← Marketecture: Get Smart. Fast.
Marketecture: Get Smart. Fast.
Episode 165: Melissa Burdick from Pacvue on the state of commerce media, plus TTD's continued woes
00:53:49
Loading summary
Transcript185 lines
- [00:01]
A
This is a paid ad from our friends at Vibe. If you're running performance marketing at scale and still treating CTV as an awareness only channel, you're leaving money on the table. Vibe Co is the most advanced streaming TV ad platform built specifically for performance teams, giving you the same self service, control, targeting, precision and real time optimization you get from Meta, but on the big screen. Right now, Vive is offering a 50% spend match up to $200,000. Head to Vibe Co and mention this podcast to claim the offer.
- [00:40]
B
This podcast is brought to you by audiohook, the leading independent audio dsp. Audio Hook has direct publisher integrations into all major podcast and streaming radio platforms, providing 40% more inventory than what could be accessed in omnichannel DSPs. What's more, audiobook has full transcripts on more than 90% of all podcast inventory. Enab enabling advanced contextual targeting and brand suitability. Audio Hook is so confident that in addition to CPM buys, they offer the industry's only pay for performance option where brands can scale audio and podcasting with peace of mind knowing they are only paying for outcomes. Visit audiohook.com to learn more. That's audiohook.com. All right, welcome to the market Extra podcast. This is Ari Paparo. We're back. We were out last week at our busy event. Eric, you're back. We're back. We're excited to be here. Did you recover yet?
- [01:42]
C
Oh man, barely. What a two day event, man. Hats off to you and the team. It was incredible. I think it exceeded every expectation from the venue to lucking out with the weather, to 1,000 people in attendance for two days, to Jeff Green closing out the show with you, preceded by Lance Armstrong. I mean, like this was amazing, man.
- [02:06]
B
Thanks. Yeah, it really couldn't have gotten better. The venue was beautiful. The weather was beautiful. The next day after the event, after we had 70 degree weather, it was raining and hailing in New York. It was like we just got so lucky. And I hope everyone enjoyed themselves. We've gotten great feedback from the people who are there and you heard probably if you're a regular listener, the last Friday we put the Jeff Green interview on immediately followed by all kinds of breaking news about trade deaths that we're going to talk about later. This is being recorded on Thursday. On Monday coming up, we're going to broadcast the Terry Kawaja session, which was amazing. He was dressed like, he's dressed like Bad Bunny and he has Spanish titles and it's really something to watch and. And then we'll Be kind of taking some of the best recordings and putting them on this podcast channel on Mondays and then also putting them on if you don't listen to it. It's a great podcast, the Ad Tech God podcast, sort of mixing and matching some of the best episodes. And then we'll put them all, all the recordings on YouTube. So if you do want to watch them at your own leisure, they're not there yet. So we need a little time. We'll probably send an email to the list when everything is ready and kind of formatted and all that stuff.
- [03:18]
C
Amazing. What was your highlight? Give me your highlight.
- [03:23]
B
I think that the Jeremiah Oyang, who is our day two keynote, did a great job on AI, the future of AI agents, et cetera. We wanted to bring him in specifically because we knew he was going to really be interesting. I talked about in my newsletter, but also we'll talk about in this episode that There was this AI demo from PubMatic and MIQ that was exceptional and Lance Armstrong was interesting. The NFL panel people said was interesting but I was backstage so I didn't see it. And then the Startup Showcase. What about you, Eric?
- [04:01]
C
Yeah, I mean everybody heard the Jeff Green interview, so to me that was, that was a highlight. Having you on stage with him, having a different conversation that I think he would typically have was really good. He said a few things that were very interesting, specifically around, you know, testing AI. We'll talk about that in the new section. And then, yeah, the Startup Showcase was awesome. We had five great startups doing all different things. The winner of the judge's choice and the people's choice was the same founder, Alicia Richardson from Crowd Access, which is interestingly measurement for experiential and in person, like a very interesting large adjacent space. And I hear we are going to meet her next week.
- [04:46]
B
Yes. I was so impressed with her and talking with her in the green room, I said, hey, why don't you come on the pod? And it's not like it was a prize for the startup competition. I genuinely want to hear about experiential marketing and she knows a lot about it. So that'll be next Friday.
- [05:00]
C
What a presentation it was.
- [05:02]
B
It was really good.
- [05:04]
C
The applause was thunder.
- [05:05]
B
It was. And, and the other startups were great too.
- [05:08]
C
Oh yeah, they were great. All of them were, were great. But it's just, yeah, this, it's a pitch comp. And, and you know, people came to me afterwards kind of, kind of wanted to talk about, you know, the, the way decisions were made, particularly on the judges. It's this is a pitch competition and Alicia pitched.
- [05:22]
B
Yeah, you got to practice, you got to make that pitch impactful in the first minute or two.
- [05:28]
C
But yeah, great products. Great, great, great, great. All five.
- [05:31]
B
And speaking of which, you're doubling down on Possible in your role as judger of startups. So what are you doing? Impossible.
- [05:37]
C
Thankfully I'm not a judge this time, so we are so Imperium is an investor just like we are in architecture in Impossible. Possible has been great, just like Architecture about embracing innovation. There is a brand challenge where if you're, you can enter into a similar pitch competition where you'll be on stage, there'll be judges, there'll be criteria on what to do and then if possible, where you never know if you're gonna have brands, agencies, decision makers in the audience. So you should submit. Just go to Possible, it'll be there. Brand challenges. Google Possible Brand challenge. The deadline for submission is Monday March 23rd, so please get it in before Monday and then there'll be ample opportunity to do some really cool stuff in Miami. So thanks for letting me, let me get this out there.
