Drill to Detail - Drill to Detail Ep.106 'Customer Studio, Hightouch Performance and the Evolution of Reverse ETL' with Special Guest Tejas Manohar

Episode Date: June 22, 2023

Hightouch co-Founder and co-CEO Tejas Manohar returns as special guest to talk with Mark Rittman about the reverse ETL market today, the evolution of the composable customer data platform and new featured in Hightouch to enrich customer profiles and drive personalization across marketing campaigns.Reverse ETL is Dead (Ethan Aaron LinkedIn Post)Customer 360 Data Warehousing and Sync to HubspotYou don't need the Modern Data Stack to get sh*t doneHightouch Customer StudioHightouch Personalization APIHightouch Match BoosterWhat's in Store for Data Teams in 2023?

If you notice, there was a LinkedIn post a couple of weeks ago. Basically, the gist of it really, and I'll put a link in the show notes, was that reverse ETL is done and it's kind of like people are moving on from that. It's a great question. So I did see that. I forgot to respond on LinkedIn, but maybe I'll go do that afterwards. People from the outside might think reverse ETL is dead. What I can tell you for sure is that reverse ETL is not dead.

Welcome to another episode of the Drill to Detail podcast, and I'm your host, Mark Whitman. So I'm really pleased to be joined today by returning guest Tejas Manahar, co-CEO of Hightouch. So Tejas, welcome back, and it's great to have you with us. Thanks for having me, Mark. I think I was on here about two years ago, so it's awesome to be on the show again. So Tejas, for anybody that didn't listen on the first episode, tell us who you are and give me a quick elevator pitch for High Touch. 100%. So thanks for having me on the show again. I'm Tejas, co-founder and co-CEO at High Touch. Before founding High Touch, I was an early engineer at Segments. So I've kind of been in the customer data space for a while. In terms

of what we do at High Touch, at a very high level, simple level, we work with companies that have tons of valuable data and insights inside of their data warehouses, like Snowflake, Google BigQuery, Amazon Redshift, Databricks, whatever technology you're using. And we help you activate that across the entire business. And what I mean by that is we've come up with this technology called Reverse ETL that allows you to easily take data from the data warehouse and bring it to all the different applications you have around your company. So Salesforce, Facebook ads, Gainsight, Braze, any of those systems, and allow it to drive action like automated personalized email campaigns or ads or telling your salesperson the right signal at the right time. And then the other thing we're known for is coming up with this idea of the composable CDP. So for the last few years, I've been writing about how, and not just writing about, but partitioning it at various companies, even in the Fortune 500, like Warner, PetSmart,

the NBA, basically helping these companies use their data warehouse as their CDP, instead of having to invest in a totally separate platform as their customer data platform or something like Segment. Okay. And for anybody that's thinking, well, I've heard this story lots of times now, this must be old hat. The thing that people don't probably realize is that you're one of the first people to actually have a product in this area and start talking about this. So tell us a little bit about, so you mentioned you're a segment and you were probably playing down playing your kind of role there as an engineer i think you were

you're responsible certainly working on on personas um how how how did give us a positive history of how you came about to to suppose have the idea around high touch and how that was fed really by ideas you had in the past and inspiration you'd heard from customers 100 so um yeah i was at segment for quite a while before founding high touch uh about four years um i left about he kind of joined when the company had about 10 engineers or a little bit less than that and then left when it was about a year and a half by before the uh three and a half billion dollar acquisition by by twilio uh which was which was an awesome outcome for the company. Good timing then. Exactly. But yeah, in terms of Segment, one of the last things I worked on there was basically the personas product. So Segment had this amazing infrastructure, which allowed companies to

collect events from their websites, from their apps, from their backends, and collect all these events in a certain segment way. So calls like track events, identify events, page events, different stuff like this, a kind of a standardized schema. And then Segment would, you know, the core product would bus out those events to a bunch of different tools that Segment's integrated with. And what the personas product was meant to do, actually, at the time, it's funny, because online i think you know segments regarded as the number one cdp or the first cdp or all this sort of stuff but

