Tim Rowe, Head of Marketing:
Welcome to Radically Candid, a cognition podcast that takes you behind the scenes with the people, personalities, and perspectives shaping how we think about streaming TV and how we approach solving challenges for agencies who are trying to build an owned and operated ad tech stack. My name is Tim Rowe and today's conversation is with Michael Lieberman, product manager of Headless Analytics and also the founder of the company that we acquired to incorporate that pixel technology. So if you want to learn about measurement, this is an episode that you cannot afford to miss, literally and figuratively. There's a lot of technical knowledge in here. There's a lot of strategic and tactical implementation that you can take away from this episode. If you're a founder, if you are someone that manages advertising spend, especially this is a conversation that you do not want to miss. So enjoy. Michael, thank you so much for being here. The question I think that's on everyone's mind internally, externally, as I talk to more folks is what is the problem that headless analytics solves for?

Michael Lieberman, Product Manager [HAT]: So the thesis behind the headless analytics was about making it really easy to collect a lot of first party data from a website or web application. and then allow you to use it where you need to use it. Meaning there are different use cases for first party data. Who's visited? What did they do on the site? Did they transact? What are the features or the components of the site that they're using? And fundamentally, the idea of the headless side of it is that that can then be used again, like I said, for different use cases. So attribution is a is a big use case. So foundationally or from the thesis, it was put this one tag on on the site and start collecting it all without really putting a lot of requirement or lift on developers or data scientists to go and do something with it. As that pertains to what we're doing here at Cognition, we've now been able to collect all that sort of first-party behavior, but also tie that back to the creative or the off-site media that's driving summoner sites. Traditionally, it was You know, you ran an ad and somebody clicked on it and came to site. We had a good idea of where somebody came from because they clicked on it. Subsequently, they did a bunch of things. You divide a bunch of pages, filled out forms, looked at images, etc. Converted if there's a transactional component. And we've extended that now to give visibility even before that click. So I show Tim an ad, he's exposed to an ad, he then comes to the site, maybe it's one of these different channels like a Google search because I was watching a streaming TV ad put in the brand that I'm interested in that I saw in the ad now I've come to site. So we've been able to extend that down. So the thesis was really about making that data collection easily, but now it goes even further by having that visibility to the things that brought you to site that weren't behaviors. Again, as I said, for headless analysts, it's about automatically collecting those first party behaviors. Now we're extending that to some of the things that don't actually involve

Tim Rowe, Head of Marketing: A true behavior meaning i clicked on something it's that i was exposed to someone excellent thank you for framing that specifically i think one of the the most common use cases that folks are trying to understand is how does my channel mix. influence and outcome. I think you shared it there. Someone saw a streaming TV ad, they then searched for my brand on Google, they clicked on my Google My Business, they took an action on my website, and then they ultimately became a lead or a sale, a customer, a prospect, somehow they converted. Understanding that journey is one of the things that headless analytics is best suited for. Can you maybe walk us through how that works technically?

Michael Lieberman, Product Manager [HAT]: So I think you kind of talked about a few different pieces of the puzzle there that are part of the journey. So there's the media that I'm exposed to. There's the media or let's call it the advertisements that I interact with. So across different channels, that could be a search. That could be an ad that I click on in my social stream, that could be opening up an email and clicking through, or really just directly typing in the brand's URL and going to site. So those would be the sources of that traffic. And then I've got all of the activity that happens on site. So again, with putting on the hat tag on the site, we're going to collect all of that, where they came from, what they're doing on site. Now you extended it to, you know, say a transaction or completing an order as we call it. And there's kind of two ways to look at how that's done. So if it's an e-commerce environment, if we're going to speak specifically about automotive, if it's a transaction online where I bought my Tesla, technically I have the transaction there and I can tie revenue to it or the conversion event. If we're talking about something that happens, you know, obviously in a dealership, I might not do any of that transaction on site. But what I do have is a bit of a journey there. And at that point, it could actually be anonymous for all intents and purposes. But now I go into a dealership, I transact, they have a series of first party data around who's whether they bought a vehicle or, you know, had a service appointment and completes a bunch of ROs there. I have that, now I can load that in and start connecting those dots. So, I mean, yeah, that's kind of the technical way it works. It's really stitching that person's journey touch points in different places all across. And that includes them seeing an ad either on one of their streaming devices or a display ad or an online video ad. So it's really about connecting that individual across all those different places to be able to go back and do a bunch of things with it. The one that obviously stands out is, how do I show that this media influenced the conversion event? you can go deeper beyond just sort of media attribution to multi-touch attribution and start looking at how those touch points work together and, you know, starting to do some fractional analysis on how all those play together to drive that conversion event. So you can kind of tackle that attribution problem in a bunch of different ways, but I think it always has to start with what are you trying to do? And obviously everyone's going to say, I'm just trying to sell more vehicles or Shoes which is hot or whatever it is but really it's about. When i start thinking about this in the auto landscape it's really about. I think that that so much of that sale happens completely outside of the dealership. and all of these touch points have to work together, that website experience, et cetera, to basically warm somebody up such that they walk in and are pretty damn close to buy. I don't know what the metrics are at that point. I'm sure you know much better, but if it's like one in three of those people that walk in the door need to transact, well, we need to do all our work ahead of time to make sure that they're walking in there with

