John Kosturos (00:00)
All right, welcome everybody to another episode for the SpringDB Data Exchange. We've got a very exciting guest today, Nick Weldon from 5x5. How are you, Nick? Doing well. Yeah, so Nick and I have gone a couple of years back and...
Nick Weldon (00:13)
Good job, how are you?
John Kosturos (00:20)
They're doing some pretty amazing things over at 5x5 and maybe we can start out Nick just by diving into a little bit of your background.
Nick Weldon (00:28)
Sure.
And five by five story is a very kind of organic bootstrapped story, is just similar to my background. So I have a background in bootstrapping half a dozen or so startups, like anything, some worked, some didn't. One of the themes with all of the, two of the themes that all the startups that I've worked on is as much as I could, I tried to stay away from formal capital, specifically venture where I could. And all of my startups had some sort of data backbone where data was a critical
elements to the service or the product.
And so rewind like six or seven years ago, we were building a solution where we were trying to create a shared experience graph, where we were trying to find all the shared commonalities between people. And then that could be used in like a sales environment. So when you get on the phone with somebody, you know, you have kids in college or you're both into snowboarding or your fans are the same sport or fraternity or whatever it may be. So in that pursuit, that is a very data heavy product.
So
I got linked up with Brian, who's how we got connected, Brian Perks. And in that, we started figuring out how to get some of our own data, first party data and some scraping and third party data. However, for the graph to work efficiently and effectively, we needed more data.
and in order to get that data, we were going to have to raise a funding round, to be able to just get the capital to secure the data licenses for such large scale data, especially back then it was even more expensive. I would say there was less just commercially available. Large scale options. so in that Brian, I had this idea of we knew a few companies that weren't data companies, but they were platforms where their platforms generated some data, but they also needed more data for their platforms to work. And the thesis we had is what if we could get
John Kosturos (01:51)
Yeah.
Nick Weldon (02:05)
half a dozen or a dozen of us together where we could all share some of our first party data into this collaboration, combine all the data together, make a big data set that we could all use to improve our products and services without enduring funding rounds or pushing funding rounds wherever it may be. So it was kind of like a side project to some extent for the main project.
Well, within a few months, what was supposed to be a side project was five hours a week and then 10 and then 20 and then 30. And then we started spending more of our time on the side project than the main one because it just had so much energy and excitement and people were coming to us. It was word of mouth spreading. And as you know, John, like with any startup, like product market fit, you know, and it kind of is happening. We were feeling the early signals of it.
John Kosturos (02:47)
yeah.
Nick Weldon (02:49)
And there were a bunch of companies that basically are out there where they were generating maybe small to medium amounts of first party data, but had an exponential need for additional data.
and wanted to collaborate on the concept. So that's where 5x5 was created. It's a data cooperative for tech platforms where third party data is a critical component to their product or service. Where these platforms on their own don't generate enough data for their solution to work independently. And they need that supplemental data.
to accelerate either the usability of it or to build new features and solutions. In addition to that, there's a lot of data out there that can't be sold based off of user permissions. So a lot of these platforms can't physically sell some of the interaction data they might be having on their platform, but they can barter and use it to...
John Kosturos (03:25)
Yeah.
Nick Weldon (03:39)
in an exchange format. So there was a whole bunch of data, valuable data, that was kind of in this platform ecosystem that never made its way to the third party vendors that people were buying from. So we got access to this like really this engagement level data, which from our point of view, the best data is when people use it, the best way to know if an email works, send an email, know the phone number works, make the phone call. And so we get a lot of that usage based data. But yeah, it's been a very organic
process and it was really built out of we had a need, we tried to solve our need and then a bunch of people organically came to us and said they had that same need and wanted to join in and then we formalized the full structure.
John Kosturos (04:14)
Very cool. So let's dive into one of the pieces there. You know, it's sound we like to ask about your ideal customer profile, but I think you guys are really zeroed in on specific types of companies. So can you tell us a little bit about like what companies you're really focused on serving?
Nick Weldon (04:35)
Sure, so we work with primarily product leaders.
in middle kind of a called large startup to middle market enterprise cohort where data is a critical component to their product or service. We focus in the sales, MarTech, AdTech space, just a lot of where our networks come from, where our data currently aligns well. We are expanding into the HR space a little bit, some fraud and risk and identity space. But predominantly our data is
is like we use analogies lot at our company. These are sophisticated data users already. A lot of times they already have an existing data pipeline. We're either going in there and helping them reduce their costs, because our co-op model is a more efficient cost effective way to get some similar data they have. Or our data is being used to build something new, because we have one of the widest diverse sets of data in this.
John Kosturos (05:08)
Yeah.
Nick Weldon (05:24)
What we found in a lot of these companies right now is there's more demand from users, from product leaders to build more products. So being just an email sending platform is not enough anymore. Or just a contact management platform or a dialing platform. These platforms need to have more features. Either that be if it's a calling platform, it needs to figure out how to integrate email or intents or...
John Kosturos (05:48)
Yeah.
Nick Weldon (05:48)
digital advertising, wherever it may be. So we're finding a lot of these leaders are having more demands from their users to expand their product offering. But yes, it's tech platforms in that kind of middle market space with a focus on sales, MarTech, ad tech.
John Kosturos (06:03)
Okay, let's dig in on the tech platforms, basically. Businesses building platforms around the data set, maybe working it into what they have or using it as an all-encompassing data set, powering their end users or their services. Now I think it's a good shift into how...
What are your ideal returns from those platforms who are consuming the data that ultimately makes it back into your graph and maintains the updates of the data?
Nick Weldon (06:33)
Yeah, this might be a good way to also explain the co-op model. So the concept of the five by five data co-op is every member provides equal value for what they're taking out. And that value comes in really the form of two categories. One's dollars. So some companies prefer to pay more dollars and we can use those dollars and reinvest them to build.
high quality things for the co-op. The other is data. And a lot do a mix. Some dollars, some data there. Most of the companies that are contributing data in don't sell their data. So someone used the analogy once with us. It's like I can put the recycling in front of my house to offset the cost of my mortgage. Great. I don't even know how to sell recycling. So a lot of these companies, this data that they're contributing in is just stored in some server somewhere, effectively creating some sort of cost function for them. They don't really have a go-to-market motion on that.
The majority of the data we do like though, so when companies do decide to get data, we don't want anything overwhelmingly proprietary. We want people to keep their own unique models or really niche data sets, because this does get integrated into our graph and we've got a lot of people from every kind of independent segment in the Marr Tech, Sales Tech, AttTech segment. What we like is usage-based data.