- [06:28]
B
Yeah, of course. That should be fun. I'll stop by, albeit possible, doing very little and hanging out and getting the nice weather. So our guest day, we didn't talk about our guests. We usually talk about our guests right away. So today I'm excited. We have Melissa Burdick. She is the co founder and president of pacview which is one of the leading vendors in retail media, commerce media. And it's a, it should be a good conversation because they have such scale that they actually have some data about what's going on and how people are using it and stuff like that. So Eric, you're involved with them, right?
- [07:00]
C
Yeah, I recently joined the board of pacview early this year. It is a really interesting company. To your point, they're kind of a, I would call them an OS for retail media. They sit across all of the different retail media platforms. They have a lot of data and they represent a really significant amount, amount of the total commerce media market. So I think Melissa's got really, really good insights. She's going to burn.
- [07:25]
B
And what's the story about your shoes and pack you.
- [07:28]
C
Yes, you know, I have to get some high quality pictures and post them. So I have a custom designed pair of Nikes that I wore at CES that had a perium branding and then the person that you know, helped me put together those added a pack view logo to them. So I wore them to their company kickoff when I was Introduced as a board member, the new board of Pack View and it was a total hit. Anybody wants like custom sneakers design, I have a connect and these things go over really well.
- [07:59]
B
Yeah. Why go to London to Seville Road to get those custom leather loafers. You go to go to Staten island and get some custom Nikes.
- [08:08]
C
Hit up your boy.
- [08:09]
B
All right.
- [08:09]
C
Anyway, this is going to be a fun one.
- [08:11]
B
Okay. Yep. Without further ado, let's bring Melissa onto the pod and should be an interesting conversation. Melissa, thank you for being here.
- [08:21]
D
Thanks for having me.
- [08:22]
B
So retail media is the hot thing. We've talked about it on this, on this podcast number of times. I guess to start you prefer retail media or commerce media. That's all. That's a big debate.
- [08:32]
D
Commerce media. Because it's not commerce media. It's also discovery. It's at retailers and it's agentic.
- [08:39]
B
It's agentic. Is it? Because I keep hearing agentic commerce is an illusion. It's not going to happen.
- [08:45]
D
It'll happen.
- [08:46]
B
Okay. Have you bought anything using AI?
- [08:48]
D
I've discovered using AI.
- [08:51]
B
Sure. I have as well. Definitely. I think I talked about this at our event two weeks ago. People say agentic and everyone freaks out and doesn't know. Everyone has their shape of the elephant on that and really drives people a little bit crazy.
- [09:07]
D
I think what we see today is not what we're going to see tomorrow. We're seeing the start of today, 2020. Nobody. You know, that was the start of Repel media when Covid happened and everything moved online. And I've heard people talk about 2026 is 202020 on steroids. So the model is changing and I think that what we have today will be very different a year from now.
- [09:32]
B
To correct the right. I mean I was putting banner ads on. On staples.com back in, I don't know, 2004 or something like that. So there's been retail media for a while. It very good.
- [09:43]
D
True or. And there also weren't a lot of eyeballs going there.
- [09:47]
B
Sure, that's. That's true. So tell us about quickly about PacVue. Like probably a lot of folks are not as familiar with it. Eric, you're on the board, so maybe you should do the pitch. Right.
- [09:57]
C
But you want the pitch from Melissa. You want the pitch from Melissa?
- [10:01]
B
Let us know what is pacview?
- [10:03]
D
So pacvia is a commerce operating system. And basically it's a place where people can go buy, purchase, measure and optimize all their media and commerce. And the best way to think about it is instead of having to log into Staples and Amazon and Walmart and Instacart and all these different UIs. You can log into one place, which is PACU, and do it all there.
- [10:27]
B
Right. Where's the name come from? It sounds like a sportswear brand.
- [10:33]
D
Isn't that the real story? You want the marketing?
- [10:35]
B
Yeah, yeah, we want the real story.
- [10:38]
D
So Xiaowei, my, my partner bought the name for $7 like 20 years ago.
- [10:44]
B
Okay.
- [10:45]
D
That and Paffy's like 8 years old. He said, you know, needs to be a very simple name, two syllables. And the honest truth is his office sat outside Picard, the Picar building in Bellevue. So P, A C and then we live in Bellevue. So View. There you go. Pack, View.
- [11:06]
B
I got it. Okay.
- [11:08]
D
Marketing story. So what's the marketing story?
- [11:11]
B
Is there a good marketing story?
- [11:13]
D
You know, we, it's got the view into the future. There's all kinds of things at View that you can have in store.
- [11:20]
C
Eric, one question just to. Yeah, yeah, for, for sense of scale. Melissa, what percentage of retail media buying does packv represent?
- [11:28]
D
Because you're across all these platforms, a very high one. We're, we have about $8 billion flowing through our, our system. So I would say a significant portion of retail media goes for us. And we are one of Amazon's top tech partners in addition to all the other ones.
- [11:50]
B
All right, that helps kind of put into perspective a great deal. So how big a problem is this for brands to get their hands around retail media? Obviously retail is very fragmented, but brands don't work with every retailer. Right. So what's the brand perspective on that?
- [12:06]
D
I think the reality is whether you work with one retailer, two retailers, or 20, you don't want to be logging into 20 different platforms. So the, the problem with fragmentation is really where we thrive. And so whether you're working with Amazon, even Critio Citrus, like this whole supply side platform kind of network, we also integrate with all of those. We integrate with over 1200 APIs, connecting all the types. So I think from a brand perspective, this really simplifies workflows. It simplifies optimization and activation for them. And they can see everything in one place. So it makes our lives a lot easier.