Tim Rowe, Head of Marketing: Everything's sort of lined up for them to sort of take it down. Efficiency is kind of what stands out. How, how efficient can I be in driving that outcome? Hey, the, the market that exists for me to sell my thing is this big. How efficiently can I get in front of all the potential people that I could sell it to engage with them and be able to attribute how each of those ads influence the next step of the buyer's journey. Something that we've talked about a little bit is. end state of using attribution to inform activation and creating this feedback loop where attribution can inform the activation strategy. Is that something that you have thoughts on?

Michael Lieberman, Product Manager [HAT]: Traditionally, there's kind of been these two groups. If you think about sort of large organizations in the past, there was that group of people that were responsible for analytics and analysis. And on the other end, there was a group of people that were in charge of activating data. email lists or whatever it might be. I don't think that that distinction exists so much anymore. I mean, even within organizations, they tend to become a lot tighter now in terms of the way they work together, because fundamentally, I think they're working off the same data set. There's the identity of that person. Who is that person that went through that journey and bought something, for example? Or who is that person that went through that journey and didn't buy something? They're probably even more important in some cases, right? So the person who's doing all the measurement the preparation kind of figuring out who this group of people are are then the ones that. Are being activated against if i know this for example these channels are bringing these kinds of buyers in and they're converting a certain rate. And we think that that's not effective vis-a-vis other channels. Well, now activation is deciding where do I, where do I spend my media dollars? So that's just one aspect of, of sort of having visibility across the journey and the attribution side of things. I think the other part is. You also sort of have a lot more enrichment about that person and their interests and so forth. Right. So, you know, when you build out an audience, you're always trying to get all the details you can to be as effective as you can when you are putting an ad or sending an email to that person. It's it's I mean, everybody's been talking about it for years. It's it's a bit. Overuse but it is that idea of hyper personalization you want to get that message to the right person right time blah blah blah yes we heard a million times but really that's what it's all about right that's how you're efficient based on you know referring back to your comment so if you can do that. Or how do you do that yes i have demographic info yes i know i'm i got eighteen kids so i need a bus right so i'm gonna market buses to you or or. Sprinter van but in reality it's much more than that so the idea of all that behavioral data. That i think is exciting from an understanding of that individual leads into the activation of it is not just measuring how many people look that blue. Ford tempos i don't even know if they make them anymore but blue ford tempos how many of those people bring back the tempo and the pinto the pinto the tempo the the santa maria there we go actually just go there's actually you just made me go on a tangent thing about all the vehicles that you bring back but i digress you're bringing back square body gmc next year just as an aside since they are bringing some things back Yes, yes, I think they are. And that's exciting.

Tim Rowe, Head of Marketing: So I didn't mean to derail you.

Michael Lieberman, Product Manager [HAT]: You did, because I started thinking about cars. I'm a GMC guy personally, so that one took me a little bit further, but I will pull it back. Again, I think the more you know about a person and their habits and what they're interested in, we can see everybody, all the images somebody clicked on. If you see that they're always looking at the interior, even think about that. I'm an interior guy. I like the interior, obviously. If those are the images that I'm looking at as 90% of the stuff that I engage with when I'm looking at a particular vehicle, well, maybe my next step should be that, right? You see how all that understanding of that individual becomes much more relevant and what I might want to put in front of them. You know, we always assume it's like that speeding car down the highway with the light shining off it and all that. But if that person is buying based on the number of dials they see or the screen size or whatever it is, well, maybe that's that's the thing that's more likely to get them in the dealership and make that transaction versus, you know, different vehicle, for example.