at the time you know there was this whole trend going on in the space called cdp that that segment wasn't actually playing in and what a cdp did was um allowed marketers to come into the tool and do things like build audiences build build user traits, sync them out to different tools. And it's kind of an obvious opportunity for a segment to expand into that space, just like many other tag managers or customer data platforms or solutions had, like MParticle, like Helium, like TreasureData. And the idea was that a CDP would kind of be the global source of truth for your company. So you would put all the data around the business into the CDP's format, so segments, users, and events and stuff like that. And then from there, you'd be able to do all the activation across your business from this one platform. Now, I would say it was truly an amazing promise. I

mean, customer data problems are huge, right? And if you can actually buy one platform to go solve all those, that's truly incredible, especially if it can work out of the box. What I saw in reality, though, is like segment and the segment model. So I think you've seen this with other players in the space like Rudderstack and so forth. It's not just segment, but the general model of handling customer data in a cookie cutter kind of way works really well for small companies or simple companies, let's say like a Shopify store or a simple e-commerce platform. But when you get into a larger enterprise, there's a few issues that come up. And those issues are namely one that, you know, Segment and Rudderstack and all these solutions have to understand your data to be able to activate them. You have to put data in their format.

So how do you want to track user event? How do you want to track a product being out of the cart, et cetera? And, you know, that's just a very tricky thing to do in a really large enterprise where there's all types of different data. You know, your business is unique and so is your data. Some of our customers like PetSmart track things like their actual pets in their loyalty system, when the pet's birthday is, or a bank customer of ours has many different accounts that a customer can have that can have many different balances, all these different entities. So that was one of the problems. And then the other problem was just the time to value. I mean, to actually set up this platform

for your marketers, which delivered on an amazing promise of, you know, time to value for marketing campaigns once set up, you actually had to go through a huge implementation process. So it could take, you know, six to 12 months for many clients and just get that audience builder off the ground. And with high touch, those are the two things we've really differentiated on. So we can tap into any data in the data warehouse, no matter the format, no matter the structure, right away and handle those complexities of your business, like bank accounts or pets or households or point of sale systems. And then we can also get spun up right away. If you're using BI tools like Tableau or Looker, you can start building audiences and right away. And of course, it might not be your perfect customer 360, but it's kind of a new incremental way of looking at that problem.

Maybe just let's start by just painting a picture of what the original incarnation of high touch was really when we first went to market. And then we can actually then start talking about what's new really since we spoke. When we first went to went to market a few years ago i think i think you were maybe one of the first people who checked out our product mark so thank you for that um but it was really a simple product i mean you come into the app hide the chap you connect your data warehouse so you know you add snowflakes or whatever it is the data source similar to how you would in tableau um and then you you put in a

sql query and you say i want to sync that s a SQL query and you say, I want to sync that SQL query. Let's say it grabs all the users that recently logged in. I want to sync it to a tool like Salesforce or Facebook ads. And then you just fill out a little mapper that says I want these columns in my SQL query to go to these columns in a downstream tool. Set a schedule and schedule it. And then your data is kind of flowing a live

sync between data that's populating your data warehouse under some conditions in the SQL query into a downstream tool. And something interesting is like, still today, you know, if you sign up for high touch, you can set up a workflow like that in your company really, really fast. We're actually just tomorrow, we're kicking off this webinar series called the 23-Minute CDP. And it's kind of a funny thing in a lot of ways because a full-on CDP is used across the whole company. It can't be set up in 23 minutes. But every company, I truly believe, can drive tremendous business impact in a couple hours of getting into our platform, which is one of the biggest selling points, especially in this market. And the reason we picked that 23-minute thing is

it's actually a stat we developed off analyzing our self-service funnel of people signing up on for a trial. 23 minutes is actually the average time it takes a new signup to go from connecting their data warehouse to creating their first sync in high-tech. Interesting. Interesting. So we'll talk about, I suppose, new features in the reverse ETL part of the product in a sec. But if you notice, there was a LinkedIn post a couple of weeks ago

from Ethan Aaron, that was it, talking about reverse ETL being dead. And he said in the post, all the vendors have moved on. Instead of being use case agnostic, helping data teams move on. And basically the gist of it really, and I'll put a link in the show notes, was that reverse ETL is done