- [12:44]
B
Do you have any stats on the number of retailers the average brand buys from?
- [12:48]
D
Yeah, on average it's 5 to 6, so 5.2 to be exact. In Europe it's a lot higher. It's over 8 because Europe is even more fragmented with all the different geographies. And so, and then we did some kind of cohort Analysis where we looked at top 100 brands. So like big enterprise brands are on more platforms. The, the mid tail are on about 2 to 3. So and then it's so specific to categories. If you're a beauty brand you're on Sephora and Ulta and Amazon and if you're a CE brand you're on Best Buy. And so it's really very different by kind of the category that you sit in. And so that's kind of the reality.
- [13:32]
B
And then when we look at like the total size of the commerce media market, you have something like let's say Doordash where you might have a restaurant there. And that's a totally different market. The restaurant promoting a delivery has nothing to do with kind of the rest of the market.
- [13:45]
D
Exactly. And especially if you're in cbg, there's a lot more options. There's doordash, Instacart, you know, you name it. So if you're, you're in different categories, you're probably on that tips the different scale of all the different retailers that you might be interested in moving on.
- [14:00]
B
Right. But Amazon is the gorilla here. Like any, is there any estimate for what portion of the retail media business Amazon represents?
- [14:08]
D
Amazon has about 80% they're like a $60 billion and then the next closest one, which honestly is pretty amazing in itself is Walmart at 6 billion. These are all emarketers numbers but Walmart's been very impressive and what they've been able to accomplish. And then there is a long tail. But that's also, you know, people want to be where their customers are so they want to be on all these other retailer platforms. It's just we make it easier for them to buy it because of, they want to be on a long tail. But it's just, it's, it's a hard thing to be. But that's why you know, a thing like Pathfee really helps with all the fragmentation.
- [14:47]
B
Sure. And 6 so Amazon 60 billion but some portion of that's not really retail media. So the DSP business and you know, the fire TV business, they're sort of retail media. It depends.
- [14:59]
D
I call it retail media. It's a retailer.
- [15:03]
B
Sure. But it's also a cloud computing company. So it's cloud computing advertising. So let's talk about like what it is. This kind of speaks to the question I was just saying which is, you know, there's kind of the market for what I would call like product listing ads is kind of different from everything else. Right. Because like the, the Kinds of ways you target is very limited. The products that you advertise are kind of, it's almost always click in versus let's say a CTV off site campaign is pretty different. So tell us about the mix there and how it's evolving.
- [15:39]
D
Well, one I think it's important to bring up is prompts are the new keyword. So things are kind of moving towards. Amazon has launched sponsored prompts. You can actually buy on prompts now. Not just keywords, there is also contextual and audience based targeting and that's really the kind of really cool stuff. Amazon is very mature and it's syncing with launching Amazon marketing cloud and the ability to create very specific audiences. They're also launching some pretty cool tech around contextual shopping behavior. What you bought before, what you're trying to buy in the future. And so that's. I think that there's some new models emerging around that.
- [16:23]
B
Right. So are sponsored prompts the first at scale AI advertising? Because we know OpenAI is just barely starting.
- [16:32]
D
Yeah, I mean that's. Have you, have you been playing around with Rufus at all?
- [16:37]
B
I used Rufus on a product purchase recently. It was about as useful as like, you know, searching through the comments and the, and the questions. It was like a time saver, I would say. What's your opinion of Rufus?
- [16:51]
D
I, I use it a lot and I'm just not as good of a searcher as some people. But I'll give, I'll give you a good example. My car is electric and I use my, I have a BMW. My husband has a Tesla Charger and BMW, BMW did some update and then my, the charger didn't work. My car. And so then I searched Rufus and I said, I have this car. I'm trying to use a Tesla Charger. Send me a new charger and it gave me the right one. So I feel it would take me a lot longer to find that charger if I had just used Amazon as like a BMW, blah, blah, blah, blah, you know, keyword search.
- [17:26]
B
Sure. Good. Good boy, Rufus. Good job. Eric. You use Rufus. Are you a Rufus fan?
- [17:31]
C
Yeah. It's great. Especially like the more niche and difficult like the situation that you have. Like I found it to be really good. I have a similar strangely automotive story. I was changing the, the air caps in on my wife's car and it was very cold. It was like when, when it was like really, really cold here in the northeast and on two of the four wheels I dropped them in the wheel well because just my hands were shaking. It was so cold when I was trying to put air in the tires and very quickly like just took a picture, uploaded it, said, you know, you're making model. And then I got the item like the next day.
- [18:07]
B
Nice. Yeah, I think Home Depot even has an ad saying that their AI will help you find a product, which is pretty nice. But anyway, back to, back to the subject of commerce media. So, you know, I interviewed Jeff Green on stage two weeks ago. We'll talk more about Trade Desk later. But one interesting thing is I asked him specifically about whether they're entering retail media. And I asked him, are you going to support product listing ads? Are you going to support targeting by keyword, those sort of things? And he didn't really answer that question. He answered his answer effectively, said, you know, we'll be anywhere there's high quality inventory. You know, sort of praising the inventory available to retailers but not talking about specifics. Do you have any intel on that, whether the Trade Desk is getting into this business or not?
- [18:50]
D
Well, I mean the way that the Trade Desk is getting into retail media is through partnering, providing the DSP of everybody else besides Amazon. I mean like that's how everyone's bringing a DSP to market is by initially partnering with, with Trade Desk, but it's on the DSP side.
- [19:03]
B
So how much desire is there on the brand side for these kind of off site deals or audience extension kind of deals? And where does the budget come from? Is it the same or different budget?