Tim Rowe, Head of Marketing: Zooming out for a minute, you'd mentioned how if someone's maybe viewing a specific product feature at a higher rate, that that might be a good creative to deploy. looking at cognition more full stack, is that then something that we could use Creative Studio and ad serving to take that insight and turn it into an ad targeting that person, getting more right, getting more specific about the message, or is that a future state that we can dream about for now?

Michael Lieberman, Product Manager [HAT]: I mean, I think there's ways to kind of do it in a less granular fashion that maybe I'm describing. Obviously, it's pretty typical. You know, you run an ad and the guys looked at 13 Buick Regals. You're going to show them more Buick Regals. I mean, that's pretty straightforward and typical. I think you can go further and maybe definitely do stuff more on the, let's say that example, like I always like taillights of cars. I'll show you taillights. Silly one. But really, I think the future state and the stuff that gets much more exciting is potentially really tailoring that creative. And I think you can use a lot of the AI tools that are available. So to generate creative that really aligns with what someone's looking at, maybe that's more effective for new vehicle sale as opposed to used. But the idea of having a number of different almost components of a prompt to generate that creative that really matches me. I'm looking at, you know, I need a truck because I like fishing and obviously have to collect this. But maybe that that's kind of the enrichment from, you know, the Amazon data that you have about that person. I mean, there's there's I am kind of brainstorming a little bit, but. I don't see this stuff is that far off in the future where you can go and kind of take these different inputs and generate an ad that looks like it's almost, it's made for Tim. Sure. Like it feels very personalized Tim's logo on the back of it, even, or, or, or something like that.

Tim Rowe, Head of Marketing: It's taking all the tags of like, well, just thinking, thinking out loud here, an algorithm like tick tock, which is based largely on tagging elements of a video to an associated behavior. How that. that then used with AI could create all of this specific content. I don't think that that world is that far away. I can imagine a world where our kids are on their tablet and then they voice prompt the type of content that they want to watch and a 20 minute show is created. All the parts exist to do that. So it would seem that we are not that far away from an ad serving standpoint. I appreciate you going there with me on it. Coming back to the e-commerce use case, and you'd mentioned social, but I know that there's a copy and paste application of the Pixel for the Shopify site. Can we talk about e-commerce? Seems like there's a lot of meat on the bone there for folks that sell other stuff online, shoes, widgets, what have you.

Michael Lieberman, Product Manager [HAT]: I mean, we've worked with e-commerce companies of, of all type, right. Uh, including some of our automotive agency partners have, you know, different verticals that they service. So anything from buying, you know, a let's step back. Cause they're e-commerce. I tend to break up into a couple of different paths. One being I sell something online, let's say physical good. The other is I sell something like a ticket. Right. Either way, the, the, the, the connecting component is a transaction online. Right. So, and I think what, what ties that all together is all of that behavior is digital. There is not necessarily, there may be, but most likely there's not the storefront, right? I don't need to worry about the dealership side of it, for example. So I'm basically conducting everything digitally. How did I get there? What did I look at? What did I buy, right? Part of the thesis and the idea of being able to collect all the info about what someone's doing from a behavioral side was always the parallel of, you know, there are some, it's been a while since I've walked into sort of a retail apparel store. They used to do it quite a bit. I'm assuming they still do. But the idea was you walked into You know suit maker and the guy knew what kind of suits you'd like and which combos go well and what your style was like it was that personal touch. So from the behavioral side that's one big part of what we're trying to understand so who's that person what do they want how do we recommend it all this data is the foundation for providing that next thing to again. Make that wholesale and move them along to get something not just the sale but also something that they're happier with right. In Ecom, you can return just like anything else. So you want to get them to find the things they want, be happy with it, really aligned to their needs and desires. That's when they're already there. But the other side of it is getting them there. So back to that point about the measurement and activation, the sort of two pieces of the puzzle. One is I run a bunch of ads to build my brand. I run a bunch of ads. And in building my brand, generate demand then i'm running a bunch of different channels that are designed to capture that demand it's much heavier on the. The paid social and the display and so forth less on the on the search side i think for for something like a calm. Is that tends to be a navigational element that we get into attribution which pieces actually are are driving the sales are driving people to side and. Any get even a layer beneath that which is i think even more important for the success of. Really anybody in the world of marketing, which is what's incremental paying to get the same customers that were about to come in to buy that pizza anyways.

Tim Rowe, Head of Marketing: Yeah.