In our opinion, that data's earned. It's hard to get. You have to build a platform that attracts users, and users have to do actions on there. So it's an earned data set. So we like email sending at volume. We like ads being displayed. We like publishers and web page interactions, phone dials, something where the data's being used, and there's some sort of feedback loopback of what happened. Did the email go through? Did it get interacted with? Did somebody open up an app?
pinged to an IP address and stuff. So we really like usage based data. And that data is really, for the most part, measured by the ton. You need volumes to really make a difference. So we email sending platforms that send us billions of emails sends a month. And we get all sorts of interesting insights. have publisher groups that are sending hundreds of millions of tag fires a month, dialing platforms doing tens of millions of dials a month. there's the volume definitely
John Kosturos (08:15)
Yeah.
Nick Weldon (08:32)
It depends on what segment, is relevant to us. But it's definitely that usage base, I would say, more so than scraped-based data. And the data tends to be more stale from our point of view than refreshed.
John Kosturos (08:42)
Yeah.
So, so you're saying basically the graph is being, there's, there's assistance from the actual co-op members to maintain the integrity of the data in the data set. And that's an ongoing thing. So. Yeah.
Nick Weldon (08:57)
Yeah, it's this recycling loop. we get data in from co-op members. We
merge it into our graph and we build this large scale identity graph that has information on kind of like core identity, business identity, device identity, behavioral identity, all linked to a singular node for an individual. So we combine that all together and then we give data sets from that graph back to the members. We use the data and then send us feedback. So we'll send sometimes, we'll think that we're right. They use it. They tell us we're wrong. It updates in the next
month and we're constantly refreshing that on a monthly basis so it's this kind of self-healing flywheel effect where as we bring on more co-op members more people validate the data as more data gets more validated it gets more valuable so there is this flywheel effect of more members creates more value and keeps it more
John Kosturos (09:33)
Yeah.
Yeah.
And I think
it actually leads me into the next two questions. Like my next logical question is, well, how often is the data published back out to the members? Like you guys are, I'd love to be a fly on the wall looking at your AWS bill to process all that stuff. But like once you're processed, how often is it redistributed? But that obviously leads us into another question. How many data sets and what kind of data sets you have? which way, which, which angle do you want to attack first?
Nick Weldon (09:52)
Huh.
Yeah.
Maybe I'll combine the two because I think there's an organic conversation on it. So what we do, we get all this data, we combine it into one singular identity graph. My vision is to have a singular node and vision slash executing on it, but is to have a singular node for an individual or an ID for an individual and everything we can possibly associate to that person. Current, present, business, consumer, device, et cetera, all back to that individual. Then from there, once I have John, I might have thousands of ad
attributes, current and past about you connected, then we create data sets off of that, which are basically just slices or queries off of that node. So an example, and I'll go into the tactical how many we have, but an example would be if I have thousands of attributes about you current and previous, I may make a consumer data set that just says, give me all the current information on John where it's name, address.
you know, does they own a home or not, and email as an example. And then that's a data set. So the thing that helps five by five create data sets from our experience much quicker than our competitors is everything connects them to the graph. The data set is just a query off the graph. Data sets can be created in days sometimes.
because it's just a new way of querying the graph and then maybe creating some filters of what's a minimum acceptable record and there's a few other things called taste test stuff that goes into creating a dataset. But the two biggest things that I'd say gives 5x5s is advantage. And I'll get tactically into datasets. One is we have arguably the strongest data acquisition mousetrap out there.
Every time we bring on a member, the most part, get more data. As we get more data, it's easier to bring on members. So we've got this kind of like always expanding data acquisition model. The second is instead of getting that data and just building siloed data sets, and you have to keep you to do massive investments, sometimes millions of dollars to build a new data set, we put it all into a graph, think data set, and then data sets are as simple as a query off that graph. So I'd say that the acquisition model and the graph,
are what's allowed 5x5 to scale incredibly efficiently, especially compared to our competitors. And it's allowed us to come out with data sets from what I've seen overwhelmingly faster than any of our commercially competitive.
peers in the space. So tactically, that's how the model works. Tactically, I 5x5 has around 15 generally available data sets with maybe a handful in a beta fashion that we're still kind of testing and bringing to market throughout the rest of this quarter and year. They're very diverse. There's a B2B contact data set that's pretty competitive to like the, you know,
People data labs of the world type type in cognizance of the world stuff like that We have a consumer data set that's competitive to like the experience and credit bureaus the world we've consumer attributes data set which is the kind of core consumer identity data set with hundreds of attributes or attached to it very much like a credit
header type data set. We have firmographic data sets that are complementary to the B2B. We have device graphs, which are very hard from our experience to accumulate, where you've got mobile ad IDs, which is phone device data, IPs and cookies all connected to an individual ID. So you can connect it to the individual and then you can connect that to businesses. So the nice thing with the graph is you can traverse through most device data is consumer driven. Connecting that
to a business is really hard. So like your mobile phone is probably not registered to SpringDB. But if I know it's you and you're connected to SpringDB, I can start to say, hey, if I want to target a C-suite person at SpringDB, that phone is going to be very valuable for targeting. Things we've learned too is
John Kosturos (13:41)
Yeah.
Nick Weldon (13:42)
Targeting B2B is expensive. So if I can find your consumer just because of the economics, you're selling usually set five, six, seven figure items in B2B. If I can connect a consumer device to a B2B persona, the ad space on that consumer device is exponentially cheaper.
So there's a lot of benefits you get to this graph to all of our graph, all these data sets talk to each other through IDs. So you can kind of like Lego at them on. We have some real more real time daily streaming stuff. We've got a tag that can be used for ad act attribution or website resolution. We have behavioral feeds. So
We have a massive publisher network where we have, I think at this point, hundreds of thousands of publishers. Now in this, it's kind of like a mini co-op inside of the co-op. One of the things these publishers want is when a publisher can give more attribution on their traffic, they can sell the traffic for a premium. So they can say these are the types of people on their website. They can sell that ad space for a premium. So we give resolution.
away to publishers so they can get that higher adder attribution and in return we're able to get weblog data so we can start to create behavioral, like a behavioral graph. So we've got a lot of these different segments. We keep expanding, two things we expand the graph on. One, we get more data in that just comes in and we're like, we could build, think of it, use the analogy like we look in our pantry and there's more things to cook out of there. The other is it's demand from the co-op.