- [19:12]
D
Well, one, what I just said is kind of a little bit prohibitive is on the minimum side. I think there's demand, it's an audience play and people want to test and learn, but it makes it hard to test and learn when you have minimums in place. And so the encouragement is to make less, less minimums.
- [19:34]
B
Right. This, this isn't in the, directly in the notes that we shared, but I'm wondering if you have an opinion about the Shop app from Shopify and how they're advertising on it or trying to get the merchants to advertise on it.
- [19:46]
D
I have less of an opinion there. To shop, we talk to everybody.
- [19:52]
B
So yeah, sure. Eric, have you heard anything about Shopify and the Shop app?
- [19:58]
C
Yeah, it was. Do we cover it? I think we covered it.
- [20:01]
B
Well, they announced, they announced a product recommendation ad network basically. But they also have this app that they're trying to encourage people to use to purchase things. I'm just kind of trying to figure out if anyone's using it.
- [20:12]
C
Yeah, as a user I have experience with it where you Will like, you know, check on the status of something that's being shipped to you and you'll get like relevant cross promotional offers. It seems to be a little bit more kind of like commerce than ad, but it's good and it's extremely personalized. I think it's a good first step. Shopify can build a beast of a advertising business should they want, using not just their own O and O app, but the massive footprint of, you know, most independent retailers. It's just a matter of, you know, do they want to prioritize, make the investment and do it?
- [20:45]
B
Exactly, yeah. They don't seem to have prioritized to date. So. So a lot of times when we're talking about shopper media or retail media, commerce media, whatever it is, they, it's like commerc. Well, well, we got this great stuff, right. You know, the consumers giving us intent and all that. But what we're really excited about is all these expansion opportunities. We talked about the DSP off site opportunity, but there's also in store digital at home. There's. There's just kind of this idea of non endemic, like we could be showing ads for insurance on retailer sites. How much of that is reality or is it still like all a bunch of experiments?
- [21:22]
D
I think that, you know, in store it. A lot of most of the purchases are happening in store and so Walmart and others are really investing in it. The other thing is it's, it's an advantage against Amazon. Amazon doesn't have an in store, well, I guess Whole Foods. But you know, it's kind of an advantage to have the whole closed loop within store purchasing. And so that's really interesting to have a very comprehensive, you know, maybe I bought it on online, I found it in store. And so having that closed loop online offline kind of reporting measurement and capability and targeting is pretty compelling.
- [22:00]
B
So I think is that media being bought in a programmatic way or an online way?
- [22:05]
D
It's, it's being bought in a decentralized way. So I think that's through the kind of retailer themselves. And that's actually kind of the problem with our space is that there's a very fragmented way that people have to buy within a retailer. So it may be a managed service for some things it may be, it's not pro. The only programmatic role play is search and display.
- [22:32]
C
It seems to be a very hairy measurement problem.
- [22:35]
B
Right.
- [22:35]
C
Like it almost, it means you almost have to have, you know, on a retailer by retailer basis to solve the closed loop attribution question. Seems to be pretty difficult to have a layer on top that you know is like programmatic access to everything in store because you need the retailer by retailer data.
- [22:55]
B
And so you recently talked, spoke to, I guess Kevl on a podcast about incrementality. So how does incrementality happen here with given Eric's point about it all being retailer by retailer, we had, I interviewed the head of Walmart Connect around Christmas time and they were just putting out incrementality for the first time. So are you dependent on the incrementality frameworks of each retailer or can you do it yourself?
- [23:20]
D
Through Packvue, we've built our own incrementality model. And so that we have our own, we also partner with people like incremental. Not the one that just got bought. A different incremental.
- [23:31]
B
There's two incrementals. Yeah, they're spelled differently.
- [23:34]
D
Hopefully they'll who bought them because then that would make everyone's lives a lot easier. But incrementality with the rise of all these retail media networks and people want profitable incremental sales growth. That's why it's such a critical factor. But the reality is the modeling is, it's not, it's not easy. And so bringing to market an incremental model is one that, yeah, so we, we have our own, but we also.
- [24:03]
B
How do you do it? Like, is it just modeled or their ability to do real holdouts? Because like most of the retail media world doesn't have consistent identity. Right.
- [24:12]
D
I mean, that's, that's why it's so challenging. And there's a lot of maybe buzz or marketing speak around it, but I think that you just have to go with one model and go with it. But that's, that's, it's a complex modeling exercise, that's for sure.
- [24:31]
C
Yeah.
- [24:32]
B
So I have to ask, is retail media incremental?
- [24:34]
D
For most advertisers, it's, it's a hard one to answer. I think that it's, it's really specific to the brand. But the reality is, you know, there's discovering something and there's buying something. And so your influence, you may find something on TikTok, but you buy it on Amazon. And that's like there's, there's not. How do you put that into an incremental model? And so the reality is, I've heard this, that Amazon throws it out. I don't know if it's true, but it takes and this is average 20 different touch points before you buy a brand. And that Touch start on Tick tock. It may go to Reddit, it may go to, you know, various different places and then you wind up converting so you know it. That's how do you measure that and is that incremental? You still have to be on TikTok to discover it.
- [25:28]
B
Sure, yeah. I, I the, the dig on retail media sometimes like well consumer shopping for your product anyway and then the retailer shoves an ad right in their face for a competitor and now you're either paying for the search you already deserved or you're getting taxed to, to get someone to buy the product they were going to buy anyway. Right. The old brand keyword versus versus broad keyword problem. So so lastly let's talk about what is going on with AI. So so how are brands thinking about this problem with in the retail media space and how are retailers thinking about it? We talked about Rufus Agentic a little bit. What's, what's the cutting edge of what's happening here?
