Subscribe
Share
Share
Embed
Unlocking Retail Media is the essential podcast for leaders and marketers navigating the rapidly evolving world of retail advertising. We move beyond day-to-day operations to explore the strategic future of the industry, covering major investment trends, the shift to hybrid marketplace models, and the existential disruption posed by Agentic Commerce. Host James Avery brings in top industry veterans and visionary founders to analyze how ground-breaking technology is transforming customer journeys, influencing product catalogs, and forcing retailers to rethink on-site, in-store, and digital media strategies to remain competitive in the modern age.
Andrew Lipsman: [00:00:00] ROAS is, it's not the right metric. I mean, it, it's not only a flawed metrics, it, it can lead you in the exact wrong direction. Um, and the problem is return on ad spend is not actually what it sounds like it is. Um, it's a misnomer. It's, it's really just saying that a sale happens. Post advertising exposure.
James Avery: Welcome to Unlocking Retail Media, the podcast where we explore the evolving world of retail media from data strategy to monetization and everything in between. This is where we break down how retailers can build smarter data-driven media networks by aligning with what brands truly need from scalable ad solutions and meaningful metrics to cross channel attribution and programmatic strategy.
Welcome to unlocking retail media. Today, I am talking to someone who I think sees this industry more clearly than almost anyone, and more importantly, he'll actually tell you what he thinks. Andrew Lipson spent over five years at e-Marketer defining the analytical frameworks the industry uses to understand retail media.
Now through his substack and [00:01:00] consulting practice, he's one of the most trusted independent voices in the space, and that independence matters when it comes to looking at the bigger picture in the industry. He just follows the data and tells you what he actually sees right now with retail media going through this really interesting moment of maturation, more scrutiny from advertisers, harder questions about measurement, the whole commerce media conversation, expanding.
I can't think of a better person to help make sense of it all. Andrew, really glad to have you on the pod today.
Andrew Lipsman: Great to be here, James. Looking forward to it.
James Avery: So I, I wanted to, I wanted to kick this off 'cause I think this is a, I think this is a fun conversation we can have. Um, I read an article, you, you recently, uh, published about, uh, you know, kind of comparing to some of the, some of the latest trends in, in, you know, I guess advertising in general, but also retail media.
Uh, to the, uh, to the ERO device?
Andrew Lipsman: Yes. Uh, well, well, it goes back to a talk I gave that, uh, architecture live recently, and it was called, uh, when the ERO is not worth the squeeze. Um, now a lot of people don't have memory of [00:02:00] ero. The, uh, it was about 10 years ago that this, um, Silicon Valley startup, it got $120 million in funding, um, was they were selling and pitching this, uh, expensive $700 juicing device.
And, um, you know, really playing it up like it was the next great thing. And then when it hit the market, they realized that actually you did not need the four tons of pressure to squeeze the juice packets. Um, you could actually just squeeze these juice packets by hand. So the whole device wasn't really needed.
Um, and basically I likened it to the magical black boxes that I think were being pitched increasingly today. Um, the AI hand wave that, that basically. You, you don't need to see what's actually going on inside. Just trust us. It's gonna just give you the outcomes that you want. We're just gonna give you the juice.
Um, and I, I think we need to stop buying into that premise. Um, there, there is, uh, a lack of transparency in a lot of cases where we don't really know what's [00:03:00] going on. Um, and advertising isn't actually performing like you think. Um, these black boxes are really just kind of tuned to. To give you the KPIs that you're going after, but those KPIs don't actually always tell the whole story.
James Avery: Yeah, I think it's, it's fascinating to talk about because I do think I, I end up, you know, rightfully or wrongfully ended up on the, like, I think the, the technologist side where we, we sit there and we think like, okay, like you should be able to, you know, like the dream is you come in, you're a marketer and you say, you know, I'm trying to sell more, more ketchup in Ohio.
Right. And, and you kind of say, here's where I'm willing to spend on it. Here's my target roas. And then we say, don't worry, you're pretty little ahead about it. We'll, we'll, we'll drive this, this, you know, ROAS result or drive this, you know, whatever result you're looking for this, this CPA target or, or things like that.
Um, and so I think, I think it's interesting to go, 'cause like I think I, I do think, like I see that on, you know, not [00:04:00] just, you know, on really on, you know, it's like. Google's version of, you know, P max or gmax, whatever it is, and Advantage, advantage Plus on meta and all these different things going out there.
