Podcast: Digital Advertising’s M&A Resurgence (with Maor Sadra)

On this week’s episode of the podcast, I am joined by Maor Sadra, the co-founder and CEO of INCRMNTAL, to discuss the current cycle of M&A in digital advertising and the industry-wide shift toward intelligence-based measurement. Maor is well-positioned to discuss this topic, as INCRMNTAL was recently acquired by Smartly.

Maor and I analyze how acquisitions are moving away from simple inventory reach toward deeper integration of data, identity, and causal signal. Among other things, we discuss:

• Why causal measurement is becoming a strategic necessity for large-scale creative and media optimization platforms
• How the transition from legacy to AI-first companies creates opportunities for consolidation across the advertising stack
• Whether enterprise clients will prioritize deep domain expertise over general machine learning capabilities when selecting partners
• If the recent investment in AppsFlyer signals a permanent shift toward a collaborative industry-wide attribution infrastructure
• What the rise of zero-click platforms like ChatGPT means for traditional performance marketing metrics and models
• How e-commerce and retail media giants are leveraging CTV acquisitions to challenge established advertising platforms
• When the current wave of AI-labeled startups will face the reality of high-volume performance requirements

Thanks to the sponsors of this week’s episode of the Mobile Dev Memo podcast:

• ⁠⁠INCRMNTAL⁠⁠⁠. True attribution measures incrementality, always on.
• ⁠Branch⁠. Branch is an AI-powered MMP, connecting every paid, owned, and organic touchpoint so growth teams can see exactly where to put their dollars to bring users in the door and keep them coming back

Interested in sponsoring the Mobile Dev Memo podcast? Contact Mobile Dev Memo advertising.

The Mobile Dev Memo podcast is available on:

• Spotify
• YouTube
• Apple Podcasts

Transcript

Eric Seufert: Welcome to the Mobile Dev Memo podcast. I’m your host, Eric Seufert, and I’m joined today by Maor Sadra, who is appearing for the—I don’t even know how many times on the podcast. You’re saying sixth time. You’re certainly at the top of the leaderboard. Welcome back.

Maor Sadra: Thank you, Eric. It’s really nice to be back. Thanks for reaching out. I always love chatting with you regardless of the occasion.

ES: I wanted to bring you on to discuss this digital advertising M&A cycle that we are in at the moment. I thought you would be a very appropriate guest to weigh in on that topic, given that you just went through an M&A process. Congratulations. Before we dive into the conversation, maybe you could introduce yourself to the audience for those who didn’t listen to the previous five appearances.

MS: I’m Maor, CEO and co-founder of Incremental, even though now I’m getting a new role as part of the post-acquisition stuff. Incremental was acquired by Smartly for those who don’t know. Before this, I was CEO for AppLift, the performance network, not the ASO company who took AppLift’s deprecated name. I’ve been in ad tech and marketing for over 25 years, doing everything from product to customer-facing roles to sales. I really like this space and I’m fairly opinionated about it. I’m glad to be here and glad to have met you almost 15 years ago.

ES: I moved to Berlin in 2014, so it would have been 2014 or 2015. Not quite that old, but at least a decade. Smartly described the rationale for acquiring Incremental as connecting real-time causal measurement directly to creative and media optimization. Walk me through why Smartly was the right home for Incremental and what’s possible now as part of the Smartly machinery that wasn’t possible when Incremental was a standalone measurement company.

MS: We were fundraising last year and started speaking with VCs. One of the VCs we spoke with told us they had this company they invested in a while ago that got acquired and they’d like to make an introduction. That was Smartly. The more we understood them, we understood that this is what we were looking for. Measurement is a very easy sell when someone doesn’t have measurement. It’s not an easy sell when you are saying to someone that you have a new ruler that’s going to measure the truth, but it’s not actually here to replace the existing ruler you might be using, like attribution.

The other challenge with what Incremental was selling is the fact that we all saw who scored the goal. In the last World Cup, Torres scored the goal. Attribution loves to tell you that Torres deserves 100 percent credit, even though we all agree that’s not the case. The ruler Incremental was selling, or the measurement solution, was to tell you that’s true, but I’m not here to replace that truth. I’m just coming to give you a different truth to help you structure the team. That was always a relatively hard sell when you don’t know if someone is interested in buying that. Incentives play a very big part there. If a company incentivizes their people based on whoever is scoring the goal, they’re never going to want a solution that tells you differently.