- [26:05]
D
The way that I like to think about it is I kind of segment this into three different areas. One is discovery commerce and that's you know, the TikTok, the Reddits and Pinterest of the world where people are discovering. There's retailers where people are also discovering and buying. And then there's this third pillar of agentic commerce which is all the ChatGPTs and the LLMs. I would say that's still a place where people are discovering and it's all about how are you showing up and ranking. There's and we know that Reddit's a place that, that pulls from so that's one of the reasons why we, we launched Reddit as the platform. But I think that's a good framing of each of those different places and each of them have a different playbook. But the, the part about the agentic commerce piece, I don't think people are actually we saw instant checkout get pulled from OpenAI that didn't convert or work very well. But now integration of Sparky is kind of the new way of thinking. So agent to agent Sparky is Walmart's
- [27:06]
B
version of Rufus which you know, I'm still upset they didn't call it Wall E but the, yeah, so that was a recent announcement. So I mean OpenAI is moving so fast it's hard to see the trends for all the static of the day to day announcements. But separately we, we know that Critio announced they have some sort of deal with OpenAI which we believe is, you know, product listing ads of some Kind right on the buy side. They're sort of competitive with you. So like that sounds like they do want products and, and ads that are product related. So you can check out. Am I misunderstanding?
- [27:42]
D
Critio hasn't said exactly what this experience is. And so I think, I think jury's out Credit has a partner. I wouldn't call him a competitor at all. But the, I think the reality is there's a lot of spaghetti being thrown at the ceiling and we're seeing what's sticking. And so what. What's today may not be tomorrow, but I think the important thing is, is being curious and learning what's going on and partnering with all these people so we can figure it out together. And so I think that's the, that's where we're living is that the models are changing really quickly. And so we'll see one thing today, a different thing tomorrow. But the reality is, is that people are going to OpenAI to research products. And so that is a place where brands should be. Because if you're searching, you know, different, you know, consumer electronic products and then the agent can recommend batteries that you need to go with it, you know, that's, that's super helpful. And so I think the starting with what the consumer journey is and what they want to have happen and have their life easier and then how do we create products around it that makes that convert?
- [28:50]
B
Right. But you haven't announced any AI integrations other than the Amazon one, which is sort of part of your Amazon integration.
- [28:57]
D
We have not announced anything.
- [29:00]
B
All right, that's a very careful answer. So there's a great spot to take a break and we have so much news that we'll come back to. They'll touch on many of these issues and related issues and more issues that you probably already know about. So let's take a break. And we're back.
- [29:17]
C
All right, we're back, everybody, with the refresh. Oh my God.
- [29:22]
B
Drama.
- [29:23]
C
Where do we begin?
- [29:24]
B
So much drama. I can't take it, man. I need to retire. I don't, I don't want any more drama. I want, like, I got into ad tech to avoid the drama.
- [29:34]
C
So last week and we'll, you know, we talked about this in the, in The Intro Architecture Live 3. It was amazing. You closed out the show with a live chat with Jeff Green. It was great. He clarified a lot of things from his perspective. He dropped something mid conversation that was super juicy, that effectively TTD is in beta with an AI solution. So basically some number of customers are going to Be able to build campaigns
- [30:03]
E
using Cloud this episode is brought to you by Indeed. Stop waiting around for the perfect candidate. Instead, use Indeed Sponsored Jobs to find the right people with the right skills fast. It's a simple way to make sure your listing is the first candidate. C According to Indeed data, Sponsored Jobs have four times more applicants than non sponsored Jobs. So go build your dream team today with Indeed. Get a $75 sponsored job credit at Indeed.com podcast.
- [30:30]
C
Terms and conditions apply to Deploy on TTV.
- [30:35]
B
Yep.
- [30:36]
C
Q. Unpack that a little bit from your understanding, because it was basically it and then that was it.
- [30:40]
B
Yeah, well, yeah, it sounded like maybe he's not supposed to say it. It was the first time they've announced anything he didn't specifically say. He didn't say Claude specifically. I think I asked Claude as for instance and he said yes, basically. Gotcha.
- [30:53]
C
Gotcha.
- [30:54]
B
So I mean, this isn't surprising. It's kind of surprising that it's taken so long because they have an API, it's used by lots of customers. If you have an API, then the next step of enabling it through MCP isn't that big a step. And so that I assume, is what he's doing, what they're doing. Trade Desk and we'll see when it gets announced what's going on. I mean, I did a little wrap up in my newsletter and it's interesting. I did a little Google searching of DSP name +mcp server and a couple of them came up. Adform has something and Freewheel announced something at the event last week. But a lot of them don't. A lot of them don't have anything publicly available right now, which is really kind of, I have to say, it's behind, like it's just behind where the customers are. It's surprising. So the Trade Desk is in beta. Probably a lot of other ad tech vendors are in beta or testing things. It'd be crazy if you weren't investing in this area.
- [31:52]
C
Yeah. How would you envision? Let's just use the Claude example and what was stated and in the ad week recap of building campaigns using CLAUDE and then pushing them into the trade desk. What do you envision that being and how is it different and incremental to building campaigns natively in TTD or other DSPs today.
- [32:14]
B
Sure. Well, creating a campaign has just quite a bit of busy work typing and clicking and moving things around and copying things and if you done it before, you can do it again. Like. And so AIs are good at that sort of Automation. So you can have, I would think maybe you have a spreadsheet that you've developed through your planning process that says like the kind of audiences and the kind of, and the dates and the flights and the creatives and the tags from, from an ad server and all that sort of stuff. And you can give that to Claude and say based on this and the, this MCP server that we think exists from Trade Desk, you know, set up my line items, campaigns and targeting to the best of your ability. Report back where there are any problems, don't set anything live under any circumstances and let us know when you're done. Right. And you've now saved like five hours of someone's life.