And you also see it more in retail media, right? Like things like auto bidding and, and how do you, how do you really just. Let somebody set an objective and then don't worry about the rest. So I think, I think technology people tend to see that as like, oh, this is great. We can, we can kind of, you know, automate this process for marketers.
But I'm like the other, what's the other, what's the other side of it? Like, why, why is it, why doesn't that work?
Andrew Lipsman: Yeah. Well, so first off, it's not a dream. It's, it's a nightmare. Um, and don't take this the wrong way, but technologists generally don't know the first thing about how advertising works. Um, it's, it's a very reductive view of the world where everything is just measured in terms of short-term KPIs.
Um, that's not how marketing works. You're, you're actually, uh, abstracting away strategy, um, understanding who you're reaching and why all these things matter. And really advertising, if [00:05:00] you wanna grow your brand. For the long term, you need to be able to, um, create those feelings about your brand. It's not just about converting somebody tomorrow.
If, if you look at every short term window about, um, optimizing for that window and converting, you know, you're just gonna get more and more targeted over time and the KPIs will look great, but you're, you'll actually shrink your brand. So, um, I don't think optimization is, is really the path to brand growth at the end of the day.
James Avery: So when you think about, uh, I think in your article too, you mentioned kind of ROAS being, being one of the, the most guilty metrics that everyone's looking at.
Andrew Lipsman: Yeah. I mean, I, I think that, um. This, this notion of optimization is about efficiency, but it doesn't really work that way in practice. The, it's only efficient when you're optimizing to the right things.
And ROAS is, it's not the right metric. I mean, it, it's not only a flawed metrics, it, it can lead you in the exact wrong direction. Um, and the problem is return on [00:06:00] ad spend is not actually what it sounds like it is. Um, it's a misnomer. It's, it's really just saying. That a sale happens post advertising exposure, and it's not actually attri correctly attributing, um, the sale to that exposure in, in a causal sense.
And so what ends up happening is these, these black box systems or really any optimization really just starts to do things that optimize for taking credit. For this, the attributed sale. So two of the most common examples of that would be in retail media. We see branded search, a lot of dollars. Just go into branded search.
Well, that's somebody who's there to buy your brand anyway. I'm not saying you don't buy any branded search. You need to do brand defense and things like that, but that tends to not be incremental. And the other big one is retargeting. Um, retargeting. Yeah. Sometimes they can tip the scales and finally convert that user that you've been after.
But a lot of times it's just putting ads in front of somebody who's already pretty far down the funnel and then you get to take credit when [00:07:00] they convert, even though they were gonna do that anyway.
James Avery: It's kind of basically like a modern version of like cookie stuffing, which I'm sure is still, is basically what retargeting is, right?
Like just being like, how do I be the last, how do I be the last view before the conversion? To get the affiliate dollars or the credit or the ROAS credit or the, you know, whatever, whatever, you know, the optimization system's looking at
Andrew Lipsman: that, that's exactly what it is. And, um, you just, you have to have a system that can't be gamed because the machines will, will game it otherwise.
And really that's what we're seeing happen, um, writ large across the, the entire digital advertising industry right now.
James Avery: Yeah. It's, well how do you, and how do you think about, so this is, I think I mentioned this on the, the last podcast, but we, we actually found when. When building kind of machine learning models, if you gear them towards just optimizing click through and roas, they will quickly work towards things like brand affinity and, and customer affinity to advertisers and, and customer affinity to a [00:08:00] product.
And so basically it will optimize towards, well, Andrew always comes to this site and he clicks on the Coke category, and then he buys Diet Coke. So he is the best person to show this diet Coke ad too, right? Like, and so basically it's the inversion of incrementality. Like we can, we can like track incrementality and conversion and the, the, you know, one of the algorithms we, we wrote would basically push up roas.
Push down incrementality because it was just, it would just figure out what's a better way to show this to people that weren't gonna buy it anyway.
Andrew Lipsman: That's exactly what happens there. There's an inverse relationship here between these two things. Um, in practice that, that usually by virtue of. Chasing roas.