Smartly has all the enterprise clients we could ever hope for. That’s pretty much the best fit for us. As founders, it gave us access to resources and clients. Smartly knows exactly which clients are interested in this because they work with close to a thousand clients who are all enterprise. That for us was the clear multiplier. The fit there is a one plus one equals three, and everyone wins from that.

ES: Do you think that’s true generally with measurement solutions?

MS: Depending on what you’re asking. I think, and hopefully we’re going to get to talk about AppsFlyer, it’s not always the same case. It depends on what sort of measurement we are talking about. Is this the referee or is this a measurement solution that gives a signal? That’s a big difference.

ES: That’s what I was alluding to. There’s a place for standalone measurement that doesn’t integrate with any other functionality.

MS: There are two types of measurement solutions. There is the referee that connects between demand and supply and acts as the one telling you who scored the goal. They have the mechanism for this. That’s pretty much what attribution does. Then there’s a measurement solution that is a signal that makes you smarter. There is a whole intelligence suite around it. I think that’s the type of measurement solution Incremental is. That’s a signal that’s valuable. If you are already in the marketing technology space and you’re selling a SaaS solution, then integrating a signal that allows your customers to be smarter is an advantage versus your competitors. In a world where you’re selling the media, that’s a little bit harder to offer an independent solution that enriches a signal because then you’re very much biased toward enriching your own signal and not anyone else’s signal.

ES: You’ve lived through several generations of digital advertising M&A. Earlier cycles tended to be about companies acquiring inventory or aggregating inventory reach. Now it seems like there has been a dedicated cycle for data, identity, measurement, and decisioning. Is AI the catalyst for that strategic change? Is it maturity in the space, a shift away from the open regime to a more closed regime? What’s driving that strategic shift into acquiring data, identity, and measurement versus just more inventory and reach?

MS: It is about intelligence. The tools of the past simply won’t survive the future unless they find how to make this shift. It’s like companies who failed at making the switch from web to mobile. A lot of companies just could not make the shift and are stuck somewhere in the past. I think most companies will fail at making this switch. I call it from legacy to AI, because I don’t even know how to call the old world. An AI-first company understands they need to let go of control and put trust in the model. They will fail many times. Debugging an AI system is a lot more hard than building software that does what it’s intended to do.

That’s one of the reasons why I think AI currently is not very public market friendly, because the market would penalize every time something happens. I think it’s a long-term investment. It’s the main reason why I think so many AI companies are keeping themselves as private at the moment, because you want to reach a point where the AI can actually grow itself and fix itself. For that, you need to adopt that mentality. You need to let go of control and not try to develop everything yourself because that’s not necessarily the scalable solution. Most of the acquisitions we’re seeing right now are trying to reinforce a moat that is definitely AI-first. To do that, you just need to connect a lot of pieces and buy a lot of pieces. It’s an ongoing period. It’s a wonderful period because for a couple of years, our space was very dry when it came to M&A or going public. When some companies went public, it ended up pretty horribly. Right now, I think we’re in a golden age again in ad tech. I’m glad for all the companies who are experiencing this.

ES: We’ve seen a lot of movement in the legacy space. You had LiveRamp, and then DoubleVerify and Nielsen just this week. LiveRamp was an interesting one. It’s difficult when you evaluate these things to tease apart the degree to which they map to a future vision of the space or the degree to which an asset that has cash flows that are probably deteriorating can still present a profitable opportunity. My sense with LiveRamp is that it is not a future-proof approach. What LiveRamp does seems to be in support of an old way of doing things that I don’t believe anyone believes has a lot of growth potential or is even something that will be dominant in three to five years. But there is still probably value there from a financial engineering perspective. With DoubleVerify and Nielsen, I also kind of get that feeling, but maybe there’s a bigger picture in terms of the market that I’m missing. How do you interpret both of those deals?