- [33:07]
C
Yeah, that makes perfect sense.
- [33:09]
B
And that's not that hard.
- [33:10]
D
Oh, it's the executioner that can be outsourced to Claude pretty easily. We've been looking at that too. And the question is, how many tokens does it take in the past? But instead of having a person do this work, you have your agent do it and then it can push it back into the application.
- [33:28]
B
Right. And so one thing I wrote about in my newsletter this week, which I think was kind of funny because one of the highlights of the Mark Destroyer Live event was that Pubmatic and MIQ did this amazing demo of using Claude to create campaigns and they pitched it as like ad cp, end to end workflow for creating campaigns. But like, it kind of dawned on me a little bit afterwards that it was really a single hop. It was going from Claude to Pubmatic. There was no, there was no. Meiq didn't have bring any tech into that demo. They were the customer because Pubmatic has a buy side solution as well. So basically they were just booking orders in the buy side of Pubmatic. It was only one hop. It wasn't the two way negotiated thing that Brian o' Kelly and others have been saying is the future is going to replace RTB and all that stuff. We didn't see a demo of that, even though there have been some demos of that. But that's not what was demoed out on the stage. And doing the one hop thing is just an order of magnitude simpler.
- [34:24]
C
Okay, this is what we talk about. I don't know what more there is to be said. The fighting on LinkedIn over this topic we're about to cover has been insane.
- [34:38]
B
So it's like broken through to the normies. Like there are normal people who have not have never met before or add like finance analysts who are now posting about this, this. Anyway, go ahead.
- [34:48]
C
Yes. So There was a memo that was leaked to Adweek who had the exclusive from Publicis stating that they are advising their clients not to use the trade desk. Publicis in. In this memo. This is all allegedly a third party audit. Allegedly found that the trade desk improperly applied fees, billed clients for tools they were automatically opted into without authorization and failed to provide sufficient transparency on media and data costs.
- [35:28]
B
The word allegedly is doing a lot of work there. We. I mean, unless we think Adweek is misreporting this. There is a memo and it does say that whether it's true, what it said or not is alleged, I guess.
- [35:39]
C
Exactly. Exactly. I just want to be careful. And TTD has basically denied everything, which is why I think the word allegedly is important.
- [35:48]
B
TTD has denied that they did not pass an audit by the company that's listed as the auditor in the Adweek story. Which is a Ubiquity is the honor.
- [35:58]
C
It's a division of ubiquity.
- [36:01]
B
Okay.
- [36:02]
C
Correct. Yep, correct. Apparently a well respected media auditor name escapes me. Everybody's talking about it. Why is this important? This is important for a few things. So number one, it was only a few years ago. Five years ago. A few years ago, Publicis named TTD the exclusive DSP for Epsilon Core id. Publicis, according to some of the publicly available information, represents 10% of TTD's gross billings. So it's the largest Holdco client. And this is just a few weeks after Dentsu and WPP quietly quote, unquote, exited OpenPath. This is a big deal.
- [36:39]
B
Yeah, it's definitely a big deal. It's like an avalanche of bad news for the trade desk. So it's at the worst possible time to have more bad news. It's been nothing but bad news for six months for this company. And there's like the. There's the story and then there's the sub text. It's similar to the the previous story, the Dentsu and WPB exiting open path that immediately had a subtext, which was maybe they exited openpath because they didn't like the transparency. They wanted to have side deals, SPOs, kickbacks, benefits, etc. And there's a general feeling like the trade desk has made some enemies over the past year or two. And maybe people are just taking the opportunity to use the press to take potshots at them to get better deals, better data, better. Whatever it is.
- [37:25]
C
This seems to be the same narrative on this one, which is supposedly this is a punch back to all of the work that FTD has done, particularly with open path to make things much more transparent, harder to run principal buying divisions.
- [37:43]
B
Yeah. Let me address the two main issues alleged in the memo. The first one was this idea that the DSP fee was applied to other things. And this is really complicated. I might do a newsletter about this because let's say you have a 10% DSP fee fee. You think that's only being charged on the media, that you're buying from the publisher. But what happens when you throw, you know, a 50 cent data segment into the mix? Is it, is there another 5 cents being taken on that 50 cent data fee? And also what about the viewability fee, which is sometimes viewability, sometimes above the line, not a bid reduction. It's usually sometimes an additional line on the invoice. Are you taking 10% of that? And then there's a difference between whether it's your data or a third party's data. Like if you go into the trade desk interface and buy data from a third party, the trade desk may take a fee on that. If it's the trade desk data, they probably shouldn't take a fee because it's already in there and the pricing right is their data. So there's a lot of complexity here. Probably DSP billing is very hard to understand.
- [38:49]
C
That was one of the best parts of your conversation with Jeff, by the way. It almost kind of like fortunately foreshadowed, foretold this because you were talking about the minutiae of DSP invoice and you can get a 30 page invoice when clients generally just want one. But now do clients ask you 30?
- [39:07]
B
Yeah, we had a chuckle on stage because we both agreed that we should give less data on our invoices to customers. And then it turns out that this happened just a couple days later. So the second thing that they're being sort of accused of is this idea that there are all kinds of extra charges that are being automatically opted into in their platform. I've heard this for years and years and years that like, unless you're careful in the trade desk, suddenly you get, you, you, you bought something. All right, it's just like they, the undercarriage protection just kind of shows up on your invoices. And the worst offender there is the so called bid shading where they use an algorithm to try to reduce your bids. They then take a percentage of the savings on that, which could be a very significant and very opaque number because they have to calculate it themselves. They're like, oh, we would have bid $5, but instead we're bidding $4. So we saved you a dollar, so we're going to charge 10% of that. Right. You're like, well, why didn't you just bid $4 in the first place? Right. Well, because you had to pay us to get a better bid. You could imagine customers not really loving that.
