep and what should I throw away? Just like we do spring cleaning with our clothes, I think we should do that at work as well with what strategies work and don't work, and then you may, you may have a whole new wardrobe.

You may find you just need a couple of seasonal pieces. But ultimately, you wanna constantly be asking, "Is this still working? Is this still delivering the impact that I'm looking for, or can I do better?" And that answer will be different depending on the industry and depending on the level of innovation within the marketer.

Nate Elliott: All [00:20:00] right, so we've got running shoes, blankets, and a whole new wardrobe. I, I love that we tied it all together- Yeah. ... in the end. Samantha, Katie, thank you so much for a fantastic conversation.

Samantha Jacobson: Thank you so much. Thank you, Nate.

Marcus Johnson: That's it for this episode. Thank you so much to our guests, Nate, Katie, and Samantha.

Thanks to the production crew, and thanks to everyone for listening in to Behind the Numbers: An EMARKETER Podcast, made possible by StackAdapt. I'll be back tomorrow discussing the big three questions for

Meta.