Hightouch's AI Decisioning, what will this mean for marketers? A chat with Tejas Manohar

Hey everyone and welcome to episode number seven of Couch Confidentials brought to you by Martek Therapy. I'm your host Matthew Neadeberger and I'm excited about having our guests with us today. Today's guest is Tejas Manohar, he's the co-founder of High Touch. After being one of the first 10 engineers at Segment, Tejas took a deep understanding of customer data platforms and together with co-founders Josh Curl and Koshish Gupta, set out to fundamentally change how companies think about data activation, known for their provocative friend, some of it friends by CDP campaign and their pioneering proposal approached. High Touch has grown from an industry challenger to a recognized leader, recently launching major innovations like AI decisioning and a solution for retail media networks. So welcome to the show Tejas and it's glad to finally have you in my podcast and talk to you one-on-one. Thank you Matthew, I really appreciate it and I'm excited to be on the show as well. Yeah, hey, listen, you went from being one of Segment's first engineers to challenging the entire CDP category. What was the moment you realized the traditional CDP model needed disruption? Yeah, great question. So first and foremost, I was a happy segment customer before joining Segment. I still remember how I got the job. I sent a cold email to the CTO, one of the founders of Segment, Calvin, who's one of the nicest people on the planet, still friends with him today. And I mentioned how I love the product and I wanted to work there and just wanted the chance to interview. That's how it all started. And Segment has a lot of happy customers, a great platform today. When I joined the initial product was the connections, what's called a connections product today previously. It was just called Segment and it's a it's a tag management solution based on steroids. So you can put their libraries like analytics JS or analytics iOS in your website or your mobile app. And you can track the actions customers are taking really easily. So one line of code you can say user clicked a button, user opened the pricing page, user signed up, et cetera. And Segment will make it easy to forward all of that data into Google Analytics, mix panel, amplitude, hotspot, the data we're asked and so forth. Now, that products phenomenal. I think it's still still, you know, one of the standard products of the market. There's more competitors today, right? I touch has an events product. There's rudder stack. It feels like everyone has an event collection product, but I think it really set the standard in a way. Segment later transitioned to be a CDP company as that category was taking up and became one of the leaders there. And CDP really meant marketing teams didn't want to just have events forated about what their customers doing to a lot of places. They wanted a data platform to operate it, right? They wanted a place to build their audiences to understand who a customer is and everything about them to build journeys that execute across emails and ads and all these places. What I saw was that CDP was a great vision that every marketing team resonated with from small companies to large companies. You know, they never really marketers have never really felt like they had a data platform for them and they had access to all the data themselves. At the same time at large companies, you know, as we got bigger at segment, we eventually sold the company to Twilio for almost three and a half billion, we got bigger. We started to go after enterprise companies. We try to close the night keys of the world, the procter and gambles and those types of companies, some of which are segment customers today. And one just consistent pattern I noticed was the vision of CDP was great, but it was really, really hard to help companies get all their data into the CDP, right? Doing that in enterprise scale, it's not just analytics events that's important. There's offline data, there's models, your data science team is building. And even just from online, you have so many different apps and websites and stuff that this requires a huge IT initiative. And the only technology I saw getting that kind of IT initiative to get all the data in a company in one place were these data warehouses, so snowflake, data bricks, Google BigQuery, Amazon Redshift. You know, I thought CDPs were growing fast, these technologies were growing crazy fast and continue to across the enterprise. And that's where I got the idea of, hey, how do we give marketers a platform that sits directly on top of these data warehouses and allows them to do the first year of CDP needs, right? Identity resolution, audience building, activating to different channels, journeys, but later on, things like AI, which we'll kind of get into today. Yeah, no, definitely, and that's the reason I really wanted to connect with you is I was really intrigued by the whole this new product you launched AI decisioning. But before we get there, as I just started MarTech therapy, I moved into an incorporated and I'm a one man army, but you need to do your daily work with two co founders and you work together with Josh and Kashi and I'm just curious, I mean, how has the growth of high touch these last two years? Because you guys grown normally, how is it influenced your relationship among the three of you? Yeah, you know, great question. A lot of founders asked me about this because she's Josh and I have