- [40:11]
C
Yeah. And this is, this is my question. I wanted to bring this up. I think, think other people might have as well, for your perspective when. So you have what's written in your msa and then you have in practice a lot of people across publicist agencies that are building campaigns that may not be even aware of the msa. And maybe they're just like, yes, you know, fine, I don't want to opt out from this, or I do want to opt in on this. And is there a theme, Is there a. A DSP game of telephone that could be part of this? Just because when the rubber hits the road, it's people making a lot of these decisions still within ui.
- [40:53]
B
Yeah. I mean, I think the bid shading is a great example where if you don't want to pay for it, if your corporate policy is not to pay for it, one option is to put that in the contract, but more likely it's not in the contract to begin with. It's a checkbox in the ui. So then you have to ask your product team, like, hey, can we have a default opt out of that of every publicity account? And then one account wants it and you end up with this, like, crazy complicated system settings system. So if it's there and some. Someone clicks on it, gets better results. Suddenly they click on it on every campaign and you're being charged for things that you didn't intend to be charged for. It's just a very complicated product. Right. And there's a lot of things you could charge for, and it's pretty hard to maintain those costs in, in that environment. And so I have a lot of sympathy for both sides on this.
- [41:43]
C
Yeah, exactly. Where do you think this goes?
- [41:47]
B
Well, I want to see the next earnings statement. I guess that would be kind of the most important next piece of data from a financial perspective. My guess is this is a recommendation.
- [41:56]
C
This is a recommendation to clients. This isn't pulling, you know, 10% of TTD gross billings.
- [42:03]
B
Yeah. I mean, the trade desk has not done great in working with the industry and in its communications over the past six. They seem to be more combative than acting like a vendor who needs to make people happy. If I were them, I would immediately do some transparency pledge and I would explain My billing and I would document it and I would just be on top of it and try to win back this argument because after all they are the customer. Right. They could choose whatever DSP they want.
- [42:31]
C
Yeah, yeah, that is a great idea and I hope they're considering that.
- [42:35]
D
That.
- [42:36]
B
Sorry, Melissa, we just rat holed on that. You're just like checking your email and stuff.
- [42:43]
D
I'm, that's what I'm doing. Getting. I think that's a great idea. But it's been, it's been interesting to watch on LinkedIn how vocal both sides are without. But I like your recommendation. I'd like to see that. But I don't, I don't think that's going to happen.
- [42:55]
B
I don't think it's going to happen either.
- [42:57]
C
It is the digital coliseum these days. Just people fighting on LinkedIn. Who knew a professional network.
- [43:06]
B
Yep.
- [43:07]
C
Okay, let's, let's, let's depart. Let's talk about a little M and A. So publicist. Also in the news, they're buying a creative AI company called adj AI A D G E A I familiar with AJ AI Car.
- [43:21]
B
No, I'm not.
- [43:23]
C
Yeah, no, me either. There's, there's several businesses like this. We were a investor in this company called Memorable, which Reddit acquired, which is in this family, which is basically like predictive data on Creative, which AI is like an amazing use case for. So the idea is to make your creative that much better in the beginning using all these signals and then optimize it over time. Seems to be like a no brainer for publicist or any agency to have tools like this.
- [43:52]
B
Yeah, so many AI companies were developed with two or three people and a couple prompts and they were pretty simplistic and they're not scaling. So there's a kind of open season for Aqua Hires. I know of another Aqua Hire that just happened in the similar or maybe a little bigger has been announced but you know, established ad tech company buying team of three to five people doing gen AI stuff.
- [44:13]
C
Yeah, creative. Yeah, yeah, it makes sense. It makes sense. This is another one that this was worded strangely so smartly which is like a creative intelligence company. Creative intelligence.
- [44:28]
B
They're more. Melissa, are they competitive with you nowadays? They like buy a lot of media. I thought smartly.
- [44:33]
D
I don't think so.
- [44:36]
B
They started as a creative. They started to serve as a templated Facebook creative like a canva for social media and then that model was not working. So they evolved and now they have a pretty robust kind of media placement buying optimization Thing that they do. CEO is Laura Desmond, who's been around our space, was CEO of Publicis, I think for a while.
- [44:58]
C
She was speaking of describes itself as an AI powered creative and media orchestration platform. So closer to what you were saying than what I was saying. So it announced that it signed an LOI to letter of intent to acquire incremental. So not the incremental with vowels. It's an incremental without vowels. No vowels with the exception of an A at the end.
- [45:23]
B
They're Israeli company.
- [45:25]
C
Yeah, yeah, cool company. I've met the founder. So it's an incrementality measurement startup being acquired by a creative and media orchestration platform. This makes a lot of sense, right? Like end to end, create, deploy, understand the incrementality, make it better and better each time.
- [45:43]
B
Yeah, I'm an investor in that through a venture capital fund and they are the number one advertiser on Eric Sufer's podcast that we sold the ads to. So we're a customer of architecture. So we're happy about that. That and I think the only thing to note here is incremental is very much an in app install kind of play. So that kind of gives you a sense of it maybe smartly is going to spend a little more time on that.
- [46:06]
C
Yeah, that makes a lot of sense. And it's just strange because normally these things get announced when they are closed, but this one is at the letter of intent stage, which is far earlier than a close.
- [46:20]
B
Could be an Israeli thing. Israeli acquisitions sometimes have weird regulatory things.