Michael Lieberman, Product Manager [HAT]: It's not an effective use of that marketing. I'm looking for incremental. I'm looking for growth. That's a big swath of things to cover and unbox, but just tying it back to e-comm, I think what's nice in e-comm, and this is something that we're actually introducing at Cognition, even on the automotive side, but you've got revenue, and ROAS, and return on investment, and marketing efficiency, and incremental marketing efficiency, and all these types of metrics are valuable. but they're that much more valuable when there's a specific dollar amount to it. So e-comm really helps you kind of see that path. I spent $8 million. I made $15 million. This is how it breaks down by channel. People that are doing this are buying and not only are they buying, but they're buying. Higher ticket items when people buy this what also they likely to buy and that informs you know that next round of what kind of marketing or ads in my running a these guys bought socks. Everybody that by socks likely buys underwear later that's the next thing we want to put in front of those people we all know we come because we're all consumers right. And we have certain behaviors that we all kind of fall to which is it's thanksgiving black friday that's when we're gonna buy stuff like that so there's all the seasonality aspect to it i mean i think that applies to all industries. But tied to that are, you know, look at how well email's doing, because I send a Thanksgiving email with discounts, right? Discounts are also a big question in e-comm. Yeah, automotive as well, but how does that play in? And does that discount work to get somebody into a loyalty or an affinity with the brand to keep coming back and buying? I mean, there are so many dimensions on the amount of data that we're able to collect. I always tend to kind of want to roll it back and then you get into the conversation of analytics maturity and even before that one of the things i tend to tell folks is you can get so much information right let's say you know for argument's sake i'm gonna Poke a little fun at A-B testing. People that are coming off of these kind of ads, but they're buying our sort of loss leader welcome first product. Then there's more to tying that into, you know, you run A-B test campaign and we know that this is result, you know, these two paths result in X and this one's better. Great. Did it correlation causation, but even more important than that, if I have a theory, that requires me to then change the site in a path that I think is better based on some kind of a subtest. And I can't make that change, it's mute. So it's almost start with what you can actually affect. Can I change budgets and channels? Okay, that's a worthwhile thing to kind of focus on in terms of any of the data collection or testing. We still lay on the foundation of collect everything because you might use it. You may decide you can utilize it and it's impactful, but it's not. Don't start with the data. It's good to have it and collect it. Start with really what the business problem is. I mean, these are the start with the first principles. People are buying stuff and they're returning it. I don't think it's your marketing problem, right? This channel is bringing the wrong people. Well, maybe that's not the channel you want to be in. I mean, yeah, I'm trying to oversimplify in some regard, but to illustrate the point, but I think it's good.

Tim Rowe, Head of Marketing: It's good to simplify because it is, it's, it's so much and it can be so many things. So bringing it back to the thing we actually care about, did it move the business forward?

Michael Lieberman, Product Manager [HAT]: That overwhelming amount of things that you can do, the amount of data and the amount of choices you have, there is the case where it makes people lazy or really just avoid it, right? I mentioned this to you before, the fear, the fear of finding out, the fact of finding out, right? And sometimes when you find out stuff, you actually have to go and do something about it. There's been this whole, you know, there are, I guarantee people that are going to listen and organizations across the spectrum size vertical that are still using last click attribution. I clicked on this organic search because I finally remembered I wanted to buy something and I'm going to give credit to that. And that's rooted. I want to say in laziness, I don't have a better way to describe it. It's 2025 because it's easy. Right. It is 2025. You know, it's not that much. It's not that much more difficult to be even a little bit better. It's rooted in something that's simple, easy to compartmentalize and understand and just decision off of that. Is that the wrong data? I'm going to say most likely, especially because it's chopping off most of what you're doing. And generally, the argument back is, oh, it's not much different. You know, it's still indicative. I would argue that it isn't even something as simple as the way that we are now pixeling creative. Right. So I'm pixeling my Amazon streaming TV video, and then I can show you a journey of an individual that just came to the site off a Google search. branded and did a bunch of stuff and bought something. And from a last click perspective, first of all, let's just look at the last click, last click perspective that go to organic search, meaning are by running a massive SEO campaign or like, cause that's, that's really the payment in any way for, for my organic search. But even worse. Or not, not even worse, but that being the lazy side of it, it's just bucketed wrong. I never had visibility into that before. So now we're taking it to a next level, but I'm using that story to illustrate the idea of just how last click fails, even without us being able to pixel it. The fact that I can now even further illustrates how flawed last click is, I guess was my, my, my way of getting to that point.