been good friends for six to eight years now. So I have known Josh for, yeah, for almost eight years. A lot of times I changed for, yeah, for six years, it's been a while, right? And we only started working on the company together about five and a half years ago. And we, you know, I actually met Josh through working together at segment. I knew we worked together well from professional sense there, but we also kind of became personal friends and I met Kashi through Josh. So friendship was really there from the beginning. And I think that friendship that we've had that close friendship from being friends, from being roommates for many years, from working. Tons at our previous jobs and working 12 hours together in the early days of the company and being in the same room all day in COVID has really been a bedrock of all the stressful situations we've been through while starting the company, scaling people, learning how to manage, having to let go of people. Conflicts around product direction, you name it, all that stuff happens when you start a company, but I think having a strong friendship from the beginning has been, has been really helpful for us. Yeah. And I mean, besides the horrors of that COVID brought globally, being stuck in a small room with the three of you, I think on the one hand, it could be, could have been very beneficial for a high touch, you, all you had to do was work. But on the other hand, you know, I'm not far from the tree. Yeah, and I, but I can imagine how sometimes, you know, how small these apartments in San Francisco are. So coming back to AI decisioning, I mean, you've taken high touch a long way, you guys started as a reverse ETL solution, then the, I think events came afterwards, which I found an interesting development as well. But then, you know, it's 2005 AI decisioning. I mean, what can you tell us about it? I've done my research, but I'd love to hear kind of what your sales pitch is here. Yeah, sounds good. Yeah, before we get into all the technology and where it's going in the vision. Yeah, let's just talk about what this is. AI decisioning is a new technology. And it's nice. Not a lot of people know about it. Less than 100 companies are using concepts like AI decisioning in their marketing today. And so what is it basically what AI decisioning is is it is an AI that solves the end to end problem of how marketers make decisions of how to engage their customers. What do I mean by that? Well, today marketers have to make a ton of decisions. They have to decide whether they should email a customer. And when I say email, I mean message in any way, questionification, something in app, email, advertised, whether they should reach out to a customer. When they should reach out to a customer should be after they do a certain event should be at a certain time of day, what of the hundred content I have in my ESP should I send a customer. What should that be informed by, you know, even besides content variations, what's the goal I should be pushing my customers towards right is this customer interested in new product categories or should I get them to buy more of the same thing or should I get them to download my mobile app, which is good for their long term. LTV yeah, it's a hard problem and you regardless of whether you think about it that way, you're making so many decisions as a marketing as a marketer and as a marketing team and as a marketing leader. And those decisions are based on a number of things right one is gut feeling, which actually is, you know, it sounds funny to say it that way. I'm not trying to say it in a in a No, no, I think yeah, it's it's really real and it's actually very valuable and we'll talk about that in a bit experience, something yeah, right, you know certain things are you know your customer if you're a good marketer to some extent, right. Second is what's worked historically, right, what is the data say what kind of campaigns perform the best. Now this one's interesting because people often have ideas of like general themes of what's working. But unless you're running your marketing program very meticulously and doing a be testing and experimentation and all that kind of stuff actually really difficult to know what's working at a more micro level and we'll get into how I decision helps there at a second and then three is. Sometimes you're incorporating outputs from your data science team or machine learning folks at the company stuff like. predictive models you know these customer of a million people has a 70% chance of churning or you know product recommendations customers who bought XYZ products often like ABC products, but these ML models tell you very answer very pointed questions and don't really answer the global problem that you have as a marketer, which is how should I engage my customers right think about it. 10 customers could have a 70% chance of churning for 10 different reasons and they could respond to 10 different marketing messages best on product recommendations, you can know that in general customers who buy product X like product Y. But how is that being informed by the last 10 push notifications you sent them about product Y or Z or does timing influence that of the marketing you're sending right now right there's so many different variations when you think about the whole customer life cycle and a lot of points by the sound of it you're taking into account so much more than just the last the last results of the last campaign in history. Exactly so much more than just