and it's kind of like people are moving on from that. But you've actually invested quite a lot of time and money, I suppose, in new and additional features around reverse ETL. So what have you done in that kind of part of the product? And is reverse ETL no longer a kind of like a, an interesting area to your clients really? It's a great question. So I did see that. I forgot to respond on LinkedIn, but maybe I'll go do that afterwards. It's funny, you know, it's hard to do marketing as a startup, especially when you're, you have multiple products, multiple, multiple kind of points you're trying

to drive to the market in this whole movement of using the data warehouse, reverse ETL, composable CDP, thinking in a warehouse native way for things like Customer 360. So, I mean, a challenge that we've had is I think we've leaned super hard into pushing all this content around using your warehouse as your CDP, the composable CDP over the last six months. And people from the outside might think reverse ETL is dead. What I can tell you for sure is that reverse ETL is not dead. Actually, over half the kind of deals that are coming into our business are for just using the reverse ETL product where there's so much value to drive from taking data from the warehouse and

using it to power workflows across the business that happens in your SaaS tools, your ad tools, Slack, spreadsheets, whatever it is. So it's definitely not dead. If anything, I think it's actually become somewhat of a standard lingo amongst data teams. I've run into data teams, even in the Fortune 500 that have built, quote, unquote, a reverse CTL script internally. So it's quite interesting in that sense compared to a few years ago when we were just coining the term and getting it off the ground. But I'll tell you in terms of

the why we've invested in our reverse ETL product so much is because it's an incredibly large part of our business to date. And even the customers who use us for a CDP context are still reverse ETLing around their company. Even companies who use big CDPs are still using reverse ETL. But the ways we've invested it are, I would say, kind of twofold. One is making it easier to activate,

do your data off the warehouse, and just easier to use the product. So that kind of goes between two things. One, there's tons of UI improvements, UX improvements that we made over the last couple of years. We really look at that stat, like how fast does it take a new user of the platform once the warehouse is connected to run their first activation? And we optimize that in the platform because we think that's super important. It's not just your first use case, it's your nth use case. So that's a big improvement area.

Another one is just like integrations with BI tools. So you can just go directly from a Looker report or a Tableau report now and say, boom, I want to... Or Sigma report and say, boom, I want to sync that data. I just analyze and drill down and found my most valuable users. I want to sync that to Salesforce for outreach so that my sales team can act on it. I want to sync it to Bray so that we can send out some re-engagement campaigns to these users. And that's a really, really quick workflow in the product. And then the other thing I would say on the idea of making it easier to use is making it easier to use for certain use cases. So we really believe in not just understanding how our customers use reverse ETL, but understanding what they're doing more broadly

with their data. So we go and talk to customers who use all our integrations like Postgres, MySQL, DynamoDB. And we were like, what are you guys actually doing with these integrations? Why are you putting data into the services? And something we realized is they're trying to build APIs off the data warehouse to do things like personalizing their app. So we actually built a first-class personalization API for that purpose that handles a bunch for you, like multi-region, global deployments, getting really fast latency and caching. And then just at a high level,

the last area is really reverse ETL is making the data warehouse a core part of the production pipelines of a company and the data team a core part of that. So we think of what does that mean around the company? Well, it means you might want things like version control with Git on these pipelines. So you can roll back changes if there's an issue. You might want things like building your own integrations and extensibility, which is very common in developers and software. You might software, you might want things like data dog integration. So just like anything, you can name it, bringing things from the software engineering workflow to the data teams is another area of focus for us.

Okay. Yeah. I mean, certainly it struck me a few, the fit and finish really the product had come on a lot really since, I mean, it wasn't bad in the first place, but certainly it's even better now, like the ability, like you said, to define a look, and then the look becomes the source of the audience, for example, or source of the kind of the data extract you're going to, you know, reverse ETL into HubSpot, for example. That was really good. I like the Datadog integration.