MS: LiveRamp was a PE acquisition, so it was definitely about financial engineering. On the other hand, LiveRamp is fueling a lot of identification on the CTV space, and that’s definitely a growing space. In a way, I think they are the dominant player on CTV identity, working with pretty much everyone. Assuming we both agree that CTV is going to outgrow linear TV—I think in the U.S. it’s already passed in ad spend—then I’m pretty sure they actually do have space to grow. They will probably also make their own acquisitions, probably in the measurement space. When you look at the Nielsen and DoubleVerify, this is definitely a measurement play and a data play because DoubleVerify had a couple of technologies that were super interesting for Nielsen to get their hands on. Nielsen is a very old legacy player coming from traditional media and linear. Taking someone like DoubleVerify under their wing is strategic. DoubleVerify is a tech-first company that does pre-bid measurement and also made a couple of acquisitions in the past. They are well-connected with pretty much all the major brands. They’re not a very performance-driven company, but when you’re going after the Fortune 100 or Fortune 200, they have the customer base. They are the de facto player you’re going to use for verification, pre-bid, brand assurance, and all the things that big brands care about.

ES: So that is a buoyant segment. That’s not just piecing together components of an ever-decreasing pie. That’s probably a pie that’s fairly stable and there’s value in bringing those pieces together.

MS: One pie is shrinking and one pie is outgrowing it. The digital pie and the AI-first pie is definitely outgrowing what’s shrinking, which is legacy. Look at Nielsen, for example. That’s definitely them playing catch-up. On the other hand, if you look at Walmart’s acquisition of Vizio, that’s incredible. Walmart has such a huge footprint of user-level data. Imagine that you can now expand that into the CTV space. You’re building a competitor to Amazon almost overnight. You take all your consumer-level data and apply it as a targeting means to a channel that is growing very fast, which is CTV. That acquisition I completely get why it’s very strategic and an amazing milestone in the industry. Companies understand the strategic value they can get from marketing technology companies, which for a while was super dry. We didn’t see any interesting acquisitions or any interesting publicly listed companies, other than AppLovin, for some time.

ES: One thing that makes me a little bit nervous about a lot of the marketing startups that are emerging now that are self-described AI companies is that they tend to be founded by people totally outside of the space. That to me is always a worrisome sign. It’s reminiscent of the crypto gaming era when you had a bunch of people launching gaming companies who were crypto-first and had never done gaming. The needs of marketers are pretty specific and require domain knowledge that goes beyond having been a power user of Instagram. I haven’t seen anything that I’ve found to be truly innovative. I wonder what the future of this crop of startups is. Some of them have raised tremendous amounts of money and I don’t see the use case or the value add. The very first wave was just startups that would do creative production. I saw no value in that. I didn’t invest in a single company doing that because I thought it would get absorbed into the platforms, and I think it has. Facebook says they’ve got eight million advertisers using their generative AI production tools. Who can compete with that? You’ve got to spend ads outside of Meta’s ecosystem, but every platform is going to offer that. AppLovin is offering it. Now you see a lot of orchestration startups emerging. Again, for the most part, that’s been a solved problem. You’re either big enough to have done it internally or you’re small enough where you’re only relying on one or two channels and they handle it for you. I don’t know that there’s a big space in the middle of companies that are spending across enough channels to need it who haven’t built it themselves. If you as a total outsider from marketing could build this tool, then a CMO could probably organize three engineers to build it using the exact same tools you’re going to use, which is Claude or GPT-4, and know the problem space better. My sense is right now we are seeing a lot of consolidation around legacy problems that is adapting to this new environment, but we are not seeing a lot of solutions emerge that are AI-first. The market evolution has been driven by AI, so you are seeing these strategic shifts in needs and opportunities, and that’s driving M&A. But you’re not seeing a lot of AI-first M&A because those tools or solutions don’t actually hue to the changes we’ve seen in the operating model. They’re just kind of noisy. Do you agree with that or do you think I’m dismissing a lot of these AI-first companies?

MS: When I look at generative AI right now, it’s a big question where this will go. There’s a lot of slop and startups that are taking what they used to use and just creating an AI version for this. I see a complete inflation of apps and SaaS companies. Being a founder today when you want to MVP something is much easier than it used to be. On the other hand, when I look at the scaled solutions, especially when I work with enterprise clients, they’re not going to work with tiny startups that are simply relying on something like GPT-4 or Claude. When you really start scaling, you go back to relying on old technologies or older practices in order to do that.