So the co-op sometimes has demanded or requested additional data sets to further their products. And then we go back to our massive co-op network and we might subsidize a certain data feed or someone's co-op membership for a certain data feed to bring it in. So yeah, we've got 15 or so data sets. They're refreshed on a monthly to real time basis. A lot are monthly, kind of the identity ones are monthly. The behavioral stuff's daily and then we have some
some true real time tag resolution stuff for ad tech. And yeah, and it all kind of connects.
John Kosturos (15:31)
brought that all
together on the updates and everything. So thank you. think you took us through that kind of entire process all the way even down to the updates. You know, I think where I want to go from there is, and I know this is a broad...
kind of universe, but because the data sets, there's some, many building blocks and those building blocks can be used for go to marketing, even for identity resolution, fraud detection and you know, stuff like that. But what are the most common platform use cases that you guys just like really.
you know, do very well at, you're powering those ecosystems in a big way now.
Nick Weldon (16:12)
Yeah, so I'll talk platform and then I'll talk maturity of those platforms because I think the two talk really well together for where we thrive. So I would say the sales MarTech ecosystems where we do probably our most of our works. These are CRM platforms, email sending platforms, dialing platforms. We do a decent amount with like DSPs and SSPs in the ad network.
A lot of these platforms, it's lot of enrichments. So CRM platforms, and you can think of some of the big CRM platforms, we probably work with most of the ones you can think of, the household names. Their users...
bring in data into those platforms and the better enriched, more recent, accurate that data is, the more value they get out of those platforms. Email sending platforms, they'll use our data to enrich, to create insights. So they might say, hey, your open rate is 42%, but in this cohort, it's 62%. You might want to spend more time in this cohort because they've now enriched that data. In the ad tech world, the whole big thing they're trying to get to in the ad tech world is users will upload data
I
think CRM data so there might be emails and stuff like that and those ad tech platforms are trying to connect those to devices the more devices they can connect to the more Ads that can be serviced the ad tech platforms make more revenue and the customers are more excited because they can fulfill more their their campaign targeting So it's really it's a lot of enrichment so that is something they're already getting There's some provisioning some of these platforms want to provision new data to users in there, but I would say enrichment is really worth
John Kosturos (17:32)
Yeah.
Nick Weldon (17:38)
the best bang for the buck is almost every one of the platforms where we do very well with already has an existing data pipeline. They know how to process large scale data sets. They're either wanting to use us to offset an existing pipeline where we might be a better cost provider. They might have an API waterfall where they want a flat rate of cost because the variable component might be hurting their gross margin.
that all of these companies know how to process large scale data sets. So they know how to take hundreds of millions of records so that they can process it and be pick the value out of it. So the analogy we use is like, we give our data to like oil refineries. Our data is not straight to the gas station. Those oil refineries refine it and make it ready for your car. But they want, you know, we're selling some of that more crude oil for large scale providers where we've seen our model
break, and I think John, this is where we do talk on SpringDB, I think this is where SpringDB comes in and very handy, is when we're the only provider in their data waterfall, and or they don't have the systems and processes already built to refine our large scale data set to make it ready for their end user, kind of gas station ready there. And that takes, as you know, there's processes, experience.
So that's where we've seen where we've created friction with individuals as they took our crude oil and put it right into a car and blow up the engine. ⁓ You and I both experienced that.
John Kosturos (18:59)
Yeah.
they're not alone in that world because I just a long story from my past in my, a past life, I actually purchased 15 different providers, data sets, and we aggregated them and we were spending millions of dollars on the data, millions of dollars curating the data. And at the end of the day, as the leader of that organization, I was communicating to it.
cloud and data engineering team that I didn't have the skills to basically understand everything that they were going through. And we would find that data sets were slow to onboard.
only certain attributes showed up and it's always like a back and forth. So it is important to, you know, make sure that when you're consuming big data sets, you do have experience around. and luckily for you, Nick, think some of your customers are the top data experts in the world, but can you give us some idea around the scale of data across these different data sets? Like how big are they?
Nick Weldon (20:00)
Yeah, so it's a little bit of like an iceberg. So what we send out is really what we've refined, even we've refined down and then Don, I work with you and you guys refine it even more. Our B2B contact data set is...
about just north of 100 million US contacts, they get 700 million plus internationally on the B2B contact side. Our consumers about 250 million US individuals. companies, we're talking maybe 10 to 15 million companies, devices, we're probably close to a billion devices at this point across.
You know, think phones, computers, CTV, wearables. On the websites, we're in the tens of millions of unique web pages now and hundreds of thousands of top level domains. So it's a kind of individual publishers. I'm kind of the raw data.
petabyte a week range, maybe more. it's a lot, because we get a lot of streaming in data, then we dedupe it and we might see the same website ping 10,000 times in a day. And then we have to dedupe that down.
We find too, like, when you're trying to understand is this IP connected to you, John, we want to see hundreds of occurrences on that IP address. So is just you logging in at a Starbucks, or is it something, you know, an office or a home where you're constantly pinging? So we do want repetitive signal to help make sure is this a fluke or not. But yeah, we're probably in the petabyte plus range a week. It's pretty heavy at this point. ⁓
John Kosturos (21:23)
Yeah,
the way I would...
Nick Weldon (21:24)
Everybody
knows petabytes is it's like it's a different level of machinery. So it's not like you can just move it with like a small tractor. Now we're talking big industrial scale tractors.
John Kosturos (21:32)
Yeah, and with experience,
what we've seen is, hey, it could be 15, 20 million companies, 10 to 20 million companies that's updated every month. And it could be 350 million people updated every month. And it could be a billion devices. And it could be, you know,
35 billion website page views every single day, right? And the cycles of those are different. But when you get them all working together, it's some very powerful.
Nick Weldon (21:58)
Yeah, the value of our data set is
if you just need a B2B data set, I'll use that as the example. There might be five or six called comparable-like solutions to us out there. So we might have five or six competitors. Then you need a B2B data set connected to devices. There might be two competitors. And then you need a B2B data set connected to devices, connected to behavioral signals. Not like one-on-one. So our value really is we're probably top-rate quadrant on one data set.
John Kosturos (22:15)
Yeah.
Yeah.
Yeah.
Nick Weldon (22:26)
Once we start adding multiple together, really isn't, there's just not a market comparable. So you have to go and get five, six, seven companies, try to get all their data sets to talk to each other. The commercials don't really map and it gets really expensive and all that. So five by five really does really well. Either we're a cheaper alternative to something in your existing data supply chain or your product roadmap demands innovation. And you need more data sets to keep up with customer demand. And that's where our data talks really well to each other and prices really well to each other.