I thought that was good. You know, make it, as you say, part of the infrastructure. And so the personalization API, I mean, so just on that point, so I worked on that sort of thing in the past a bit. And one of the things about that was that you need to have very low and predictable latency for the queries. How are you handling that? So you've got a website that needs to be able to refer to the personalization API in real time to personalize content. How are you ensuring that the latency is low enough really for that? Yeah, a hundred percent. So we've built the API internally. It's backed by a kind of distributed key value database on our end. So, you know,

we make it really, really seamless to take data from the data warehouse and expose it in the personalization API. So, you know, let's say you have something like a propensity to buy. What's the score, propensity score that this customer is going to buy another product between zero and one? You have that in your data warehouse. You want to use that to offer a discount to low propensity users and see if you can push them over the line. Well, with Hightouch, you can easily offer that discount by reverse ATLing low propensity users into Braze or Salesforce Marketing Cloud or Facebook ads and running campaigns of those customers. But let's say you want to do that

in your e-commerce experience or in your mobile app, you need an API, right? That's the fundamental building block is you need an API that can take a user and return you if they're propensity. And what Hightrush allows you to do is stick a SQL query that pulls all the users and their propensity scores and click. I want a personalization API off that. And then now you have an API with really low latency that you can use in your apps.

Now, in terms of how we actually do that, we don't query your data warehouse or issue that SQL query whenever data comes in. We're caching that data in our own distributed key value data store in the backend. And we have kind of like almost like a CDN

in front of the API so that basically, you know, if someone hits it in India and then someone hits it in Australia, you're going to two different servers and not your US servers so that it's really, really low latency,

if that makes sense. Okay, yeah, it makes sense. Absolutely. So there's all the stuff that I saw to do with reverse ETL, but then I noticed there was a new business tier and a customer studio feature in the product. So just give us a bit of a background of give us a bit of a background and high level kind of, I suppose, overview of what customer studio is within, within high touch.

100%. So I imagine this podcast is mostly listened to by data professionals and, and, you know, data teams. So at a, at a high level, I mean, customer studio, we basically founded it because we realized, hey, there's more people Studio, we basically founded it because we realized, hey, there's more people around the company than data people who know SQL that want to do this reverse CTL stuff. That's not actually how I'd pitch it to a marketing team, but basically, marketers, they want to get their hands on the data. They want to say, hey, I want to build a custom segment, find customers who recently didn't make a purchase, and I want to send it to Facebook. I want to send it to Braze. I want to send it to these tools without having to go to the data person. grows farther and people start working together through sprint processes and tickets and all these things that no one really likes in the end of the day. The future is self-serving data, and we wanted to enable people to do that directly in the Hydros platform.

So basically, Customer Studio, it started off as audiences. That's what it was initially called, but we renamed it as we started to think broader. But the idea is you can come into the platform and get this visual UI that's quite aesthetic, easy to use, very marketer friendly. It's been a huge focus of ours from the beginning. And that visual UI lets you do things like build audiences, just using kind of drag and drop filters and stuff like, I want to find people who went to the website but didn't check out, just stuff like that. And then you can preview it, you can see some stats about it, and you can activate it by syncing it to all the different channels that we can reverse ETL to. So any ad network you can imagine,

any ESP, marketing automation platform. And it's fully a self-service flow for the marketer so that's really the differentiation there with the reverse ETL product okay okay okay so that that's that that's and that I suppose conceptually that's quite familiar although as you say new to high touch um the product but then but then I noticed that you've got you've got things in there around say AB testing so randomized audience splits and high-touch performance in there. So maybe, again, for anybody listening, well, people listening, what is that and why would that be relevant to reverse ETL market and marketers and how does it work within high-touch?

Yeah, 100%. So, I mean, take things like the randomized audience splits or performance. There's all these questions and tasks when it comes to data. Let's say specifically marketing data within the context of Customer Studio. There's all these questions and tasks that data people can do really easily in SQL. They're like, oh, yeah, I want to do a split test. Let me use this entile function or percentile function or whatever it is inside of Snowflake. Or, oh, yeah, I want to see how this audience is performing.

Let me copy paste the SQL query into like another report and tableau and see how the performance is. Easy questions for a data person to answer. Incredibly, you know, long waiting, long cycle questions for marketers to answer. Because sometimes they don't have those technical skills to operate in the data directly and they have to go through the ticket queue or ask a question or should be responded in a day or an hour or whatever it is so we just you know look at these most common tasks talk to marketing teams pretty simple ask them what they're waiting on data from a data perspective what they wish they could do themselves they told us in addition to activating audiences they want to build experiments themselves They want to be able to split their audiences, send them to different tools, have a control, do things like stratified sampling, know how the audience are performing in the end of the day, understand the performance of it as well without waiting for another report to be built. So we've just encapsulated those workflows in the high touch