It’s too soon to call out generative AI. We’re just in the infancy of that. I do think the level of orchestration there is definitely a big plus. Maybe what we’ll see is companies who build an orchestration layer to orchestrate multiple MCPs, so you can completely unload your vision to the point where the system can execute things for you. But we’re not there yet and that’s not the type of acquisition we see right now. When it comes to AI components we’re seeing right now, it’s reinforcement learning and intelligence. That’s the level of acquisitions we see today. On the creative generation side, it was pretty obvious that the platforms would have their own solutions. That’s not what’s getting acquired right now.

ES: I feel like a lot of the analytics solutions coming to market now that are supposedly AI-first are just recreating stuff that works well. You either talk about orchestration across channels, which I feel like for a certain size of advertiser was a solved problem, and there’s just no opportunity there to serve that with a SaaS tool. Anyone for whom it was not a solved problem can’t pay for it, and everyone for whom it was a solved problem doesn’t need it. I’ve also seen a lot of these tools that are recreating player profiling or lifecycle optimization or even stuff like monetization optimization for the end user. I’ve seen stuff that does on-page optimization with context for e-commerce. I read about one two days ago. It exists. It’s not a new thing. Someone knows how to build ML tools and thinks e-commerce would like to optimize a page for a person they don’t know anything about yet. That would sound like a good idea. If you did zero research, you’d think that’s a blue ocean opportunity. If you did five minutes of research, you’d understand that there are a lot of tools that serve that use case. Maybe they’re not built by people that came out of DeepMind and maybe they’re not even using that sophisticated of an ML methodology, but nonetheless, they do the job. Your goal is not inventing this category; your goal is to demonstrably outperform those existing tools to the extent that you win all the business. That’s what makes me nervous.

MS: Then you’re talking about the SaaS apocalypse, apparently. I had one client in the last 24 months who, ahead of a renewal, came to me saying they were not willing to commit to a yearly contract and were not willing to raise the price. I told them I wasn’t having this conversation with anyone else because people who are smart enough understand that you cannot get this through some sloppy made tool. If you want to try, go for it, and when you come back, we’ll talk about new pricing. I literally had just one company who thought they could take advantage of this SaaS apocalypse and let us build our own sloppy tool. Most people see beyond that and understand that just because something looks UI-based and fine, it doesn’t mean that there’s actual content behind it.

On the other hand, if you are a startup today and you’re looking for tools, and you have an option between a tool that costs $5,000 or $20,000 or a tool that costs you 20 bucks a month, of course it’s very tempting to go for these sloppy tools. I’m not that worried about that. When people start seeing some sort of scale or when they see that they cannot scale with these tools, they eventually get to the actual good tools. These enterprise-level built tools, which have a lot of ML and AI in them, even if it’s not generative AI and maybe wasn’t created with Claude, that’s where you see the most value. It’s something that was built, thought of, and tested through actual customers.

ES: We’re agreeing with each other. My point is that the spate of AI-first startups that are being founded by people who themselves are AI-first from a discipline or a domain expertise standpoint are trying to replace systems that work well. They just don’t know enough about the space to know where the real opportunity is. It’s not that they can replicate these tools quickly and easily and steal the market. I don’t think they can. The market’s already very competitive. They may think it’s blue ocean because they just didn’t do any research about the space. But you do see these companies where people come out of OpenAI or DeepMind and they’ll raise at a $100 million post for an idea that you know as a practitioner on its face is not an unsolved problem, it’s not a blue ocean opportunity, it’s being serviced by multiple tools at this moment. They’re not actually adapting to the new marketing environment that has changed fundamentally as a function of AI. They’re entering this already solved portion of the market with a tool that is not differentiated. A lot of times you look at these people’s backgrounds and you question whether they have that deep domain expertise in ML in the first place. It’s just that they came out of Google or OpenAI, but they were a year out of undergrad and they studied something unrelated and they were doing something not that technical at the previous employer. You see a lot of that. The big point I’m making is that I see the market evolving, I see a lot of the M&A happening and doing a lot of consolidation and adapting to the new opportunities. I don’t see a lot of the startups emerging to serve the new market reality. I see this is just the AIfication of existing tools that didn’t really need to be reinvented.