products history and transactions or the results of the last campaign you really if you want to optimize the problem with decision in making marketing decisions you have to think super holistically and you need to have a lot of data so so how does that goes way beyond got feeling exactly that goes way beyond got feeling. So how does that work how's that possible well what we do is we're giving marketers a platform right a decision is a platform just like an audience builder in a product like I touch or braze or sales for us or like. Journey builder and one of those tools a decision a new platform for marketers and it's a very powerful engine where the first thing they can do is set up their goals you know AI and machine learning can't work without clear guidance of what it's trying to optimize towards. So here is a step where marketers get to put in things like I'm really trying to drive cross cells at my company even though cross cells hasn't been our most successful campaign over the last couple of years. We have done analysis to say that that is the highest value thing we can drive customers towards for us as a company if our customers want to be driven towards that we should drive into it because if someone engages in a second product category. That's where twice as much as a purchase and just the kind of interrupt there you mentioned analysis is that what AI decision is doing is there were recommending saying hey you need to do a cross cell campaign. Yeah a great question so AI decision does an element of that but honestly we're not replacing in house data science teams in house data analysis work in house marketing strategy that your team is setting. What we're helping you do is once you have a goal like that so you know you want to cross cells are worth. You know this much to your business downloading a mobile app is also an important goal. Getting people to activate offers something i'm trying to drive my customers towards we look at all that and we look at all the content you have in braze and sales for us and those different tools that can help get people to take those actions and we figure out. What should you be sending to each customer if anything as well as when and on what channel to drive them towards the highest overall TV so. For examples you know cross cells can be the most important thing but if you have your customers don't care about your other product categories are aren't responding well you should pivot your marketing for those customers to something else like downloading their mobile app and maybe they'll care about cross cells later. That's just really hard decisioning to do today in the paradigms we have where we're building these linear journeys or these you know isolated audiences of customers when each customer is really different and we should be building a journey for each of them based on everything we know about them and that's what AI decision is really doing. Using reinforcement learning is the main technology that we use for this so when you say reinforcement learning i'm again trying not to sound naive but i really from a kind of from a business perspective is it mean that if i launch a campaign and i set the goals i set the audience i'm et cetera and the content does that mean that AI decision will kind of self correct is that what you mean with reinforcement learning it'll try to find the most. Optical path on an individual user level yeah that's exactly that's exactly it so we're both trying to find the most optimal path on an individual user level and get smarter over time so learn based on how your users are again not just a historical product purchases but learn based on how your users are engaging or not engaging or responding to your marketing and to your communications with them yeah so that's one but then we're also learning across all your users so. Yes, we learn on individual user level, but we also look at everything that makes that user what they are regenerating tons of features on them from all your data slicing and dicing it in different ways to feed our machine learning models and we are learning across users as well yeah so the AI decision is also generating its own data to take into the town on future campaigns that's interesting go I noticed that again as a individual user with something like chat GPT that if I ask him hey what do you know about me it's it's. It's storing that data so it can make sure that the content or the results that it delivers is tailored to my preferences and it's interesting to hear that that's kind of the same case with AI decisioning it's taking every time it interacts with a specific user the results of that or the the steps taken and the results are then being pushed back in so that it can be reused for reinforcement learning for the next time that's that's that's pretty amazing I mean. You know I know this is kind of a luring a kind of a sales response out of it but from one CDP fan to the other what are the kind of results you're seeing the responses and I know that you have a great sales team they're doing a massive amount of work but just just on just on the download are the results promising yeah yeah so thanks for God will dig our sales team yeah I know that I've seen a few past clients that I did segment implementations with I went back. To the website and I've I see that most of them been replaced by high touch so oh wow yeah that is that that is one of the trends that's happened yeah um so uh yes we're seeing pretty promising results in the