Tim Rowe, Head of Marketing: It's a great point. And I believe we call that traffic reclassification. Is that correct? That's that's how we refer to exactly that of, hey, you would have counted this as one thing. But if you knew where it started, you might count it a little bit differently. Can you talk to me about the collaborative game theory that this is built on top of it? Because it almost sounds like a a blend of multi-touch attribution meets a media mix model. It's kind of the best of both worlds. Can you can you walk us through how that came together?

Michael Lieberman, Product Manager [HAT]: So the idea of multi-touch attribution is rooted in understanding which touches should get credit for that conversion. Now, let's look at some of the pieces the way things were done before. You know, you had your Google Analytics and you had, for the most part, last touch conversions. Obviously, that didn't account for different other touches higher up in the funnel, especially towards the top of the funnel. So, There was that, there were also your platform. So I'd go into Facebook and I would see how many Facebook conversions were reported. I would go into my AdWords and I would see how many AdWords conversions were reported. I'd go into my email tool. The point is at the end of the year, you have a thousand conversions and there's 800 from here, 700 from here, 600 from here. It just doesn't add up because everybody's saying credit for the same thing, which means they're all touched that final conversion. So the idea of multi-touch is figuring out a way to kind of give credit to all the touches along the journey. Sort of the first set or the first introduction of attribution to most people was that drop down in Google, which said, do you want to do first click? Well, it always defaulted to last click. And then it said, oh, you want to go to first click? Or do you want to kind of look at them all the same? Or do you want to do the first and the last one? And it gave you the nice view in there. And you could pick from that cool looking chart. And really it meant nothing. Because fundamentally, first click, last click, they're all a rule. And the rule is arbitrary. And it's not rooted in anything other than, okay, we want to look at it this way. So that started with the idea or led to the idea of people coming in and going, well, I'm the brand guy, so let's do first touch. And I'm the paid search guy, let's do last touch. whatever kind of aligned with their objectives or what they wanted to do. And then it got into a little bit more of the custom ones, which made that fight amongst different channel leads, assuming they're multiple channels, even more. Our approach with game theory was to introduce something that basically abstracted away all those biases. It removed the double counting. It abstracted away the biases of people arguing for how should we do attribution. And the way that we do that is, again, with this model. So I can give you the quick explanation of how it works and what it means. But fundamentally, it's about eliminating those obstacles to using that data to better align where you're going to put your money or your media dollars. Collaborative game theory is about looking at how each touch point along a customer journey provides incremental lift along that path and giving it credit based on that. So I'm going to give you the super simple example. It's the casino example, but it helps illustrate it at a very simple level. Then keep in mind you're going to abstract that to hundreds of people or thousands or millions of people across dozens of touch points. Now you get a whole bunch of data to start making sense of it. So the way it works is like this. So think about it. We're going to take down MGM. So we fly down, uh, we're going to take Carson with us and you and Carson are going to play first. I'm gonna play third, but you start. So you play whatever game, blackjack, whatever you care for, and you win a hundred dollars. So you can tell that all the winnings of that a hundred go to you, right? You're the only player. So really there's nobody to divide that up with. Now Carson flies down, he plays with you as well. And you guys win 150 bucks. You won a hundred bucks on your own. Carson joining made your team win or your coalition $150.

Tim Rowe, Head of Marketing: So I had a hundred adding Carson added another $50 to our net total as a team.

Michael Lieberman, Product Manager [HAT]: Think about them as individual games, right? So the first game you want a hundred on your own, the second game where he joined you or hands hands or games, whatever you want to call it. Second hand, you want 150 bucks. Why'd you win more? Because the marginal contribution of adding Carson to the game. Cool. Now I'm going to fly down to Vegas. My nickname when I'm in Vegas is organic search. That's what they call me down there.

Tim Rowe, Head of Marketing: And now they're definitely going to blacklist you from the MGM.