What you're often chasing is cheap impressions in front of people who are already gonna con convert and the incrementality is terrible. Meanwhile, if you do the things that really move the needle on incrementality, you're probably buying expensive media, um, and the ROAS looks terrible. It pushes you into things like premium tv, [00:09:00] um, or, uh, sponsored brand ads or sponsored brand video.
Those are more expensive forms of retail media. Um, but guess what? They work. They really do work. So, um, I think that's one of the big lessons is if you want, if you really wanna achieve advertising effectiveness, you need to be willing to pay for it. Um, otherwise you get what you pay for.
James Avery: And how do you, do you think this is a more of a problem on what retailers are offering or what brands are asking for?
Like, I'm often kind of sitting there thinking like, okay, well sometimes we'll talk to a retailer and they're like, the brand is coming to us saying they want four X roas. And it's like, right, you can optimize towards your four X roas. Um, but the brand, you know, the retailer's kind of at the mercy of the brand at that point.
Uh, you know, 'cause that's what the brand is pushing for. But then also I think retailers are quickly, you know, are probably a little bit guilty here as well because, you know, there's lots of systems out there saying, Hey, we'll just automate everything for you and, you know, put in a price and we'll, we'll generate you roas.
Andrew Lipsman: Yeah. One of the, the interesting things is I, I thought that [00:10:00] brands largely were unaware of some of these issues, um, because it seemed, you know, they're the ones who are spending the dollars. And if you're spending dollars inefficiently, don't you wanna change that? Um, but the problem is. Individual people's incentives within brands actually steer them in the wrong direction, right?
Their, their goals, their objectives are oftentimes around cost efficiency. In roas, you have procurement within brands, and their whole goal is to, to drive those CPMs down. Um, so by virtue of doing that, you're, you're basically always looking for low cost and high roas, and that's actually what's driving them away from, um, incrementality and, and true effectiveness.
So it's institutionalized. And I used to th think of the brands as the victims. Um, but really they're, they're victims of themselves in, in large part,
James Avery: yeah. I mean, to some extent, right? Like the, the, the black box programs from Meta or, you know, Google wouldn't, they wouldn't be there if brands [00:11:00] weren't using them.
Like, there's definitely a, you know, like drug dealing, uh, like kind of aspect here where, like, whose fault is it? The drug user or the drug dealer, right? Like, it's kind of a, a two-way, two-way street.
Andrew Lipsman: Listen, it's designed to make their lives easy, um, and to produce the metrics that they want. So it, it's, it's very easy for brands to adopt that.
Um, but you have to think about what you lose in, in the long term. Um, you, you're not seeing what's actually working. You're losing that visibility. Now, the funny thing is about optimization systems in a, in a, a better scenario, I would say. Well. You could correct it if you could optimize towards long-term brand building metrics.
Yeah. Or, you know, LTV, things of that nature. But guess what, it takes too long for the lifetime value metric to accrue, right? You can't optimize to something that you're not gonna see fully manifest for 12 months, let's say. So you can't optimize to that in the short term. So there's, there's an inherent imbalance between short-term horizons and optimization and [00:12:00] what brands really want, which is to optimize for long-term growth.
James Avery: Yeah. When we do, we do see this with, you know, when we look at the more sophisticated brands when they're buying from our customers, what they do versus the less sophisticated, right? Like, and I think you go down the sophistication model, it is all, how do I automate everything, put in a budget, go. Um, when we get into the more sophisticated brands, you know, we can track whether it's like new to brand, right?
Like, oh, Andrew buys Diet Coke all the time. He is never bought Diet Pepsi. Right, and we can, you know, you can then go in and say, I only wanna target people that would be new to brand. That based on all the order data you have the last, you know, three years. You know, he's never ordered Diet Pepsi, so we're willing to pay more for that.
But you know, RO is gonna be in the, you know, in the dumps, right? Because you might be a super loyal diet Coke drinker, that's never gonna change. But if you did change, you know, you're probably worth a 12 pack a week. I don't know how much diet Cho do you drink? I mean, I'm putting a lot on you here, but like, you know what I mean?
Like
Andrew Lipsman: I'm zero over Diet Coke. I don't know why anyone drink Diet Coke. There's a lot in there though. But, um, so first off, new to [00:13:00] brand, um, that's another imperfect metric, right? It probably correlates with. With incrementality. Um, but it's just a proxy and it's probably works very differently if you're a low awareness brand versus a high awareness brand because new to brand only means you're new to brand within that retailer channel, right?