industry uh we are working with fast growing up and coming brands to like whoop if you know the fitness brand product we've the fitness band um okay no they they still need to uh into the Dutch market I think you're not the ones okay got it got it got it yeah you know I'm not going to do it. It's basically but um uh you know we work directly with their life cycle marketing and CRM team and their VP of CRM marketing uh if uh if uh and she's a huge advocate for the iterable platform which you you probably know about and um you know one thing that she pointed out is that not just the lift so they're seeing 10% plus lift on some of their key cross sell initiatives in terms of incremental cross sell so in terms of actual sales. But what's more interesting than that is that um you know we talked earlier about how AI decisioning is actually creating more data it's running all these experiments automatically for you through reinforcement learning so today if you need to answer a question as a marketer like do customers who are younger respond better to uh content the morning versus the evening and does that change based on the type of content we're sending them yeah that's a really hard question. And answer you probably haven't run an experiment to answer that question AI decision has it's continuously running literally millions of experiments through the reinforcement learning technology and surfacing those insights to marketers in our UI um that can help them not just make decisions about how they want to change their CRM marketing content and strategy but can influence the whole marketing strategy like we've seen CMOs of companies really interested in those learnings and they're taking it to other things like billboards and brand campaigns and just the internal understanding of a customer. and something that if a sad that really sucks out to me and I'll quote it in this case um but it was basically something to the extent of they saw more learnings in six weeks of using AI decisioning that in the previous 12 months of experimentation as a team and they do a heck of a lot of experimentation and and I would almost say that that's a given when you're using technology I mean I've done the A/B testing in the past and the amount of effort that goes into planning working out on hypothesis. generating the content programming these A/B testing tools uh putting it live preferably not on a Friday uh sometimes working on a Saturday to fix something that you've seen go wrong uh that needs to be correcting and only to find out a week later that your results are maybe getting a 0.01% uplift because your hypothesis was all wrong it's it's great to hear about how high touch is now using that uh AI uh as a part of the engine I guess around it. So the question I wanted to work on the weekends too and we're from the weekend yet I mean I'm have kidding but and then and they're not it's not unionized either no um he said that yeah yeah yeah it's a such thing that's that's your proud around uh unions but I get fired if I say that in the US oh yeah no no I won't put you in that spot but you did mention billboards and and and other channels the question I have is um I mean. What does it take to kind of set up a campaign around AI decisioning I on the website I read about four steps you have the flow as you mentioned it you have your goals your content your audience and uh something called guardrails that I like to get into but I'm thinking about the output you um is this something that you could put into emails into your uh set top box set top set top boxes difficult word I used to work for a TV provider but. To get advertising on these digital television uh uh oh S's et cetera because that's also what high touch has been downwind is it yeah um is it kind of delivered on any platform. Great question so we believe in our text and ecosystem game uh we're not the ST we're not an ad channel or not your DXB we integrate with those steps um luckily uh as you know we started to I teach as a CDP as a composable CDP we already integrate. With almost 300 different marketing channels and ad channels et cetera but yeah we have those native integrations that while you don't have to buy our CDP product to use AI decisioning it's completely independent we have companies using AI decisioning on top of sales forces CDP or treasury data or other systems um we do use all those integrations from a technology perspective so that we have a lot of integrations that are a decision AI decisioning product now um and so basically we think about as you go into AI decisioning. And that's the hub for you as a marketer to input your initiative so cross cells downloading mobile app loyalty activation of offers those are things i'm trying to drive to as a marketer. You also input all the actions that you're willing to take on customers you don't build those emails out in high touch you continue to be building them in your ESP like Adobe or so stores or uh or a damn system that you're using like a and then you just enabled them through like drop downs in high touch select the ones that are relevant to different initiatives. Um if you want to get fancy you can let AI generate variables to put into those emails to you like a discount offer between zero twenty percent or experiment with different levels of discounts yeah yeah or just or to experiment between five different CTAs um or you can just send the email as it's right yeah yeah um and then uh and then. From that point onwards you can get you know with