product. And I think that's, again, you know, compared to like, let's say a CDP or like a more specific solution. High touch, we can really do things like showing the performance of campaigns really easily, because all the data is in your warehouse. Like when you think about lift data, attribution data, you know, pulling back the stats of how your performance campaigns are running from something like Facebook or Google, that stuff usually lives in the warehouse. If it doesn't, there's a great ecosystem of tools to bring it in there,

like funnel and super metrics and five trend, or Google can do this for free for big query customers, actually. So, yeah, I mean, that's, it's, the problem is that not the data being in the warehouse, it's really, it's not accessible to marketers. Okay. So this doesn't cover sort of incoming traffic to the website and splitting it in real time. This is about taking an audience

of kind of users or customers or whatever in the warehouse and applying a split to that, isn't it? Is that correct? Yeah, that's kind of spot on. We're really open to any data from any data source that goes into the warehouse. And stuff's getting into the warehouse a lot faster than it was years ago, as we both remember, Mark. I think you've been at the data space for a long time, right? Amplitude, their sinks go to the warehouse, I think, every 10 or 15 minutes at this point, in terms of new events. With Snowfall customers, we've seen people even getting it even faster. And yeah, I think that the whole world of how quickly you can get online data into data warehouse is really evolving. I've even seen stuff in the minutes. Okay, okay. So and also you've got in that you've got in the product a way of

visualizing the uplift and the revenue uplift in there so i guess it's probably early days of the product and and so on but are you looking at things like i suppose statistical significance and that sort of thing in there or is it really more of a a basic feature at the moment that we're built out over time how does the how does that kind of part of it work the the uplift the uplift kind of visualization and so on? Yeah, 100%, 100%. So there's kind of statistical nuances and significance that need to be thought up on both sides. One, in terms of picking the samples, and then two, in terms of basically understanding if there's actually significant lift. The first side is something we 100% play in.

We have a lot of different sampling methods. Random is obviously the default one, but there's stratified sampling, all sorts of different things. In terms of the second one, helping you understand, was this a good experiment run in the first place? We have some basic capabilities there today,

but that's an area that, frankly, and happy to be open about this on the show that we're investing more and more in helping, you know, marketers understand things like incrementality, you know, is there actually a difference being proved by this campaign? Or do I have too many conflating campaigns running at once? So we really think if companies are able to centralize, you know, all their activation and analytics around one core platform around the whole company, the data warehouse, then it provides the right breeding ground to build these capabilities. But we're still early in our journey here when it comes to analytics.

And I would suggest that's where partners come in as well. Yeah, 100%. But no, seriously, I mean, I think that's where a good services partner can really kind of make the difference as well, because the tool is there, it's got the capabilities. It enables these things to be sort of, you know, much more accessible and much more kind of, you know, easy to do. But then a good partner is the one that says to you, actually, it works with you to say, right, is this the right way to do it? How do we interpret these results and so on?

Totally. I mean, technical literacy and, you know, experimentation literacy is a level beyond that too, I would say, is not available in every organization. And, you know, and experimentation literacy, it's a level beyond that too, I would say, is not available in every organization. And most of the largest organizations we work with all need partners to help with those things. But if we look at the, I suppose, looking at the CDP market in general, right? So we have, since you came on the show a while ago,

we've had a segment on the show talking about reverse ETL and Vudderstack and so on. And there's also sensors out there and so on. How do you see, I suppose, the CDP market going at the moment and the reverse ETL market? And is it one that is kind of growing? Is it one that's getting differentiated? I mean, how has it changed since we spoke, do you think, a few years ago 100 percent um so answering the simplest of questions first is it growing yes it's definitely growing i think the reverse etl market's growing i think the the cdp market is growing uh both traditional as well as the kind of new new cdp uh composable cdp um all this stuff is growing the demand to use data and derive outcomes with data has never been