John Kosturos (22:47)
Yeah.
Yeah, the one thing I really like is, so if you're building a product, you know, when you get all the five by five data working together, you can build audiences, you can do emails, you can do telephone, you can send direct mail, you can build integrations into meta and...
LinkedIn and Google retargeting and every ad network you need and push audiences for digital advertising, you can resolve people that are on your website or viewing your ads. Like it just gives you a great foundation. then like what we've helped companies with is okay, the foundation set and they're still data junkies. And so are their customers. like, I want healthcare and I want education.
Nick Weldon (23:26)
Yes.
John Kosturos (23:36)
So we like stitch additional data sets because you know it's not always just about being cheaper it's like how do you create a mode in this new world of competitive solutions popping up and AI and it's like create very functional foundational data sets and then verticalize and and and that's where we've been really successful with you know five by five
partners and stuff like that because you you guys really do offer a great foundation.
Nick Weldon (24:04)
I look at
that's exactly why I see the partnership on it is we give them the foundation to start with kind of that core identity, which is where you start. And then that's probably good for half the market right there. That's enough.
John Kosturos (24:15)
Yeah, maybe more,
yeah.
Nick Weldon (24:17)
Yeah, maybe
even more. And then there's the other half to third that need very targeted add-ons. And you guys know how to talk really well with our data. So you know where the gaps are. It might be a small business segment that they need more focused on, or health care. There's different attributes they need. We have some health care data, but definitely not the depth that you guys can. And you know to blend it together.
really well to our data set. yeah, that's absolutely. And then the other thing I'd say just through the partnerships you've created, John, is our data might be 80 % refined to what people need. And you have created partnerships and solutions where instead of putting the pressure on a company to figure that out, which some have already, these bigger companies have, there's a large cohort that.
John Kosturos (24:55)
Yeah.
Nick Weldon (24:57)
We all wouldn't recommend they build that the systems and there's companies out there that you've partnered with that can refine it. So you're either refining the five by five data, the extra 20, 30 % the end person needs to go into their product or service and or you guys are bolting onto it. So just adds more complimentary value and they can service their customer needs. And we can't at five by five go do that because it doesn't benefit enough of the overall average of the co-ops. I have to balance investment over overall benefit.
John Kosturos (25:24)
Yeah.
Nick Weldon (25:28)
We can't get very niche because we have so many use cases
out there, but our vision with 5x5 with the core and with our partners is for no product leaders roadmap to be limited by their access to data. That's access, both like literally how do we give it to them. That's licensing terms. So do they have the rights to do it? A lot of data providers might limit what you can do with the data sets. And then that's the types of sets. So as you just mentioned, like there's a lot you can build off 5x5. And I don't think there's many use cases.
John Kosturos (25:40)
Yeah.
Yeah.
Nick Weldon (25:58)
from a broad standpoint, we don't have data for. And then if we don't have it, we go to you, John, and you come and fill in the gaps. So that's really the vision we'd had. That's what wanted. We wanted data. And then we want to make it in a way that it's affordable. So you don't have to go overwhelmingly gray. So those were kind of our tenets. And that's where the two of us have just meshed together.
John Kosturos (26:01)
Yeah.
Very
cool. So Nick, where are you guys going next? Well, before you answer that question, let's go to domestic or global coverage. Where are you guys really focused in terms of bringing in data? Are you focused on US? Are you in any global environments?
Nick Weldon (26:36)
So yeah, up until about November, at least publicly, were just US based.
In November, we were able to release in a beta fashion that we've been expanding aggressively on the international side. It's a cohort that we've seen be wildly under invested in from the data ecosystem. think a lot of it happened during the kind of GDPR side where everybody just took a step back, us included, not knowing the risks and where it was trending and what the regulatory frameworks would be there. I think a lot of us are feeling more comfortable for specific use cases with where that's.
what's happened and where it's trending from a regulatory standpoint. We're seeing that it seems like the EU is being a little more.
called partnership minded, especially in a business standpoint, for how to use that data for business purposes. think it's very restrictive. So quick answer is we have a B2B contact data set that is like our US data set that's globally, from our point of view, of all the data sets I have right now, I think that's probably our most...
Probably one of our best data sets right now, just because there's not a lot of competition and alternatives there right now. So if you're looking for international B2B contact data, there's just not a lot of places to go look, especially in comparison to the US. from what we're seeing, I think we one of the best specifically there. And we were able to use our co-op model and build, go from like, think, last to, if not first, right near first in that cohort. So that's one. From there, we're taking our US playbook.
and just internationalizing it where it makes sense and where we can compliantly do it. So we have device graphs that are coming out like this month or in May, this month. So it's going to be maids, cookies, IP addresses, IP to hem, IP to company. So you can do international device targeting there, which is I think a lot of people are looking to do that.
Those counts will be publishing not as big as our US counts, but again, it's like compared to what's commercially available. That's a big statement we use a lot on is how is the data compared to what's commercially available? We all want perfect data, but as we know at scale, it's hard. I think our device data will be right up there with the top commercially available solutions and it'll talk well to our B2B contact data, which will talk well to our US data. Everything talks well. So it's like, it'll be a lot of co-op members right now.
And it's our fastest growing segment, I would say, in five by five right now is international from a percent of growth standpoint, who are upgrading, who've been waiting into international data and testing it. So same thing, they're going through testing processes and the testing has been coming out quite positive.
So that's probably the hottest growing section right now, where I see us like over the next six to 24 months is to start expanding more into the fraud prevention and identity space. It's kind of the same type of data that's used in ad tech. Like for the fraud prevention space, that's for like e-commerce platforms. They want to understand, hey, is this device seem right, address, et cetera. Same kind of thing used for marketing, just higher standards.
John Kosturos (29:15)
Yep.
Yeah.
Nick Weldon (29:21)
well in fraud prevention, you kill it in advertising and marketing because it's a higher quality standard. So that's probably one of the next big areas that we were in, but we're going to be investing more into that segment.
John Kosturos (29:25)
That's cool.
Okay. So I
do have two more topics and then we'll get into where you headed next. And I know we're getting kind of late into the game, but you guys have so much to cover. One of the things that you mentioned earlier was terms. And you know, if you're a platform or you're an agency and you need to consume data at mass and you need to forward license it to an end user, it's not always a straightforward game when you're going out to data vendors. Like you really.
You understand how to negotiate terms with vendors to be able to use their data in your product or service.