some guard rails in place you can get AI decisioning clocking and experimenting across your customer base and automatically just triggering sends right in tools like. Graze or Salesforce were iterable and firing off these messages different customers um and learning from that and optimizing your overall program so that each customer is really getting us to most relevant to them, but I would say the most important part asked. Um pastures optimizing the performance of these campaigns which we've seen great lips on you know about ten percent or for who as well as we have a forty percent stat on a pretty important cross the initiative for a specialty retailer in the US to actually have fortune five hundred company. Um and besides performance one of the most valuable things we hear from everyone is the insights right actually being able to go into the UI and see what different cohorts of customers are responding well to what different types of marketing tactics content etc. Because the biggest lift you're going to get is from changing your marketing strategy and your content strategy based on those insights actually even more than orchestration a lot of times so that's super valuable. Um and honestly what we see is that it kind of simplifies a lot of times people think this new product is going to make my my workflow getting a campaign out more complicated and have more steps. We've actually seen to some extent it's simplifying it what I see in a lot of marketing teams, especially in larger companies is that they're always working so hard to hit the deadlines right to get a campaign out it requires three to four people on average. So often a marketing leader involved in the strategy exactly some sort of CRM or life cycle marketer involved at an IC level there's a marketing office person who configures the campaign there's a data analyst who pulls the segments. There's someone else who helps you analyze the results of all these incredible efforts yeah there's all these people involved that are highly skilled highly intelligent people. You know that data analyst that data scientists can be go building really interesting propensity model they can go figuring out what's the value of one purchase of a product versus another for a long term. TV they can be figure out pretty interesting stuff but instead they have to build audience segments on and with AI decisioning. Markers can really just focus on the content and the actions are trying to drive towards in the strategy and then AI can figure out all the nuances of who to send it to and get smarter over time and surfaces insights back to everyone from the data analyst to the marketing team to see about the company. I mean AI's so this is a question that I think about quite often and you're I mean we just read about Mark Zuckerberg releasing another 5% of the workforce mid-level coders. Is it is it going to be kind of a double edge sword for a lot of these companies using AI decisioning I just just taking a little tangent here. Personally my belief is that it's going to help productivity so I don't think there's necessarily a reason to fire people it's just going to help improve their. Productivity and like you know the based on the results that you shared so far also the the performance that they're reaching so. Will there targets get higher will the pressure still be there is that I know it's completely off topic of the product but I'm just kind of. Thinking about the the the impact it might have. Really great question. Yeah it's interesting start to think about how these things play out but we can all think about them for fun and play a game that we'll see what happens it is fun I think that. I think that AI decisioning honestly doesn't doesn't pose much threat to to the labor force right I don't think it's going to be shrinking marketing teams if you look at. What marketers are spending a lot of time on where they can add value the organization is running time on a lot of things and there's a lot of places they can add value the organization and they don't they aren't necessarily always attached to some of these tasks like audience building or journey building and they're always interested in tools that can help them do those things better is is what I've seen. And that marketing office person that marketing tech person is configuring is campaign can now with something like AI decision and go explore how technology can help move their business forward right is it. Gen AI technology that can help them generate more variance and move faster than dependency on their brand agency is it something like I know there's a product wonder can that I've been looking at recently you know hearing from customers it's pretty cool it it brings in an IP graph to find email addresses of visitors on your website and then helps you. Outbound email those people and sales first marketing cloud so you can email people anonymous visitors on the band and cart people are seeing pretty good look from that stuff so totally irrelevant to AI decisioning but the idea is that marketing office marketing tech person can go think about how do you technology to move the business for it instead of configuring. Each one off campaign in Salesforce every day yeah so I think it really is going to make these these teams more productive make them more powerful trade every marketer into a data scientist in a way. And I hope that it does actually I mean increase the the goal is a little bit over time because we're seeing are pretty impressive no