higher and more serious in terms of the impact to businesses than it is today there's a lot of change in the ecosystem right third-party cookies going away there's there's lots of lots of tailwinds here but in terms of how we see you know differentiation how we see the market evolving. I think Hightech was founded on the principle that the data warehouse is, I won't say it's going to be the single source of truth in a business, but it's going to be the closest thing to it. We actually are putting out a blog on this, but I think the whole idea of a single source of truth is kind of flawed in a lot of ways. It's always a journey, but the data warehouse in a business, no matter if you're a 200-person company or a 20,000 or 200,000-person company, is going to be the closest thing to that single source of truth and

is also the most powerful technology when it comes to operating on data. And our goal is really to democratize access to that to the whole business, especially marketing teams. So I think the part that's become obvious over the last couple of years is that the data warehouse is an important piece of the puzzle that you can't ignore. Companies like Braze or Salesforce Marketing Cloud that may have previously felt it's important to lock in their data and not make it accessible in people's data warehouses, go build your own connectors for that, are now leaning in and just saying, hey, if you want to get your data into Snowflake from Braze,

click a button, we'll make it happen. And that's a big trend across the whole SaaS industry. And then, yeah, as you mentioned, the CDPs, Segment, MParticle, Treasury Data, no matter what they are, they're not sitting still. They're recognizing their customers are demanding some sort of integration data warehouse, and they're adding those kinds of capabilities. But I think what ultimately will help HITA shine, and we continue to see in deals where we compete with the segments

of the world, the CDPs of the world, is that there's a difference between being integrated with the data warehouse and being centric, warehouse really, you know, being able to use the data warehouse to the fullest. And let me give you an example of what this means. So Segment, while it has an integration with the data warehouse, can only understand data that was built for Segment, or is, you know, in the Segment format, you know, goes back to a track event, or an identify call, or a page call, or, you know, product added to cart kind of event of the e-commerce schema. That just doesn't work in a really large company scale, especially with everyone's own unique complexity to their data. So how I took this kind of secret sauce in the customer studio, as well as bringing it to more and more parts of the product, is being able to sit on top of any

data in the format it already is, and making it usable to marketing, which is incredible benefits in the time to value part, but also allows marketers to personalize at a level they could never do before on things like, hey, I want to not send an ad campaign to these customers. If the product that we're advertising is actually out of stock, I want to wait and then send them an email when it is in stock. And just stuff that people couldn't really do, journeys that people really couldn't build before due to a data availability and a data understanding problem. And that's where we'll continue to differentiate and build more.

So the market has been interesting recently, I suppose, for sort of SaaS businesses. And you've got, maybe I suppose, maybe there's three types of SaaS business that I'm referring to here. So you've got like yourselves who are very, I suppose, specialized in a certain function and arguably, you know, very much aligned with a certain kind of set of users and a certain set of use cases where there's value you can provide and so on.

Okay, so you've got, I suppose, not to say niche providers like yourselves, but maybe sort of focused or whatever. And then you've got maybe the likes of maybe Segment that's covering a lot of different sort of, you know, a lot of different sort of markets and so on there. And then you've got companies like, say, DBT Labs, where arguably it's more kind of infrastructure. And certainly in the last couple of weeks has been, you know, I suppose over the last six months has been layoffs and all that kind of thing. So I suppose is high touch the kind of company is for a reason the sort of single focus on one certain thing um you know which companies and which sass businesses

do you think will survive going forward and what's the market been like really as a founder over the last kind of six months if you're asking yeah yeah totally great question um and this is a topic that honestly more companies should be should be talking about totally. Great question. And this is a topic that honestly, more companies should be should be talking about. So first thing I'll say is a clarification before moving into the market question is just as a clarification. I don't really, you know, believe in this whole, everything's unbundled forever, you know, go pick your best of breed software and string them together kind of world. I do believe that data is a multifaceted problem and the data team will buy tools like dbt,

marketing team and martech team and marketing analytics team will go buy tools like high touch and those tools will work together regardless of whether they talk to each other. But I don't necessarily believe in these ideas of like, oh, you know, I need to build my whole company around this like modern data stack

or modern marketing stack and plug all these different tools together. And I think you can kind of see that in our product direction at Hikage. Actually, you know, bundling is generally good for buyers. And that's why we have the reverse ETL, the personalization API, the customer studio. We recently launched a product

called match booster that allows you to do live ramp like things so you know let's get on to that let's get on to that at the moment actually yeah yeah for sure happy to chat more about that but um you know bundling is a good is generally actually a good thing for customers i frankly do believe that what i don't believe is good is is good is making you buy all the products at once, making you buy stuff you don't need. Just like these software suites have become impossible to understand, or actually, in addition to that, lower quality products, companies where they focus on a lot of things and they can't drive the right focus in each of those products.