Nick, tell me how you guys, because you're kind of unique. I mean, there's a handful of companies in this space that just sell the platforms and you're like, not after the end user consumer, right? So tell me how you guys make it seamless for a platform to just buy data, onboard it and start providing it to their customers.
Nick Weldon (30:28)
Yeah, so there's two buckets I would say. One is the licensing terms. One is the actual access to the data itself. So on the access side, for the majority of our data sets, we're giving you a full file in your environment so you can query it once, 10,000 times, 100,000 times, however you want to query it. So that gives you the flexibility from a product leader. The other side of this too is on the licensing terms.
there's key pillars that are important to when you're buying data. What are your rights? Are there derivative rights? How do you forward this on your end users? Do you have to delete the data if you leave? And so we have standard licensing terms. We also have flexible ones where people can pay more for certain rights that they want. So it's just like when you're looking at the data side that
Where's the liability stand? Like there's just a bunch of these, these different things. We're pretty good to, we don't get a lot of red lines on a lot of our stuff. And then we do have modular ways where somebody needs some extra rights. We have the ability to do that. There's just a cost component to that, whether that be because we take on more risk or they want more value out of that data set. And then we can use those funds to go either invest more in compliance internationally or build stuff with it.
John Kosturos (31:38)
Yep.
Nick Weldon (31:41)
So I put those are the two buckets if you're looking at it and they're different for your sit your where your phases in the business so your earlier stage Access might be a different demand than when you're late stage The thing too is when we have a lot of data sets which a lot of these companies were finding need more types of data Adding more data sets on is just a skew ⁓
John Kosturos (31:51)
Yeah.
Yeah.
Nick Weldon (31:59)
So you don't have to manage seven data sets with seven different licensing models with seven different access models. And that's where I think the complexity gets really tricky where you can use this in certain use case and that not a use case and all that. That's why I also think having one kind of MSA stack, one access stack simplifies that down quite a bit.
John Kosturos (32:18)
Yeah, I think just experience working with you and then other providers is number one, that's what you guys do. So your paperwork is structured toward enabling these types of platform use. Whereas like others, like got to tell them that it's an end user agreement. Like you got to get to the right department. It's a whole game.
Nick Weldon (32:30)
It was all built around that, yes.
Well, also have a lot of ones. Well, big thing too is a lot of the companies
out there, they look at it as if you're using more data, they could have charged you more for the data. lost a lot of you.
John Kosturos (32:48)
Yeah.
Nick Weldon (32:49)
In our case, we get the usage back. We want you to send 10 billion emails, because then we get 10 billion email signals back. So we have an incentive system that's really built around making it easy for you to get the data, use it as much as possible. And we actually capture the upside value, not through charging you more or restricting you, but through the contribution. So that's, say, one of the big differences why our model didn't have to change, where a lot of companies might have had to.
John Kosturos (32:56)
Yeah.
Yeah.
Nick Weldon (33:14)
or they have this kind of usage-based model, because if you grow a bunch, they want to charge you a lot more. Versus us, we'll either give you more data sets, we keep coming up with more, or we get the usage back. So that's helped us stay very true to that model.
John Kosturos (33:16)
Yeah.
Yeah.
Yes.
All right, last question before I get into the future. Intent is a big kind of keyword in our space. And I've gotten used to with the traditional providers seeing a business type intent where a business is surging through multiple individuals on a topic and...
Now I'm going in and building buying groups and just doing AVM advertising and just, we call it whitewashing that company and just trying to get in front of everybody who's relevant. Tell me how you guys are kind of enabling this idea of categorized intent that breaks down to a potentially individual.
Nick Weldon (34:12)
Yeah, and I would actually say we've got 90 % of the ingredients, or 100 % of the ingredients, but we take it 90 % of the way to intent. So we're not trying to say this person is in market for this solution. What we are is we're an observation network. Almost everything we see is deterministic there. So a couple things. One, we look at almost everything in our graph person out. So the devices, their attributes, even their behavior. And then if you want to get it to a company, you just roll it up.
John Kosturos (34:21)
Yes.
Yep.
Nick Weldon (34:36)
into a company. our behavioral graph is person out, which lets you get person level intent. But what we're saying is we're not saying this person is in market for this solution. What we're saying is we've observed a tremendous amount of web pages that are connected to this topic and connected to this person. So we're saying this person has observed a tremendous amount of content related to this topic.
it's the co-op member gets to decide what is intent. So is it, I'll use buying a home on the consumer side. Buying a home on the consumer side could be, looked at homes, mortgage rates, moving companies, neighborhood safety.
That might be true intense. If it's on the company side, might be multiple people at that company over a certain seniority level have observed it over a certain amount of time. So we're getting like what five by five, but we try to do everything. It's a massive observation network. saying we've observed even when we say an email is good. We're not saying this email is good or bad. We're saying we observed an open or click on this date.
John Kosturos (35:34)
Yeah.
Nick Weldon (35:35)
You can decide is that a high enough level activity or recent enough for what good, because what might be good to you might be bad to somebody else. It depends on what they're, they have a Flamin' Yone kind of steakhouse or is it more their wide scale and they want to go more maybe McDonald's. Both can make a lot of money. So there's different versions of what good is. So first is it's person-based, which is very unique and hard to do. Second, we have a massive network of publishers. So we're seeing a tremendous amount of behavior and content.
Third
is we have a very good, this is one of the probabilistic but high confidence probabilistic things, it's a categorization engine there. So we say, hey, this page was about this piece of content. And then we just link it together. So if you're building and you get to build your own intent model off it, but we're taking a 90 % of the way there, the hardest part of that, the two hard parts of pulling off a person-based behavioral product, one is you need a massive network of publishers.
John Kosturos (36:14)
Mm-hmm.
Nick Weldon (36:29)
And that's an earned network there. So that's one. The second is you have to be able to connect that massive amount of web behavior to identity. And then that identity to other useful attributes. Be it consumer, B2B attributes. So...
John Kosturos (36:38)
Mm-hmm.
Nick Weldon (36:45)
For like what we're seeing is it gives individuals the flex or companies the flexibility to build their own little secret sauce connected to it versus a black box, which is like some of our counterparts. A, they're not person based, which is limiting. And B, it's a black box. what they may say this company is in market for intent. You don't know what that means. Was it a hundred observations? One, C-suite, number of employees. It's a secret sauce.
John Kosturos (36:52)
Very cool.
Yeah.
Nick Weldon (37:11)
So it's hard for you as a product leader to build off that because you don't know what went into it. Worse, this person observed a lot of content related to this. You can roll it up and decide what you think a good score is.