absolutely and it does remind me of a quote that I read in the customer 360 book and there's a quote in there saying AI is not going to be. Is it you know it is not the threat for the for the workforce it's the human with AI skills is going to replace the human without AI skills so we're going to be using AI decisioning are going to have an edge on the workforce working with this you know new cutting edge technology. I think I saw that's very similar quote maybe is on LinkedIn yeah and that really resonates with me I think that the marketer is in any profession who embraces AI and figures out what are the ways it can be used. By me and my team to drive the business for it and educates the team etc are going to be the ones that become the leaders of marketing tomorrow we've seen this when new channels come about right in the retail world direct consumer e commerce loyalty mobile the marketing leaders that that saw those things as the rising tides of the market and jumped on it and realized that. Those bets are going to change the course of my business, which also means you're right to just you know that. Composable exactly to an extent to an extent and are the ones that are getting promoted and being speaking at conferences and getting job offers in their inbox and and to be honest right CMOs that purely focused on brand and messaging and positioning are going a little bit out of style right the ones that that that didn't expand their repertoire to. To more than that right that's still very very very very very important I don't think it's going away with AI anytime soon. No, definitely those those types of CMOs will stick to still be around for another 10 years I see it the way that I see it the way that you mentioned yeah hey listen we've talked a lot about the product now I'm curious kind of what what's driving everything I. The documentation mentioned snowflake cortex i work in more tech i've done a little bit dabbling with data engineering but to be able to explain in layman's terms what's not you know why do you need to use something like snowflake cortex or just even take one step back what is snowflake cortex can you kind of double about that or kind of shed some light on what that is exactly yeah for sure so. What do we use what is snowflake cortex and what do we use it for in our product first and foremost I will say this is an optional part of our product it's for snowflake customers. And for companies on other data platforms kind of built similar technologies for us what data bricks and Google et cetera but what what is it means it's a new term lock 11 heard of it even I hadn't heard of it until a year ago I think that's maybe where they announced it. It is basically a native version of some of the lm type capabilities you can see in my chat GPT directly in the snowflake platform so if you have unstructured data like text like email copy in snowflake. You can use cortex to run lm prompts on that data and basically what we're using it for is actually looking at that. Unstructured data so one of the things I just do is when you connect to channel tool like raise or Adobe et cetera we are basically connecting to that scraping all the campaigns out of that tool. Putting it in your data warehouse which is the database of the system you don't have to think about it and we're using snowflake cortex on top for snowflake customers and other elements for other customers to to build attributes off of those campaigns so. This campaign has a discount offer at this campaign has this kind of tone this campaign has these characteristics and those features that we're building those characteristics of different campaigns content push notifications creative that we're building and and putting in your warehouse are actually used to create even more interesting correlations and insights. From an editing perspective because we're able to make insights on how different customers respond to discount offers for example instead of just thinking that email 50 is completely different than email 200 which is true but there's something similar about them. So those tags are extremely important and and I think LLM's actually have provided technology companies like high touch a way to generate those tags off unstructured data in a way that's way easier and way better than I think we've had before. That's pretty amazing listen I think AI decisioning is like I like I mentioned before I'm seeing different companies take different approaches and I think this and I'll have to say that I mean analytics with their their CDP opening up that you know the profiles into a data layer for content management systems use is insightful the same around AI decisioning it's one of these solutions that you know 10 years ago. When I was using tools like optimize to do a beta I wish it was there because the the pace at which you can start generating insights about the effort that you're putting in into optimizing campaigns or at least the customer experience is phenomenal and I I'd like to schedule another call sometimes and actually kind of see it at work you know how do you what does it look like how do you go to the processes and just kind of see. The tool at it's fine is that would be awesome is that we should get it get you a demo yeah definitely hey another another article that was released two weeks ago around off site media networks or retail media networks and in the article said that and I was I've been hearing this a lot in the conversations that I've been having. There's