Michael Lieberman, Product Manager [HAT]: Right. Organic search guy. So you went $100 with Carson's $150, and I joined the game with still $150. Meaning I didn't add any, there was no marginal contribution for me joining the game. So the parallel are your marketing channels, right? Very simple example, but think about all those different touches to the conversion. If paid search plays on its own, we know what the total winnings are. It's it's done across all the games. If paid social plays on its own, we know what the winnings are. Now we have the foundation to go. How does paid search and paid social play together and what are the winnings? And fundamentally it's an algorithm to determine. how to assign the credit to those different channels or touches depending on how you set it up. Which just sounds fair. It's known as the most fair model. When we were investigating approaches to tackle multi-touch attribution, there are really a couple of different channels or different models. One being Shapley value, which is collaborative game theory. It's actually won a Nobel prize for the math behind it. Most people know game theory from the beautiful mind movie. This is not that. It's similar to it, an extension of it, where as opposed to people doing game theory to win a game, this is how coalitions play together to win a game and how you divide that credit. So similar area, different approach. But yes, it's fair because I've abstracted away the things that I mentioned. I'm not looking at everything individual and double counting. There is no double counting. I'm not giving somebody the ability to go and say, well, we're going to weigh the last touch and the first touch 20 and 50. We have abstracted away that bias. And now it's, it's the, it's the machine or the algorithm that's doing it. And it's based on the actual data. So that's how we approach it in a very different way than a rule, etc., and removing the bias and making it fair because it's not based on opinion at that point. It's based on the actuals. The caveat that I think is very relevant and valid to say, which is it's directional. It's not accounting. When we talk about attribution, if the goal is to account that last red penny to this channel, you're never going to get there. So you want to get to a point where you're, you're setting things up such that you have a good insight into the directionality of the channels. It's not designed to be an accounting system. It's to inform where to put money, et cetera, in terms of media dollars. It's not to account that this $30 that we spent on this one Facebook ad resulted in this one conversion for $80. And that was a return. It's. You're, you're looking at it in a little bit more aggregate. Otherwise it gets so, so granular that you're never going to be able to make decision off of that.

Tim Rowe, Head of Marketing: Sure. How hard is it to integrate? I mentioned the copy and paste Shopify example earlier. I was just observer in that setting, but how hard is it to get the, the headless analytics tag live and collecting data?

Michael Lieberman, Product Manager [HAT]: So adding headless analytics to a site is 20 seconds.

Tim Rowe, Head of Marketing: Okay. What's the caveat?

Michael Lieberman, Product Manager [HAT]: Shopify is the connector. You put it on, you just register it and it's in. Regular dealership site, 20 seconds, add the tracking script. And I know you're tying this to the conversation we saw with the cut and paste. That's because we are introducing a new approach to setting up Shopify. But again, it's still a 20 second cut and paste. I mean, you saw us do it live during that call. That's what's involved in doing it. Now the difference is. When you're collecting a transaction online for an e-commerce scenario, you're waiting for the transaction to happen there. It's not, I clicked buy, which is an event that you're tracking. It's, Hey, the credit card went, they processed it and it came back and told us, here's what we charged you. And it went through and that's the event you want to track. So that's the little bit of additional for something like a Shopify or an e-commerce shop.

Tim Rowe, Head of Marketing: API. Is that something else? If I have. a way to ingest the data? Can I connect this via API?

Michael Lieberman, Product Manager [HAT]: So the way that we provide access to it is in its most raw form, but we give clients direct access to the database itself. And this is how we rolled out the headless part, meaning we ingest. So we do all the process of ingesting and cleaning and organizing and setting up all the data in this nice database schema. And you can directly connect to that data and do what you want with it. Back to the points we were talking about before, you want to put your own reporting layer on it and build your own dashboards. Great for reporting. You want to connect a reverse ETL tool on it and use that to push audiences into whatever, Facebook through Cappy or Amazon or whatever it is you want, use that there. So that, rather than sort of building the structured API, we made it so that you can basically set something up immediately without having to write any connector code or rely on the API. There's still API, but this is kind of the, Again you can build a full rollout composable cdp if you will based on that data and in a matter of hours.

Tim Rowe, Head of Marketing: I think what's cool about that specifically is. Transparency is this big T buzzword of the year, but that's actual transparency is, no, no, no, we're gonna give you the raw data and enable you to build the views that you want and are most useful to you and your business. I think that's… That's pretty transparent.

Michael Lieberman, Product Manager [HAT]: I agree. I think that's a that's a selling point to what we're doing in terms of not just showing you 800 people were exposed to this and did this action. It's here are the 800 people that did these actions that were exposed to this, which I think is a it's just a different paradigm for a lot of people.

Tim Rowe, Head of Marketing: And not not behind some firewall, you know, just our dashboard, you can't do anything with it, you know, except for look at it like, no, no, here, here's everything that happened. And we want to give you the data and show you all the cool ways that you can, you can turn this into into insights, and then close that feedback loop and actually be able to action on them, activate the next capsule. Michael, I can't thank you enough for the time that we've spent together. I know that we'll have lots more conversations like this, but I think this is a great starting point for folks who are trying to understand headless analytics. So again, thank you.