John Kosturos (37:15)
Yeah, if you're a product leader, can almost just...
Yeah. Like if you're
a product leader, you could just build that into your UI and like let the customer create their own like, yeah, that's pretty cool.
Nick Weldon (37:27)
It could be a feedback loop based off campaign performance. Hey, it's like,
we've seen people build regression models off campaign performance that it might be multiple topics unrelated, like the, you know, that you would have expected. And that's where the true intent comes from. It's people looking for homes with blue jeans. Great. I don't know. So that, lets you, you don't have to, you could build a training model that trains off of campaign performance to then create the intent. And that's probably where the best people are.
John Kosturos (37:38)
Yeah.
Yeah.
Yeah.
Nick Weldon (37:55)
now in the car.
John Kosturos (37:56)
One of the things we've actually done for that is, I mean, one holding a week's worth of intent files is one thing, but holding six months or a year is a whole other ball game.
Nick Weldon (38:06)
This was terabytes of data today.
⁓
John Kosturos (38:08)
Yeah. And then
like processing against it. So like we built look back periods that allow them to process quickly and no cost per query type of access. And yeah, I mean, I like the way you explain that though, is that you again, providing the building blocks for a platform provider. Yeah.
Nick Weldon (38:25)
It's a massive observational network. We're observing
this behavior and then we use our graph to link the behavior back to identity. And then from there you can traverse. Where like where you come in, John, in a really positive way from my experience is one, just like you can make the data exponentially more usable. So way more companies can access to it because it's a heavy data set. So I'd say the amount of companies that can process that amount of data daily is very few. We work with those very few. So you make that data accessible to a much wider range of
companies that normally wouldn't get access to it. And the second is some of those companies might need additional data sets layered onto it. So because it's just behavior plus.
John Kosturos (39:00)
Yeah. Or they don't want to pay AWS 50. Yeah,
they don't want to pay AWS 50 grand a month to process all of it. Yeah.
Nick Weldon (39:06)
Exactly. Well, we
use this analogy all the time. It's like, if you look at behavior, a lot of people are in market for Ferraris, but who can actually afford them? And that's where there might be supplemental data sets to qualify the intent beyond just the behavior. So that's the other side of that keto intent is looking at the behavior and then qualifying. Does that individual or company actually have the means to acquire what they're looking at? And that's...
John Kosturos (39:17)
No.
Yeah, one of the cool things
that we've done is...
you get, you you circumvent the different intent categories and then you create an audience and then we merge it with a social affinity audience. So we can see, okay, what are these people have affinity for? And then we took it a step further and said, who influences those people? So you could do like celebrity endorsements down to these people that were showing research on a topic, but you have to stitch like a social graph
to the other signal graph, to the identity graph, but when you do those things, it's super powerful, like transformational.
Nick Weldon (40:06)
And that's where almost every company we're seeing, whether it be with our data or other data sets, they're layering multiple data sets. If anyone could just buy a data set off the shelf and build a seven, eight figure company, there's no competitive edge.
John Kosturos (40:16)
Yeah.
Nick Weldon (40:19)
there. So all the ones we're seeing, they're building stuff off of our data and or combining multiple data sets there. So like we think there's a lot of value. If I have research behavior that says you're into golfing and then there's survey, a completely different location based or survey based stuff that says you're into golfing, that's completely independent. The odds are very low that you're not into golfing because it's just too, so like starting to do like almost like a Venn diagram of overlying data.
John Kosturos (40:38)
Yeah.
Yeah.
Nick Weldon (40:46)
Almost any company we're seeing that's not doing something on top of regardless of the use case or really even size, those are the ones we see struggle. The ones that we see skyrocket quickly have a strategy to take our data and or other data sets and build some sort of recipe on top of them. To give feedback loop, there could be multiple ways to pull that off.
John Kosturos (40:55)
Yeah.
Yeah. Even if it's their own first party information, right? Like censored or whatever it is. Yeah.
Nick Weldon (41:11)
But if you just take
it and then just build a little UI in front of it, that should be a lifestyle business ⁓ there. Which is okay, but you're not gonna probably build a meaningful seven, eight figure business without building a model on top of it. And that's where you guys come in. You make that way more accessible to early stage companies. And then for companies that need niche stuff, they can come to you and say, can't service. We refer them to you all the time.
John Kosturos (41:16)
Yeah.
Yeah.
Yep.
So we only have a few more minutes left, but where is 5x5 headed?
Nick Weldon (41:38)
Yeah, I put two buckets that we're focused on, or three buckets, okay. One is just bringing out more co-op members to enhance and improve our current graph and data sets. So we're always looking for more of what we have to improve. So that's one. Okay, the second is a TAM expansion. So taking our graph and bringing it, it's still platform-based use cases, but bringing it into fraud tech and the kind of identity tech space.
It's similar data sets, but they have their own nuance, things they need for what they consider quality, usability, et cetera. So I'd say the fraud space is probably gonna be one of our largest expansions over the next 12 to 36 months. It's also a harder space to be in. So I think it'll just overall improve everything we're working on. The third is, so five by five, we have a very horizontal data strategy. We keep adding on more data sets off our ground.
So we probably, I think definitively, we have the widest data offering in our kind of wholesale.
data access category. Most of our competitors have very siloed. They build up one, maybe two or three data sets. We have 15. So we have a horizontal data strategy. One of the things we're working on right now, and we've just acquired our first company and we're going to be acquiring more, is acquiring companies in each of those maybe data categories that are specialists. So these might be smaller companies that did a really good job at one data set, but their revenue has maybe stalled just because they won, but have gotten really good at that.
quality or sourcing or something very niche or a model or something that like would take us a lot of time and effort to build. And because we have so many data sets, we can't invest maybe that time or capital into that. So we have private equity partners. We can leverage different vehicles to acquire these. So.
We've acquired one. I wouldn't be surprised over the next 12 to 36 months we acquire a handful more companies that specialize in maybe device data or consumer data or maybe some validation vehicle or international. So we're actively looking. So there's companies that fit kind of that.
Maybe like three to 10 million ARR, they're more bootstrapped, a little bit of profitability, and they have one of those specialties. Like we're an active looker. So yeah, I'd say improve what we have, a new target market segment or an expanded target market segment. We're still in there, but we're expanding in there. And then acquisitions, those are probably the three big buckets that should keep us incredibly busy.
John Kosturos (43:48)
Very cool.