MS: I completely agree with you. A fundraising strategy is you put an AI label on yourself, same as a couple of years ago you put an NFT label on yourself or a couple of years earlier you put cryptocurrency or web 2.0. Fundraising strategies—you will have a shoe company saying they are generative AI for whatever. Founders want to raise money and start their company. When I mean an AI-first company, I truly mean a company that adopts an AI mentality where you understand that you cannot control all the things the software will do and the software often will hallucinate and mess up, and you’re fine with it because you understand you’re going for the greater good. You’re going for a long-term vision. That’s what I consider as AI-first, not any company that says they are an ad network but an AI ad network. That doesn’t mean anything. I completely agree. I don’t understand why everything needs to be AI when it’s not.

ES: I want to talk about the AppsFlyer investment. What’s your interpretation? I know you’re on the ground and probably have a good sense of what the consensus read on the investment was.

MS: I’m really happy for them. I know Oren and Reshef for many years, even since they were in a Microsoft accelerator. This is a huge success. I hear that a lot of people were cynical about it, saying they are no longer independent. I completely think the opposite. Because they are such a big player in the industry and they have 60 to 80 percent market share in mobile attribution, they’re the referee. It makes sense for the system to sponsor the referee, because if you don’t have the referee, then this is complete and utter chaos. I completely understand why they managed to pull it off. Amazing that they did. All the players who participated in this investment understood that they have much more to lose if AppsFlyer would fall. If AppsFlyer ended up either in the wrong hands or reached the point where they’re in financial stress and had to cut more people and the market goes into chaos and suddenly you don’t have one major third-party attribution player but you have 15, then you’re reaching a point where there’s complete chaos where you cannot trust the referees. Even though Google and Meta are not the biggest in terms of actual money invested, they would have so much to lose. I completely get it and I also completely get why they are still independent and unbiased. Oren’s explanation to it was very honest and transparent. I also get the cynicism in the industry.

ES: That was my argument. It was too big to let fail. All of those companies, those four—my next question was going to be why these four, but you just explained why these four.

MS: I don’t know why others didn’t join. I don’t know if others were offered. Snap, TikTok, Apple Search Ads, X—we both know which ad networks they were all defending against. I don’t know why specifically these four and not others participated.

ES: They have the most to lose. Snap, I don’t think they had the latitude to invest in something like this right now. But they benefit from the status quo. The status quo works for them and for advertisers too. The point I wrote in my piece was it was too big to let fail. I think people interpreted my piece to mean that I thought they were in financial distress. I said in the piece they’re profitable. They’re not in financial distress. They’re very well-positioned as a going concern. The point wasn’t that they might be in financial distress. The point was that PE ownership clearly wanted to get out. What happens if they say okay, well let’s sell to this company that isn’t committed to independence? The fact that they took this investment from a syndicate shows that they are committed to neutrality. If they weren’t, they would have just sold to one party that would have owned it and controlled it and not been neutral. That would have been the direction to go in. My point was more about the fact that they’re comfortable, but if the PE firm says we want to sell, you’ve got to cut expenses because we need to juice the profitability, then you’re talking about cutting people and service degrading potentially. That would be a problem. Or a company buying them who was not dedicated to neutrality. I think this is the best outcome for all market participants, including advertisers. It wouldn’t make sense that the people that invested got preferred treatment because who would they favor? These are the four biggest app install players outside of AppLovin. So they would favor them? All four of them? Which one of them gets the most favorable treatment? It doesn’t make sense to interpret it in that cynical way.

MS: I completely agree with you. There was this scene in Oppenheimer where there was a potential chain reaction where we all end up killing ourselves. The scene there was if that’s what your math shows you, share it with the enemy and hope they also don’t want to kill everyone, including themselves. This was pretty much it. These four players understood that they have more to lose if AppsFlyer ends up in the wrong hands. If you take this analogy of a referee, a referee can easily throw the game, can easily sell the game. If a referee wanted to optimize profit, there’s so much evil they can do. When you think of an AppsFlyer, if they didn’t have this investment and they really needed to optimize purely for more profit, faster growth, and so on, there’s so much bad that they could do if they just wanted to. I really appreciate the outcome here and I’m very glad for them.

ES: What’s the future of that space? This does seem like it codifies the status quo. We probably don’t see any dramatic innovation now because the point was to keep everything the same. Do you agree or do you think the MMP space looks radically different in three years’ time?