- [46:26]
C
Yeah, good point. Or it could be a scoop. Okay, maybe we get into something for Melissa. So walmart OpenAI shaking up this Agentix shopping deal. So there was big news. Or Walmart was part of OpenAI's instant checkout feature in ChatGPT. Apparently that didn't go that great. People were having weird things show up in their cart. Wasn't connecting to Walmart's data generally not a great shopping experience. So instead it's embedding Sparky directly into ChatGPT. So they own the customer data, they own the experience. They can handle checkout on their end. What is your sense on this whole thing? Is this now putting a real wrinkle into agentic shopping? Is this just like, you know, hey, early examples didn't work. We'll retool and come back. What's your take here?
- [47:18]
D
I think part of it is one none of the retailers really wanted that anyway. The partnership with OpenAI where the transaction happens there, lose all the data. So this is Way better for the retailers. And when I could see Rufus being happy to be there, whereas they, you know, Amazon's been a holdout with OpenAI, but because they want all the transaction and data to happen on the retailer site. I think it's a couple things. I mean, one is people weren't buying on the platform. It wasn't a real marketplace. You know, to do that's really hard. And so it's kind of that fail fast and pivot mentality that I think is really important. Maybe this comes back, you know, in a year from now and they actually do it. But this is definitely the way retailers want things to happen. They don't want to give up the data and the customer. They want that to happen on their site. So this is, this is way better win win partnership than it was before.
- [48:10]
B
Isn't this a little bit more like what Google announced with ucp, where the retailer controls the checkout?
- [48:15]
D
I feel like that's a little bit different on the Google side, what they're doing there.
- [48:20]
B
Sure. I know that with the sparky thing, they're talking about a chatbot inside a chatbot, which is a little weird, but you know, whatever works. Right. Whereas I think the UCP was much more about checkout itself. Like it was just like the end result checkout you can control remotely.
- [48:39]
D
Yeah, exactly. I think that's the interesting thing about this space is that it's changing on the daily. And so part of it is how do you play in all these places and what are the Playbooks? And the Playbooks are different in each of them. And so that's kind of the exciting times that we're living in right now.
- [49:00]
C
Yeah. And I think the business implications are pretty significant. Again, all this stuff is so early and happening so fast. OpenAI went from having a potential share of commerce revenue and potentially, you know, if you model it out, I'm sure they did like significant share of commerce revenue to now. Like maybe they get some fees for the distribution that they're providing, but not a cut of, you know, every transaction that might have happened on ChatGPT via, via Walmart. So very, very fast changing space.
- [49:30]
D
So, speaking of question for you guys, do you think it's better to be first or a fast follower as a brand kind of activating in these places since it's changing so fast. So like, should you go and be an advertiser in OpenAI first or should you wait and be second?
- [49:46]
B
It's a good question. I think it depends on the investment level. I think the media dollars Aren't don't matter that much. It's whether you have to do any kind of work, any sort of, you know, integration, landing pages, changing the way you operate to participate, that would make me think you'd be better off waiting. But if you just have to write a check, then why not?
- [50:07]
C
Yeah, I'll take the other side of it. Presumably, if it's just as simple as like being part of the initial rollout of a brand new ad product. If it's something as significant as the first ads on ChatGPT, I think there's real advantage to being there. Getting some of the PR value, getting some of the data, getting a good understanding of this stuff and then growing from there. I'm a big fan of this and as, as you both know, every brand has innovation but just for this, this intended experience, to be first, to get the learnings, to be part of these early programs, I think it's a, it's a well trodden path. Speaking of, of failing, but maybe not failing fast. Let's pour one out for the Metaverse. So Meta, formerly known as Facebook.
- [50:54]
B
Yeah.
- [50:55]
C
Changed their name to Meta because of an. An effectively what became an $80 billion bet listeners. Ari just poured out maybe some water for the Metaverse malt liquor. So Meta seems to have pulled the plug on the VR strategy related to the Metaverse, which I think was seemingly everything with the acquisitions they did and all the investment they did and we're going to be focusing on mobile and AR instead. I think this was a massive investment that went nowhere. $80 billion and then the other side. I mean good for Zuck, just like taking the L, moving on, focusing on AI and continuing to I think deliver some results. That guy is not afraid to fail in public.
- [51:47]
B
He runs the company like it's his corner grocery store and he could just do whatever he wants. He's like, let's put out the mints today. No, let's take the mints back. Oh, that cost us 80 billion DOL dollars too much. Oh well, we'll sell more mints tomorrow. Like he just does whatever he wants. It's amazing.
- [52:01]
C
Here's my question. So somebody on X posted this and I, and I'm sorry, I can't credit them. I don't have the, the, the, the screen name but they reminded me of the fact that people were during crazy times. Covid crazy times. Buying real estate in the metaverse.
- [52:19]
B
Sure.
- [52:20]
C
And there was an article that linked to a person spending almost half a million dollars to get real estate next to Snoop Dogg. In the metaverse. What happens to that?
- [52:30]
B
I'm sure. I'm sure it's on the blockchain somewhere. He's got proof of it.
- [52:36]
C
I think so. Sorry. For metaverse real estate purposes, Maybe that's a good place to end it. Ari, what do you think?
- [52:45]
B
Yeah, I think that's probably. We've done our. We've done our duty here. Well, there's a great conversation. Melissa Bur from pacview, thank you so much for joining us. We loved hearing about what you're up to, the origin of your company name, and, you know, your perspective on all this stuff. So thanks again.
- [53:03]
D
Thanks for having me.
- [53:04]
C
Thanks, Melissa. Take care. See you next week, everybody.
- [53:08]
B
Thank you for subscribing to Market. New interviews are added every week at Marketing and your favorite podcasting app.