Absolutely. And that's also a question if you're buying data, you should ask your data providers, like what is your strategy to continue to grow your data or to continue and invest in your data. it's exciting to hear that you guys are making that a top priority and we will be keeping an eye on you. And hopefully next time we get together is in Austin or in Miami in person. But Nick, it's been incredible having you on today and we look
Nick Weldon (44:22)
That's right.
John Kosturos (44:26)
forward to kind of continuing to watch you guys's journey. Absolutely.
Nick Weldon (44:30)
Thanks, John, and we appreciate the partnership. It's been a fun journey
together.
All right, welcome everybody to another episode for the SpringDB Data Exchange. We've got a very exciting guest today, Nick Weldon from 5x5. How are you, Nick? Doing well. Yeah, so Nick and I have gone a couple of years back and...
Nick Weldon (00:13)
Good job, how are you?
John Kosturos (00:20)
They're doing some pretty amazing things over at 5x5 and maybe we can start out Nick just by diving into a little bit of your background.
Nick Weldon (00:28)
Sure.
And five by five story is a very kind of organic bootstrapped story, is just similar to my background. So I have a background in bootstrapping half a dozen or so startups, like anything, some worked, some didn't. One of the themes with all of the, two of the themes that all the startups that I've worked on is as much as I could, I tried to stay away from formal capital, specifically venture where I could. And all of my startups had some sort of data backbone where data was a critical
elements to the service or the product.
And so rewind like six or seven years ago, we were building a solution where we were trying to create a shared experience graph, where we were trying to find all the shared commonalities between people. And then that could be used in like a sales environment. So when you get on the phone with somebody, you know, you have kids in college or you're both into snowboarding or your fans are the same sport or fraternity or whatever it may be. So in that pursuit, that is a very data heavy product.
So
I got linked up with Brian, who's how we got connected, Brian Perks. And in that, we started figuring out how to get some of our own data, first party data and some scraping and third party data. However, for the graph to work efficiently and effectively, we needed more data.
and in order to get that data, we were going to have to raise a funding round, to be able to just get the capital to secure the data licenses for such large scale data, especially back then it was even more expensive. I would say there was less just commercially available. Large scale options. so in that Brian, I had this idea of we knew a few companies that weren't data companies, but they were platforms where their platforms generated some data, but they also needed more data for their platforms to work. And the thesis we had is what if we could get
John Kosturos (01:51)
Yeah.
Nick Weldon (02:05)
half a dozen or a dozen of us together where we could all share some of our first party data into this collaboration, combine all the data together, make a big data set that we could all use to improve our products and services without enduring funding rounds or pushing funding rounds wherever it may be. So it was kind of like a side project to some extent for the main project.
Well, within a few months, what was supposed to be a side project was five hours a week and then 10 and then 20 and then 30. And then we started spending more of our time on the side project than the main one because it just had so much energy and excitement and people were coming to us. It was word of mouth spreading. And as you know, John, like with any startup, like product market fit, you know, and it kind of is happening. We were feeling the early signals of it.
John Kosturos (02:47)
yeah.
Nick Weldon (02:49)
And there were a bunch of companies that basically are out there where they were generating maybe small to medium amounts of first party data, but had an exponential need for additional data.
and wanted to collaborate on the concept. So that's where 5x5 was created. It's a data cooperative for tech platforms where third party data is a critical component to their product or service. Where these platforms on their own don't generate enough data for their solution to work independently. And they need that supplemental data.
to accelerate either the usability of it or to build new features and solutions. In addition to that, there's a lot of data out there that can't be sold based off of user permissions. So a lot of these platforms can't physically sell some of the interaction data they might be having on their platform, but they can barter and use it to...
John Kosturos (03:25)
Yeah.
Nick Weldon (03:39)
in an exchange format. So there was a whole bunch of data, valuable data, that was kind of in this platform ecosystem that never made its way to the third party vendors that people were buying from. So we got access to this like really this engagement level data, which from our point of view, the best data is when people use it, the best way to know if an email works, send an email, know the phone number works, make the phone call. And so we get a lot of that usage based data. But yeah, it's been a very organic
process and it was really built out of we had a need, we tried to solve our need and then a bunch of people organically came to us and said they had that same need and wanted to join in and then we formalized the full structure.
John Kosturos (04:14)
Very cool. So let's dive into one of the pieces there. You know, it's sound we like to ask about your ideal customer profile, but I think you guys are really zeroed in on specific types of companies. So can you tell us a little bit about like what companies you're really focused on serving?
Nick Weldon (04:35)
Sure, so we work with primarily product leaders.
in middle kind of a called large startup to middle market enterprise cohort where data is a critical component to their product or service. We focus in the sales, MarTech, AdTech space, just a lot of where our networks come from, where our data currently aligns well. We are expanding into the HR space a little bit, some fraud and risk and identity space. But predominantly our data is
is like we use analogies lot at our company. These are sophisticated data users already. A lot of times they already have an existing data pipeline. We're either going in there and helping them reduce their costs, because our co-op model is a more efficient cost effective way to get some similar data they have. Or our data is being used to build something new, because we have one of the widest diverse sets of data in this.
John Kosturos (05:08)
Yeah.
Nick Weldon (05:24)
What we found in a lot of these companies right now is there's more demand from users, from product leaders to build more products. So being just an email sending platform is not enough anymore. Or just a contact management platform or a dialing platform. These platforms need to have more features. Either that be if it's a calling platform, it needs to figure out how to integrate email or intents or...
John Kosturos (05:48)
Yeah.
Nick Weldon (05:48)
digital advertising, wherever it may be. So we're finding a lot of these leaders are having more demands from their users to expand their product offering. But yes, it's tech platforms in that kind of middle market space with a focus on sales, MarTech, ad tech.
John Kosturos (06:03)
Okay, let's dig in on the tech platforms, basically. Businesses building platforms around the data set, maybe working it into what they have or using it as an all-encompassing data set, powering their end users or their services. Now I think it's a good shift into how...
What are your ideal returns from those platforms who are consuming the data that ultimately makes it back into your graph and maintains the updates of the data?
Nick Weldon (06:33)
Yeah, this might be a good way to also explain the co-op model. So the concept of the five by five data co-op is every member provides equal value for what they're taking out. And that value comes in really the form of two categories. One's dollars. So some companies prefer to pay more dollars and we can use those dollars and reinvest them to build.
high quality things for the co-op. The other is data. And a lot do a mix. Some dollars, some data there. Most of the companies that are contributing data in don't sell their data. So someone used the analogy once with us. It's like I can put the recycling in front of my house to offset the cost of my mortgage. Great. I don't even know how to sell recycling. So a lot of these companies, this data that they're contributing in is just stored in some server somewhere, effectively creating some sort of cost function for them. They don't really have a go-to-market motion on that.