published some research and they're saying that you know 42% rise and add spend on retail media networks according to the E marketer in a fall or November 2024 report. Did you guys already think about stepping into that really you know the off site media network and niche with high touch or you know it was a feedback from companies I mean how do you how do you come up with the idea to jump into that. I can totally see from the outside and how that's how that looks a little random right what's high to doing in helping companies create a media network yeah that's different than a CDP or absolutely anything else is different but it's similar right I think our goal is to help our mission right is to help companies use use data and use AI to to drive growth yeah and when you think about growth one one side of growth is how do I get more customers another side of growth is how do I get customers. To engage my brand that a deeper level how do I get them to use all my products you know I'm more for me come to my site more often and then other side of growth is just other ways to generate revenue from from my customers and from my business and I think media is becoming a bigger opportunity for not just media companies but for every company and retail media is a really good example of that where you're seeing typical e-commerce companies or retail companies like Best Buy now makes significant business lines. Off selling ads to some of the companies that are selling products at that's incredible yeah right and I think there's two ways there's two ways that we see people. selling those ads one is on site so when you go to a company it's like Best Buy's website and you Best Buy itself can show ads to customers on the website the other way is Off site which is Best Buy is a lot of data about different customers of Best Buy and they're always running ads actually to try to get customers to buy more at Best Buy. But they can also run those ads when partnership with some of the companies that sell products on Best Buy just as an example like LG or a Samsung or Apple and you know promote things that those companies want in the advertisement and they can use all the data they know about their customers what they're interested in to influence that. So they're building audiences and thinking those to Facebook Best Buy data just as hypothetical example that's Off site almost working like a media agency right but has this very unique data that no other company has which is your retail data. It just be locked up in Google and Facebook but now the retail websites are offering it in the same way to be able to target their users. Exactly and even even your in store purchases can contribute to the data mode that you have as a company to serve as a as a media platform or a media agency for brands that are selling products right even your offline purchases can help with that. So what we're seeing is that a lot of companies across a lot of industries, particularly retail are getting into this media business that third way to make money with marketing and and the first thing they have to ask is like how the heck do I do this from a technology perspective. On site media, it's a bit tricky. Usually requires multiple solutions. You got to think about where those ads go on your website is affect your search. All sorts of different things. We do have a role there. We help companies like bull.com, for example. We have billions of millions of revenue. We have a case study with them online. We help them with the decision around which adds to show which people right off all the data and their data warehouse. But we have to work with GAM and other systems there. But off site media is kind of interesting where what is it actually doing to do outside media. It really just like building really rich audiences and data sets about their customers. So something like best wise customers and shipping that off to Facebook ads who will ads to trade desks even an email platform maybe to put some put an ad inside of their email which is a lot cheaper. And again and and sending ads based on that first part data which is what something high ditch is already pretty good at. And we have additional products we built around and we have a lot of insights that are useful for marketers. But with some more things we've added can be really useful for retail media ads specialist. So it's a problem we saw we were just curious people we kind of saw like hey some of our customers are investing in retail media that seems like really interesting opportunity let's go learn more about it. And once we got into it we realized there's a lot of overlap between what these companies are doing what these teams are doing and what are other teams at the company that we're serving are doing. And then I think there's an opportunity to help help both teams out. We're driving revenue and growth from from data now there's something we've had to add tons of new integrations and new product feature isn't you know any any opposition like this is a big effort. But there's so much that we've built that's been helpful for it at the same time. I can imagine that I mean once you get started with with one customer it's it's a learning experience not only for the customer itself but also for you guys for high touch and for the developing that product. I mean as you say you guys got were stuck at a room to product development during covid period trying to come up with best best idea to move forward with high touch. But now you guys are also