But it's not inherently bad at bundling. I would say the two things that really matter in this market as a startup that's been adapting and selling over the last 12 months in the downturn with companies across the stack is really focusing on one, time to value. And then two, what is the actual value you drive? So companies that just provide infrastructure, they can do well. They can actually do well, even infrastructure companies like DBT,

and they are doing well. That's because they do provide tremendous value to data teams. And they provide a step function on lock that just wasn't there before. But they need to get really good at conveying that value. And that's why you've seen us focus so much on the marketing use cases and what we're driving in companies in our marketing itself.

And that's an adjustment that we've made. That's an adjustment Twilio has also been making. And you can read about in their quarterly reports online. It's kind of an adjustment that everyone's making across the stack. The second thing I would say is time to value. So let's say you deliver value. Now, how quickly do you deliver value? Because a lot of people are only focusing

on the most important initiatives right now and don't want to start new, big, huge ones. And that's actually where high-touch and the composable CDP and reverse ETL has really been picking up because we can deliver value really, really quick and prevent you having to do this huge customer 360 or CDP initiative.

So that's where we've been doubling down. Yeah, that's interesting. So we've been involved in a few initiatives around customer 360 that it's easy for them to either get mired in too much detail or kind of, I suppose, just never really happened really. And I suppose it's how you focus on that it's how, how you get to say, how you focus on that time to value, how you focus on getting a certain sort of thing done and getting, and getting it clear early on where the value is. I mean, how, I suppose, how, how, how, how much activity do you see in the new CDP market at the moment? Is it happening,

or is it more a case of kind of finishing off existing ones? I mean, what's, and do you see, have you seen any kind of, I suppose, uptick in the market recently? Just curious. Yeah, yeah. Great question. So as a fast growing startup, you know, perspective is always biased

due to the growth rate of the company in general. But overall, I do see that the CDP market is growing. There's more interest in CDPs, more RFPs being opened for CDPs than ever before. But the ones that are getting closed unsuccessfully, getting shut down effectively, are the ones that are just kind of oriented around wanting a bunch of different features. You can see these RFPs when you get them. You can immediately recognize them. They just have all these words from the CDP vendors.

Identity resolution, third-party data, first-party data, this, that, you know, just pounded, you know, with buzzwords versus the ones that come in and say, we need to increase repeat purchase of our customers. We have data in our data warehouse that we think can lead us to propensity and the highest value customers to focus on. We need to talk to software vendors and partners that can help make this happen. So really what we do is, we don't believe it.

No one's stupid, right? These marketing teams, marketing tech teams that are opening to CPR, even the ones filled with buzzwords are not stupid in any sense way. All these capabilities are very valuable, but in this market, for everyone's benefit, for our benefit, for the customer's benefit, we have to figure out what are the

actually most valuable things that can drive the most value in the lowest time. Because that's what matters in the end of the day, especially when people are strapped for resources. So we help our customers with that and we identify those as fast as we can while making it clear that we are, you know, a good long-term solution for the client due to our flexibility at the platform. So that's what we've been doing. So lastly, I want to talk to you about really in terms of the product is something, again, you mentioned, I think you mentioned earlier on audience match, right? So that sounds very interesting to me. So talk to us about that feature and how it works and what problem that solves. 100%. So people have been using reverse ETL and CDPs, composable CDPs for advertising for the longest time.