MS: The MMP space is infrastructure. It’s pipes and you need pipes. Everyone needs pipes. Pipes are fine as pipes have been. You went to Italy to a very old building. Hopefully it was connected to pipes, and pipes haven’t changed for the last hundred-plus years. You don’t necessarily need to change infrastructure if it’s working as is. Attribution has flaws, it works, it does what it’s supposed to do. Leave it. I’m really glad for the outcome here for everyone involved. I don’t think necessarily that they need to change it.

One of the reasons why I started Incremental was knowing that an attribution solution can’t go and offer what Incremental was offering because they cannot circumvent themselves and say, “Oh, remember all these conversions I was counting? Here’s another tool that tells you that they were all wrong.” That gave opportunities for other companies like Incremental, but it did not circumvent AppsFlyer. No one in their right mind would say, “I don’t need to use attribution anymore. Let me just use these additional or other measurement solutions that look at things from a completely different lens.” No one did that and that was never the pitch.

ES: That’s under-appreciated and I misread that. I thought the era of MMP-style attribution is over going into 2021, and that’s not what happened. What I recognize now though, and what a lot of people underestimate, is how much innovation is happening around that. You’ve got core MMP attribution, you have it, you always will, but you’re building a whole ecosystem of measurement around that yourself and you’re layering on tools like Incremental and your own proprietary tools. You’re using the MMP as this attribution signal, but there’s a lot of other stuff that you can build on top of that—synthetic measurement, call it, that adjusts for it or modifies it in terms of your own business intelligence and decision-making. That has flourished. The underlying foundation is rock-solid, not going anywhere. It’s the Roman aqueduct, to extend the Italy example. You’re always going to have that, but inside your own building you might do calcium removal systems and a whole bunch of more sophisticated stuff to pump the water up to the top of a six-story building. But the MMP space, it seems though it’s probably frozen in time from here on out. You’ve got the major players, which would include Branch, AppsFlyer obviously, Adjust, and then some smaller solutions like Tenjin, and there’s not going to be big evolutions here. These companies have core businesses that are very strong, there’s demand that’s not going away, and they will exist for as long as mobile app installs exist.

MS: They can all complement with additional services. Singular is very strong in the cost reporting side. Kochava is definitely on the VIP velvet glove service to some of its customers. They all can find their edge and they all have a reason to continue existing.

ES: Let’s jump into ChatGPT ads. Whenever I do a post on just a summary of what ChatGPT just launched, it does really well. There’s certainly a lot of interest in ChatGPT ads and it’s funny how polarizing they are because you get people that seem really dedicated and committed for whatever reason emotionally to this idea that ChatGPT ads are going to fail. Not only are they committed to that idea as an outcome, but they’ll argue that it’s already happened. They’ll argue that ChatGPT ads are already dead; the idea didn’t work. I don’t know if you saw this, but eMarketer published this estimate of the total size of chatbot ads being $10 billion in 2030 or something, and $1 billion this year, and ChatGPT’s target revenue was $1 billion this year and I think they’re on track to beat that. I called it an alternative information space because if you live in the primary information space, you see ChatGPT pushes out new functionality almost monthly, it seems like they’re scaling really well, and then the alternate information space is people saying it’s not working. Talk to me about ChatGPT ads. What are you seeing? What are your clients doing with it? How are they finding the experience? How are you seeing the platform evolve?

MS: Incremental was the first company to publish actual analytics on ChatGPT and performance. I can understand why people think it doesn’t work. It’s a zero-click, zero-touch platform. Ads are there, but similar to social, but way worse off, people are not going to be clicking out as frequent. If you are relying on last-touch top-of-the-funnel metrics, it’s going to look terrible.

There is a very meta analogy here. Incremental is an advertiser on Mobile Dev Memo. The first month we were ever live, we got few leads, to the point where I was thinking we should not continue. But then each of these leads was super high-end, sophisticated advertiser which we ended up working with for long. So it’s crazy high CPL, incredible returns. ChatGPT ads is pretty much the same. You’re not going to get a lot of top-of-funnel, so you may not get a lot of installs and you may actually not even get a lot of purchasers, but the ones you will get are going to be so qualified that the LTV will compensate. In a world today where the whole user-level tracking and measurement is quite difficult, for you to get to the point where you’re spending a thousand bucks and you know that from that thousand bucks you only got one install, but that one install actually ended up spending a lot more, takes a lot of guts, belief, and different types of measurement. Attribution is doing what it’s supposed to do, but if I judged my quality of return from Mobile Dev Memo advertising—very meta, Eric—based on leads, then I would have never been as an advertiser for the podcast for the last two years. It’s the same equivalence there.