Um, and really just, you know, for the period of time in which their data goes back. So maybe it's only within a two year timeframe. Um, so, so it's not a, not a perfect metric, um, to begin with.
James Avery: Yeah. So, so thinking about, you know, so we're, you know, we're in the business of kind of helping retailers build. You know, ideal, you know, the ideal retail media platform.
So thinking about it from this measurement standpoint, from from, uh, you know, what brands should really be looking for. What, what should that system look like? Right? Should we be focused on, you know, ioas, should we focused on, you know, I think, you know, like the long term, I agree with you. Like the, the ideal world would be okay, we can actually map all your brand objectives for the next two years.
But you know, the, the time horizon make that [00:14:00] makes that difficult. Um, what do you think, what, what should retailers be kind of aiming for when we think about these, these optimization systems or targeting or, or things like that?
Andrew Lipsman: Yeah, so, um, first off, just eventually getting off of roas. Um, you know, it's not a good metric, but I'm a realist.
I know it's gonna be around for a while. And so what I'd like to see brands do is start to learn. Um, to, to measure with I roas in mind, see what the ROAS is. One, start to learn that those two things don't have to be correlated at all. Um, and ultimately kind of break the addiction to roas and be making your optimizations on incrementality.
The whole industry will argue over what's the right method of incrementality. I would say it's often context specific. Um, there are some methods that I would say are not very good. There are some that are very good in general, if you're using a pretty good method of, um, calculating IRO as you're gonna move your brand in the right direction.
Um, so that's, that's the first thing I think that's essential. Um, and then looking [00:15:00] at other metrics that help tell a, a bigger brand story, um, there's not enough of a, a view of market share. That's an outcome metric. It's not just about, um, you know, driving that ROAS metric, but it also, it accounts for the competitive set.
You could be doing a great job strategically, but if your competitors are coming in and spending more than you, you may not be able to gain market share. So market share is one I think's undervalued. Um, and then we just mentioned it before, but lifetime value, actually it's usually, it's it's long-term value.
You're not really, in most cases over the lifetime, but in the world of CPG for example, you know, 12 months is a, is a good. Um, longer term value metric. If we can start to build those into, um, the platform, I think that's gonna give brands a much better sense of what they should be doing.
James Avery: Yeah, that makes a ton of sense.
And the, uh, the second one you, you talked about, right, so IRO ads makes a ton of sense. Share of, share of, uh, kinda share of category or, or share of, uh. You know, that's something we started to see movement on that I think is [00:16:00] really interesting. 'cause I think, I think Amazon's done some work there, but really being able to show people, Hey, this is your, this is your share of diet Soft drinks being sold on Amazon, right?
Yeah. Because your number might just be going up into the right, because Amazon's number going up into the Right, right. You're just selling more because there's more people buying from Amazon versus like, are you, are you outpacing your, your competitors or what's your, what's your percentage of that category?
I think it's a really interesting metric that I wish, I think more brands should look at.
Andrew Lipsman: And, and actually I should mention the, the corollary to that, which is share a voice, um, right. Share a voice is how am I spending it against my competition right now? Right? That will drive some short-term value, but it's also gonna drive that long-term value.
The market share is more of a, a lagging indicator. Um, so I think ideally you're looking at both things in your managing to them.
James Avery: So going, going back to our original, uh, you know, the, the juice algorithms. You know, if, if the, if the I, you know, these optimization system, right? If it's, if it's built towards these other metrics, do you think that that kind of makes more sense?
Right? Like [00:17:00] if we're, if we're optimizing towards I ROAS and things like that, then, then, then, you know, like the AI and the, the machine learning can help out here, but it's really, it's really the ROAS is kind of the root of all evil in this scenario. That's
Andrew Lipsman: exactly right. ROEs is the root of all evil. I'll give you a another example here at these systems.
So one of the things that. The black box algorithms do is they will default to including, um, not just Google and YouTube, right? Those are premium channels, but also the Google Audience network, which is Open web. And frankly, it's a lot of junk out there. Um, that junk gets a lot of bots. Bots click a lot. Um, so if you're looking for things like clicks or it's a lot of cheap inventory, it can help boost roas.