The basic use case is you build an audience off your data, like customers who signed up but didn't purchase, customers who dropped off, etc. And you sync them to an ad platform where you can say, target these customers and put this much budget towards it. Or better, don't target these customers. They already made a purchase. Don't send them more ads. It's a waste of money. And how these capabilities work is that they rely on Facebook, Google, TikTok, whatever, being able to match the identifiers you're sending them with the identifiers they have. So an email from your database matches an email in their database or phone number to their phone number, right? And that's how these capabilities work. And that's how they're able to find who to show or not show these ads. But sometimes, you know,

a lot of times, actually, I would say companies don't have the right identifiers to, to drive those match rates. So I'll give you two examples. Um, one would be like a B2B companies. They have this problem all the time, which is, you know, they're collecting people's work email addresses, which you can use for ads on LinkedIn, but you just don't match many customers. When you send a list of work email addresses to run ads on Facebook, but you just don't match many customers when you send a list of work email

addresses to run ads on Facebook or Google. Personally, email address does way better. So what Matchbooster unlocks is that we've actually made five, seven plus partnerships, built sort of a identity graph on our end of those data partnerships, sources like Newstar, Axiom, all the types of big, big data partnership companies you can imagine. And then we basically allow it so that when you create a sync in Hightouch to an ad network, we can actually enrich the emails, phone numbers, names, et cetera, you have with additional identifiers from these data sources, all merged together and cleaned up by high touch, all out of the box and send the enriched data to the ad platform, which means, you know, you can start using work emails to target on Facebook. Something you couldn't do before. And Facebook's way cheaper than LinkedIn for a lot of companies or use emails to target on TikTok, even though it's just a mobile app and TikTok is really only good with phone numbers. So it's a pretty powerful capability.

Okay. So what's the privacy angle on that? How does that work in terms of people's expectation that you do that and so on? A hundred percent. So now actually storing that third-party data and managing it does come with a fair amount of liability and compliance overhead. So that's something we take off our customers back, if that makes sense, by managing the relationships with these third-party data vendors for them, by having the contracts, relationships, secure environment all set up properly on our end. And that's actually a big reason that companies use something like a live ramp in the first place and want to outsource this task, whereas they often build automation

around first party data internally. So that's one part of it. The other part of it is that we've kind of audited the whole data supply chain as well. You know, what data can be used for what type of activation, which data sources are compliant with things like GDPR. And that's something we are experts in and are able to help our customers with, if that makes sense. Fantastic. Well, so just to wrap things up then, what's coming on the roadmap then with Hightouch that you can talk about? And on that same topic, what is the next thing that you're thinking about trying to solve really? Big picture, it's really, you know, there's really one big thing driving the roadmap this year, which is how do we help the marketers get their hands on data and use data across their whole workflow? From building audiences to managing them,

to activating them, to analyzing how they're doing. So I think you'll see more and more capabilities coming out from Hightouch to make that process easier, easier than ever before, and kind of a streamlined workflow for the marketing teams to not have to rely on data teams or technical people to get their jobs done. And then the other thing we're thinking about is how do we help companies get their data warehouse prepared or use data

they already have even more easily inside of high touch? Because data preparation, data engineering, it's not always high in resources at a company. It's in fact high in demand. It's the opposite. So we're thinking about those problems and, you know, how do we give a framework for that to our clients? And what's the obligatory AI feature that you've got coming along sort of soon then that you must be putting into the product? Or maybe being slightly cynical, you know, where do you think AI fits into what you're doing?

Totally, totally. So, I mean, I think the opportunity for using large language models to build audiences and to do analytics is actually quite obvious. And I do think there's something powerful about what we've built at Hightouch that can actually make that work versus not work. There's a lot of startups online that say, I'm going to generate SQL queries for you using chat GPT. And that's not just going to work on top of a data warehouse that has like 10,000 or 100,000 tables in it. It's going to query the wrong data. It's a total mess to navigate something like that,

even as a human. It's not a computational problem. It's a data cataloging problem, if that makes sense. So where I do think it can work is on top of whitelisted data, on top of the BI reports, on top of the models, on top of the exact understanding of what data matters and how does it relate to each other. And that's what we have in Hightouch

in order to set up our audience builder. So I would say the obligatory response is that we are exploring those things. Okay. And so how do people find out more about Hightouch then? 100%. So you just go to high touch.com um so the word high like a high five the word touch like i don't touch um and then go to the.com and we'll be there okay and will you be at snowflake summit in a few weeks time i will be at snowflake so if anyone wants to

grab a coffee or go to an event together there, hit me up. Fantastic. Excellent. Well, it's been great speaking to you again, Tijaz. It's really good the way the products come along and it's great to get your insights into the market. Thank you very much for everything and stay in touch and best of luck in the future. Thanks, Mark.

I appreciate you having me on the show. Thank you. you