in a position to collect a lot of feedback from your customers into further further develop individual products within high touch high touches offerings it's. Yeah so kind of to round off slowly because I know we're kind of reaching reaching the time. You guys have come a long long way for me personally correctly if I'm wrong the first time I caught touch with high touch was around reverse ETL you guys were doing a lot of reverse ETL work. We always have the CDP vision but yes reverse ETL yeah exactly and now you're at AI decision and you're connecting to retail media networks I mean it's. I think this is kind of the new era of CDPs that were entering and you guys are also a carving out a unique space beyond traditional CDP capabilities. I mean was this to what point was this always part of the plan. Yeah we're really good question so I think the plan was that and still remains that. There are a lot of opportunities that data and AI I think can be disruptive to how business drives growth and the steps the order which features rebuild all that stays dynamic. And all that we keep learning every year we didn't think we put out some products like AI decision or master in that exact way when we started the company. But did we feel like there's a general going to be a general opportunity around providing more intelligence to marketing teams on top of this 12th of data. That we have now that we're connecting to their snowflake their data bricks their Google BigQuery yes and we decided to figure that out along the way. So I think never say never in terms of the opportunities out there but we really feel like there is a lot of potential to build a hub for marketing and digital teams for all things data related and help them really extract insights from that data and use it to inform every part of their marketing strategy. So that's kind of the way we're thinking about it and we'll pursue more opportunities along those lines. No exactly and I mean you guys have all the markings all the products all the modules to become a pack of children what you guys are remaining composable so a lot of respect for that and just kind of round off the question and tell me back off if you say Matt you're going a little bit you're overstepping a little bit. You guys started with a kind of very provocative marketing. I think one of the most popular articles you guys have erode was the friends that friends by CDP in my eyes you guys have now become really you guys establish yourself as market leaders. I mean how has the this change kind of matured your company from being kind of seeking the limits of marketing and now shifting to a more market leader. I don't want to say authoritarian but kind of you guys are becoming an authority within CDP world. Yeah good question. So I think when there's something really provocative that's worth saying we might still say right I think no one was really talking about no one was really talking about the shift that we're seeing in the market that we were seeing in the market a few years ago and now has become more of a norm around CDPs sitting directly on top of companies at a warehouse. And I think some of that catchy maybe provocative maybe childish depends on you ask messaging helped grab the attention of a lot of people online. Oh absolutely did open my eyes at least in a positive way. And I think now we have a lot of customers were focused on educating them on new use cases and putting out valuable content on that like helping them understand how to pursue opportunities like retail media or help to how do you think about ways AI can actually drive impact today. Today that isn't pie in the sky and isn't futuristic that's where we look at it is practical use cases we're going to be putting out a lot more content about that so I think the right content it's the right message for the right time. There's nothing so you're provocative we might do in the future but right now I think we're really just focus on educating the market on we think there's a few topics that people know they need to pay attention to they don't need anyone being provocative about it people want to learn about AI people want to learn about retail media. People want to learn about composable by the information out there is not reaching everyone and there's so much more information that can be written so that's why I think there are a few companies out there that definitely stole your stole the magic a little bit to a couple of days ago with the treasure data's trade up program. The right time just with all the market in turmoil at the moment I think they played into it as well by kind of not necessarily kicking people shins you did get the attention and you did open it up a lot of eyes like I said in the beginning I was quite naive around composable drinking from the wrong cool it completely convinced that package will be the way future but slowly but surely that's. It's left my system and I'm seeing the benefits of composable not to sound like a follow here but I'm very respectful of the work that you guys are doing in the community and I look forward to seeing high touch take a kind of a command position and in teaching us more about how to do better with CDPs and how to get more value out of our business using solutions like high touch. So thank you very much for that and thank you about a lot for your time today I know you're a busy band and I look forward to that demo of AI decisioning and what you guys have more in store in 2025. Awesome thanks for having me.