You're actually actively optimizing towards the worst quality inventory. That only happens in a ROAS view of the world in an I roas view of the world. Right? Bot inventory can't drive incrementality. Yeah. It's not gonna work. It's gonna zero out. [00:18:00] So it, it would automatically, um, it would automatically start to root out the worst performing inventory.
So I, I think that's, you know, that's one of the, the strongest cases for why, um, you should move towards IRAs.
James Avery: No, a hundred percent. Yeah. The, uh, the old, uh, checkbox on Google. I remember even 10 years ago, I would go to our marketer and say, open up our Google campaign. Like uncheck that box that says also, also run this across the, you know, whatever Google display network.
Uh, 'cause it would do, even if you're doing keywords right? You're like, we're wanna target a keyword for retail media platform that would enable that and, and do a keyword ad somewhere, and, you know. You know, it wasn't working at all. Right? But it was, it was driving, driving, uh, you know, fake clicks and, and kind of, or cookie stuffing, right?
Like just showing that, hey, this person didn't even see it, right? It showed up on cnn.com where they're looking at something else. But, but now you get that last view attribution.
Andrew Lipsman: Yep. And, and the smart advertisers, I mean, they know that the boxes to, to check out of, right? Um, but not every [00:19:00] advertiser is this sophisticated and there's a lot of, uh, new ways.
They're always the, they're always looking for new ways to, um, to game attribution. So, um, it's hard to stay ahead of that curve. So, um, even the smartest brands aren't gonna be optimizing to the best metrics, though. Some of them can at least, um, minimize the waste by optimizing towards much better metrics.
James Avery: And how do you, how do you think, you know, as we think about optimizing towards the better metrics, and maybe this is kind of taking a step out of just the, the kind of ROAS ERO line here, but thinking about AI in general. You know, I think a lot of people are talking about, you know, agentic, commerce, all of that sort of stuff.
But instead, I'm actually more interested in, in how do you view, you know, the capability of ai, you know, improving a brand, you know, brand marketer's life, right? Like where, where are the places where they, they can use it to do more effectively, do their job?
Andrew Lipsman: Um, I think of AI as a tool. Um, I think that anytime there's this impulse to set it and forget it, um, let, [00:20:00] let AI do the work.
I think that's where you're gonna run yourself into trouble. And I, I would argue from an advertising standpoint, at this stage of the game, you are way more likely to, um, engage in bad marketing and advertising waste money, um, than get any benefits of efficiency. So letting AI. Do my creative for me? Well, one, I don't think that's how branding works.
You, you don't wanna have a million iterations of creative. You're not trying to personalize things to everybody. You, you, it's actually more impactful to have a singular brand that everybody sees and understands in the same way. Now I can use AI to do, uh, iterations of different variants so that my ad formats fit in all the different platforms and channels.
I think that's totally reasonable, but I don't think you want it creating its own creative. Um, there was that example from a couple months ago with the sort of, uh, young male apparel brand that, that, um, on meta all of a sudden was seeing variants That was like trying to sell. [00:21:00] Sweatshirts and sweatpants to grandma's.
James Avery: Yeah, yeah, yeah.
Andrew Lipsman: That's not the brand that, that you
James Avery: want. This is the algorithm. The algorithm figured out Andrew. Trust the algorithm. Yeah, yeah, yeah.
Andrew Lipsman: Yes, yes, exactly. Um, and, and right. It's probably optimizing to the clicks and maybe the older people are just more likely to fat finger clicks. Like literally are, they're the ones on
James Avery: Facebook, right?
Like it's, yeah.
Andrew Lipsman: So, uh, so I worry about letting it, um, you know. Giving it the reins on, on creative and cr, you know, generating the creative and then also on the optimization for the reasons we talked about before. Unless you have really perfect information, um, and data that it's optimizing against, it's way more likely to go astray and start basically spending your money on, on the worst things.
James Avery: Yeah, totally makes sense. Um, I think it's a, it's a different narrative than the one you hear out there that, you know, AI is gonna replace. Kind of all these different jobs, but I think when you, when you take a step back and you think about what does real [00:22:00] branding, what does real brand advertising look like?
What is, you know, what do these things look like when somebody knows what they're doing? Then, then, yeah, this, these things aren't gonna be automated anytime soon. Right? By the time they're automated, we're dealing with like a GI, and they're a whole host of other problem.
Andrew Lipsman: We, we have an entire business culture built around the ideas of efficiency, right?