ES: It’s early days, but they did integrate with AppsFlyer and Adjust on app install tracking. I could see subscription products, like “give me the best run tracker app,” maybe that performs well. But for high-volume installs, I imagine it works similarly to Reddit. E-commerce I could see doing pretty well and that would be click-based. The point I made when they took money from Amazon was like clearly Amazon wants to advertise here. They don’t want ChatGPT and other chatbots scanning their product catalog and excluding them from the process for obvious reasons, but they do want to buy ads there. E-commerce could do pretty well. I’m moving back to Austin and I’ve got an office there and I’m decorating it and I use ChatGPT to do research and to coordinate stuff, upload pictures, and it works really well as that kind of tool. Have you seen e-commerce perform there?

MS: We see e-commerce, travel, lifestyle products, healthcare, BI tools. We see pretty much everything other than mobile app installs at the moment, but I think eventually they’ll also go there. What they need is more tools, which I’m pretty sure they’re developing or working on, even though they keep telling the market on the one hand no, but on the other hand they’re releasing everything we expect them to release. I don’t know if it’s the advertisers who are working with them telling them they need to release this or if simply the teams—they are hiring people from the industry, we know many of their people—they have a written playbook, they just need to follow that.

ES: Anyone who is surprised at the direction the platform is taking just wasn’t paying attention. Look at who they hired. It’s people that you would expect to be doing it and they know exactly what they’re building. I can’t surprise them with how I guess what’s coming next because it’s very obvious and linear. When they introduced the standard events library and they said they’re not optimizing towards that yet, it’s just a measurement tool, well then I know what’s coming next—conversion optimization. But we needed the events library first and CAPI first in order to build up to bootstrap the data to do that. You can just pencil out the next milestone and trace it, you know exactly where they’re headed. Why would they take this radically different direction and do an ad network? It doesn’t make any sense. They’re building Meta Ads.

ES: Talk to me about channels that you’re seeing that are ascendant in the portfolio. I had the CEO of Mistplay on recently, the rewarded mobile app install space is booming. What else? Reddit? Pinterest? What channels do you see that are interesting that maybe you wouldn’t have guessed are doing well?

MS: From my perspective and looking at a lot of spend, four to five billion dollars of ad spend in the last 12 months, we do see CTV growing in terms of share of wallet. Meta is definitely back on where it’s supposed to be and took some share of wallet. When it comes to UA network, it fluctuates because performance trends go up and down, seasonality goes up and down. There is no single platform I’ll name other than OpenAI Ads or ChatGPT Ads, which we saw from zero in February to a sizable number already in July. I’ll be surprised if it doesn’t end up ranking within top 20 across the board by December this year. Mainly because when they open up the door to everyone and started offering support to third-party measurement and MMPs, that’s how you grow. When it comes to UA network, that’s definitely a lot of gaming, but today gaming is maybe 30 percent of our customer base, so I’m sure we’re not necessarily seeing the same trends as other companies would be seeing.

ES: Maor, it’s great to have you on, as always. Talk to us about Incremental. How can people go and become customers of Incremental and talk to people about Smartly?

MS: Incremental is on incremental.com. Book a demo, speak with me, speak with the team. We also recently added Aurora, which is an LLM part of the platform. People really like using that because you don’t need to think hard anymore, you don’t need to look at numbers at the dashboard, you just ask the questions and get answers based on your marginal data. Smartly works with pretty large enterprises for workflow automation, campaign automation, and creative optimization. There is a bridge between Incremental and Smartly where if you wanted to include an always-on incrementality signal, Smartly can optimize towards this on all the platforms they’re connected with, which includes OpenAI, Meta, and Snap. They’re very strong on the social and CTV side. That’s smartly.io and we’re at incremental.com.

ES: Maor, thank you so much.

MS: Thank you, Eric.