Just the idea of squeezing out profitability. Um, but. That's only good once you're actually growing first. Right? And, and so we've, what we've really lost in the last decade or two is a, a real sensibility here around marketing effectiveness, advertising effectiveness. Um, you know, it's kind of a shame in the us.
I grew up in that school early in my career, um, and I, I believe that those principles are, are still. You know, they're tried and true and, and will exist forever. Um, and you'll still talk about it in the UK and Australia. They still have a, a strong culture of marketing effectiveness. We basically lost it here.
Um, we, we've gone, you know, we've really [00:23:00] overcorrected towards pure performance and pure KPI driven marketing.
James Avery: How do you, and that you think that's affecting, you know, is that affecting the health of brands in the long run? Right. Are we seeing like more historic brands? Kind of be driven into the ground through poor marketing.
Andrew Lipsman: Yeah, we, I mean, we've seen some strong examples of this happen with, uh, Nike. Very recently, right? Um, they brought in a, a digital guy, uh, John Donahoe from, uh, from eBay, and all of a sudden they moved, they shifted into, you know, purely D two C and performance marketing. Um, but Nike was all about this brand.
People kept wanting it and buying it because the brand was so great, and so they, they just chipped away at the brand, um, you know, for these short term metrics. And all of a sudden they found themselves in trouble. So. Nike's a great brand, right? It's, it's, it can rebound. It will rebound. Um, but we do see brands start to target themselves into obsolescence, and they, they can absolutely become less [00:24:00] relevant, um, with generations over time.
So you, you need to maintain that cultural relevance. Um, and that means showing up in the channels where you can actually reach people.
James Avery: Yeah, I mean, I think, I think your other example in your article, um. Was like the rebrand of the, the orange juice container. Um, yes. And we, we always, I always joke about this, my wife's an NBA and I'm, I'm not an NBA and so I always blame these things on MBAs, like where my daughter, we were all in the car the day and my daughter was like, you know, why would, why would somebody do this?
I think it was like the Cracker Barrel logo or something like, another example. And I was like, oh, some NBA came in and ran a spreadsheet and showed that, you know, these changes would, would drive X amount of, you know, save X amount of cents or, or drive X amount of like new margin. Um, but really it is, it's like, it's kind of the over optimization and not having just the sensibility of when you go into the aisle and you see the big orange with the straw in it, you know, it's Tropicana and you've been buying that for 25 years and, and trying to update [00:25:00] it and make it sleek, kind of killed the, the one thing they had going for 'em.
Andrew Lipsman: Yeah, I mean, you know, part of it is that when leadership comes into new positions, they have to make their mark. So a new CMO comes in. Maybe they have to do a rebrand that's not really needed. So the incentives aren't, aren't always that well aligned. Um. But yeah, I mean, I think the strongest brands really don't change that much over time.
You know, they go through, uh, evolutions. Maybe they, they, you know, update the font and, and little aspects of it, but it's not the sort of thing where they're, they're totally upending what they do. Um, I think it's very easy. Um, within the confines of a business, especially an insular business, to lose touch with the consumer and what actually drives their decisions.
Um, so in this case with Tropicana, right, it was this heuristic that just every, it was very identifiable. It's a commodity product. Um, and it made people reach out for you and they rebranded to something that they thought was cooler and more modern and more innovative and actually look generic to people.
James Avery: Yeah. [00:26:00] Makes a ton of sense. So how do you, so, uh. You know, taking it back a little bit more to the retail media side of things, um, and maybe take a little bit broader lens. Like, you know, you, you watch the space closely, you work with a lot of different, you know, brands. Uh, where, where do you think, where do you think retail media is headed?
Like, what are the next, what are the next couple of years look like in, in retail media?
Andrew Lipsman: I mean, the big shift that's happening right now, and, and it's a pretty exciting time, is we're, we're starting to move up the funnel. Um, I think that retail media has been pigeonholed into being just a performance channel, e-commerce, and actually a lot of CMOs actively ignore it for that reason.
Even, even though they're, they're responsible for it, they're kind of like, let the e-commerce people deal with that. Um, you know, I, I say that retail. The future of retail media is the future of all media, just because it's the foundations of retail media is better targeting data based on people's purchase and shopping behavior, um, and the ability to close the loop [00:27:00] on sales and understand performance better.
So it's, it's evolving into all these upper funnel media. And I think that's where we're gonna see some really interesting things happen the next couple of years. The one that are, you know, I think is very imminent is performance tv. You know, Amazon has this targeting data, closed loop measurement, and now they are underpinning.
Um, a huge part of the ad supported CTV market. So this is a market that's, that's ready to explode. And guess what? This, the CMO can't ignore it. Um, in-store media that's finally becoming real and tangible, and we're seeing rollouts of these digital screens across, you know, Kroger. Um. And, you know, Walmart's probably not far behind because whatever Kroger does, Walmart's not gonna be a laggard too long.
So, um, I think we're gonna see that market materialize in, in a much bigger way. So there's a lot of excitement. Um, and then, you know, the other aspect of it is like, even as these new ad markets get created, there's so much [00:28:00] infrastructure that's required. There's so much u unique measurement that's required, um, that it's also a, a time of great innovation in some of these areas.
James Avery: Yeah, absolutely. Like on like going back to the CTV side, like we, I had James Barrow from Universal Ads on, right? Like the, they're, they're kind of making, you can buy premium TV this way, right? Not just, you know, I think CTV has sometimes had a bad name for, you know, some of the same issues the Open Web has.
Uh, but with people like Universal Ads opening up, you know, the, you know, you can buy whatever you can buy directly. You can now buy through, you know, a CTV type. Uh, lens. And I do agree, like if you bring the data to bear and you can actually show attribution, hopefully incremental, uh, you know, ROAS here, uh, then, then that, that's a big opening.
Um, and in-store also, like in-store, I think, you know, also has these measurement challenges though, right? Like how do you, how do you think about in-store measurement and, and the right way to, to, to kind of bring that into that experience?
Andrew Lipsman: Yeah, well let me, let me talk about performance TV really [00:29:00] quickly.
'cause this is kind of, uh, funny. What, what's happening right now. This is how we mess up as an industry. Um, really promising ad markets. So because it's born out of retail media, it's often looked at through a performance list. So a performance lens. So you're looking at TV inventory, which has always been looked at through campaign delivery and brand building metrics.
We're now looking at it through a performance lens. Um, but that ROAS looks terrible because the, the premium media is expensive, so it just drives down roas, all else equal. The only way you can get the ROAS to pencil out is do a lot of TV retargeting, so we're already going down the wrong path of, of optimizing to performance.
What I'd like to see is that one, we have incrementality baked into that measurement. You're gonna get a truer picture of the effectiveness and we can't forget about the brand building. So we have to be able to layer in as part of the KPI. The, the branding effect. Um, if we do that, we're gonna be in a much better place to, to optimize the [00:30:00] media, um, in-store.
Thi this is like in-store measurement's, kind of a white whale of mine. Um, I, I worked in, uh, an early provider of, of in-store retail media for, for a brief time, and one of my first projects there was to basically build the measurement system for, to measure in-store media. This was, um, in-store audience measurement.
Brand lift measurement and sales lift measurement. And, um, there's a lot of difficult challenges because the data has to be location based in order to do this. And then also you have to understand specific exposure within the store. Um. All I can say is there, there have been a lot of technologies that have been tried.
They haven't worked very well. They, they all kind of have, uh, Swiss cheese elements to 'em where they're just missing core components. So they're based on, um, major assumptions. There's some new technology that's coming to market. Um, we'll, we'll start hearing about it more later this year, um, that kind of cleanly and elegantly solve [00:31:00] for a lot of these, um, problems.
And we'll give you a clean read on impressions and ultimately be able to drive. Um, that closed loop measurement that you need from specific ad exposure to whether the brand was purchased in store and be able to do it, um, through, through an incrementality lens.
James Avery: Yeah, that's awesome. I mean, there's also like the, you know, this ties into the kind of brand building piece where we just talked about with CTV or in store, you know, and I had, I had Andrew Evado on, uh, you know, a couple episodes ago, is the answer really just.
MM, right? Like, are we going? Is, is that really the, how brands should be looking at this and, and how your marketing, you know, doesn't really matter as long as you can run it through a, a proper MM.
Andrew Lipsman: The, the key is a proper MMMM. Um, now I'm, I'm generally gonna follow what Andrew says. He's a, he's a super smart guy, but there's, there's a lot of problems with how MMM is put into practice these days.
A big part of it often is that, um, you know, it will bias us towards priors. Um, in other words, like [00:32:00] whatever them MM has said before, um, it will, will basically tell you, you know, if, if you've always been, um. Wedded to, to advertising on TV or meta or YouTube. It tends to bias us towards those examples. And it actually has been systematically understating retail media as a result of that.
Use better methods, um, whether it's, uh, you know, a true test and control, incrementality measurement or um, uh, an econometric method like causal modeling. You're actually gonna see that some of the most undervalued media channels right now are retail media.
James Avery: Interesting. A good, that's a good headline, uh, to share, right?
Like, I think that's all the retail, that's all the retail media people want, wanna believe. And so I think it's, it's, it's good that it sounds true.
Andrew Lipsman: Well, and, and that's, and this is, this is coming up, uh, time and time again. And to me it makes sense from a marketing effectiveness standpoint when you have high quality media.
And you bring it, um, to something that's commerce adjacent or close to the point of [00:33:00] purchase, right? That's actually where advertising can be at its most effective. It's not always put into practice that way in retail media, but if it is, um, it can be highly effective. You know, meanwhile we're comparing to some of these other digital channels that are just so saturated at this point and, and really over invested.
The incrementality is not there, um, like it used to be. So, uh, you know, I tell RMNs all the time, you, you should be trying to graduate brands to thinking about IRO as, because you win. All else sql, you win in that worldview right now. You know, you've sort of benefited from ROAS measurement. You've taken the easy money from it, um, the last few years.
So it's not like it hasn't been to your benefit. But long term, there's a, a much bigger part of the market that you can take if you start to align more with incrementality.
James Avery: Right? And it's, it's also making sure that who you're being compared to is measuring it that way. Right, because like if you're, if you're, if you're looking at meta outing the ROAS number and you're [00:34:00] showing just an I ROAS number, you're gonna look bad.
But if META'S actually showing an I ROAS number as a retailer, you're gonna look great. Right? That's gonna have, that's a much bigger impact.
Andrew Lipsman: That's the exact complaint that, that I will hear. Um, rightfully so. You know, this is the, the practical application of the data is if everyone else is doing ROAS and you're comparing against these channels and they're showing us six or eight roas, right?
And you're a retail media network and you're showing a four or five roas, that's pretty good. You'll, you, you'll be on the plan or you could show a dollar 50 IOS. That's actually really good. That's a true 50% return on investment, right? But that dollar 50 doesn't stack up to those much bigger numbers and you get cut from the plan.
Like that's how illogical it is. But that, that's what's happening. So to some extent the Armand's have to play the ROAS game. I just don't think it's in their long term interest to do so.
James Avery: Well, I think it's just a matter of like, how do we, how do we keep pushing up to the brands, right? Because, uh, it's a matter of maybe it's showing ROAS and I roas.
Right, and saying like, Hey, if you're [00:35:00] comparing us against meta and you have plans towards roas, this is your roas, but this is the number you should be thinking about. Go back and ask your, you know, ask your other providers what your IRO as here is. Right? Compare it to apples to apples. Um, but I think it is more about, you know, you have to be able to, as a retailer, probably provide both, right?
You can't just just show one.
Andrew Lipsman: Like I said, I mean, ROAS isn't going away. Um, the easy metrics, they, they, um, live long past their expiration date and you still have agencies measuring display ad campaign effectiveness based on clickthroughs. That's crazy.
James Avery: Cpm, that's my favorite one. Just how is that
Andrew Lipsman: effectiveness me metric?
James Avery: Yeah. Yeah. I can get you, I can get you all the CP m you want like a nickel, but you're not gonna get any, anything out of it. Yeah.
Andrew Lipsman: Yeah.
James Avery: Awesome. Well, Andrew, this has been great. I think you've given everybody a lot to think about. Uh, and I think there's, there's a lot we can keep pushing in the industry towards incrementality, towards IRO as, uh, so thanks, uh, thanks for joining.
Andrew Lipsman: Thanks so much for having me, James. [00:36:00]
James Avery: Thanks for tuning in to Unlocking Retail Media. If you enjoyed this episode, don't forget to subscribe and share this show with your network. We'll be back soon with more insights to help you navigate the future of retail media. See you next time.