The majority of the data we do like though, so when companies do decide to get data, we don't want anything overwhelmingly proprietary. We want people to keep their own unique models or really niche data sets, because this does get integrated into our graph and we've got a lot of people from every kind of independent segment in the Marr Tech, Sales Tech, AttTech segment. What we like is usage-based data.
In our opinion, that data's earned. It's hard to get. You have to build a platform that attracts users, and users have to do actions on there. So it's an earned data set. So we like email sending at volume. We like ads being displayed. We like publishers and web page interactions, phone dials, something where the data's being used, and there's some sort of feedback loopback of what happened. Did the email go through? Did it get interacted with? Did somebody open up an app?
pinged to an IP address and stuff. So we really like usage based data. And that data is really, for the most part, measured by the ton. You need volumes to really make a difference. So we email sending platforms that send us billions of emails sends a month. And we get all sorts of interesting insights. have publisher groups that are sending hundreds of millions of tag fires a month, dialing platforms doing tens of millions of dials a month. there's the volume definitely
John Kosturos (08:15)
Yeah.
Nick Weldon (08:32)
It depends on what segment, is relevant to us. But it's definitely that usage base, I would say, more so than scraped-based data. And the data tends to be more stale from our point of view than refreshed.
John Kosturos (08:42)
Yeah.
So, so you're saying basically the graph is being, there's, there's assistance from the actual co-op members to maintain the integrity of the data in the data set. And that's an ongoing thing. So. Yeah.
Nick Weldon (08:57)
Yeah, it's this recycling loop. we get data in from co-op members. We
merge it into our graph and we build this large scale identity graph that has information on kind of like core identity, business identity, device identity, behavioral identity, all linked to a singular node for an individual. So we combine that all together and then we give data sets from that graph back to the members. We use the data and then send us feedback. So we'll send sometimes, we'll think that we're right. They use it. They tell us we're wrong. It updates in the next
month and we're constantly refreshing that on a monthly basis so it's this kind of self-healing flywheel effect where as we bring on more co-op members more people validate the data as more data gets more validated it gets more valuable so there is this flywheel effect of more members creates more value and keeps it more
John Kosturos (09:33)
Yeah.
Yeah.
And I think
it actually leads me into the next two questions. Like my next logical question is, well, how often is the data published back out to the members? Like you guys are, I'd love to be a fly on the wall looking at your AWS bill to process all that stuff. But like once you're processed, how often is it redistributed? But that obviously leads us into another question. How many data sets and what kind of data sets you have? which way, which, which angle do you want to attack first?
Nick Weldon (09:52)
Huh.
Yeah.
Maybe I'll combine the two because I think there's an organic conversation on it. So what we do, we get all this data, we combine it into one singular identity graph. My vision is to have a singular node and vision slash executing on it, but is to have a singular node for an individual or an ID for an individual and everything we can possibly associate to that person. Current, present, business, consumer, device, et cetera, all back to that individual. Then from there, once I have John, I might have thousands of ad
attributes, current and past about you connected, then we create data sets off of that, which are basically just slices or queries off of that node. So an example, and I'll go into the tactical how many we have, but an example would be if I have thousands of attributes about you current and previous, I may make a consumer data set that just says, give me all the current information on John where it's name, address.
you know, does they own a home or not, and email as an example. And then that's a data set. So the thing that helps five by five create data sets from our experience much quicker than our competitors is everything connects them to the graph. The data set is just a query off the graph. Data sets can be created in days sometimes.
because it's just a new way of querying the graph and then maybe creating some filters of what's a minimum acceptable record and there's a few other things called taste test stuff that goes into creating a dataset. But the two biggest things that I'd say gives 5x5s is advantage. And I'll get tactically into datasets. One is we have arguably the strongest data acquisition mousetrap out there.
Every time we bring on a member, the most part, get more data. As we get more data, it's easier to bring on members. So we've got this kind of like always expanding data acquisition model. The second is instead of getting that data and just building siloed data sets, and you have to keep you to do massive investments, sometimes millions of dollars to build a new data set, we put it all into a graph, think data set, and then data sets are as simple as a query off that graph. So I'd say that the acquisition model and the graph,
are what's allowed 5x5 to scale incredibly efficiently, especially compared to our competitors. And it's allowed us to come out with data sets from what I've seen overwhelmingly faster than any of our commercially competitive.
peers in the space. So tactically, that's how the model works. Tactically, I 5x5 has around 15 generally available data sets with maybe a handful in a beta fashion that we're still kind of testing and bringing to market throughout the rest of this quarter and year. They're very diverse. There's a B2B contact data set that's pretty competitive to like the, you know,
People data labs of the world type type in cognizance of the world stuff like that We have a consumer data set that's competitive to like the experience and credit bureaus the world we've consumer attributes data set which is the kind of core consumer identity data set with hundreds of attributes or attached to it very much like a credit
header type data set. We have firmographic data sets that are complementary to the B2B. We have device graphs, which are very hard from our experience to accumulate, where you've got mobile ad IDs, which is phone device data, IPs and cookies all connected to an individual ID. So you can connect it to the individual and then you can connect that to businesses. So the nice thing with the graph is you can traverse through most device data is consumer driven. Connecting that
to a business is really hard. So like your mobile phone is probably not registered to SpringDB. But if I know it's you and you're connected to SpringDB, I can start to say, hey, if I want to target a C-suite person at SpringDB, that phone is going to be very valuable for targeting. Things we've learned too is
John Kosturos (13:41)
Yeah.
Nick Weldon (13:42)
Targeting B2B is expensive. So if I can find your consumer just because of the economics, you're selling usually set five, six, seven figure items in B2B. If I can connect a consumer device to a B2B persona, the ad space on that consumer device is exponentially cheaper.
So there's a lot of benefits you get to this graph to all of our graph, all these data sets talk to each other through IDs. So you can kind of like Lego at them on. We have some real more real time daily streaming stuff. We've got a tag that can be used for ad act attribution or website resolution. We have behavioral feeds. So
We have a massive publisher network where we have, I think at this point, hundreds of thousands of publishers. Now in this, it's kind of like a mini co-op inside of the co-op. One of the things these publishers want is when a publisher can give more attribution on their traffic, they can sell the traffic for a premium. So they can say these are the types of people on their website. They can sell that ad space for a premium. So we give resolution.
away to publishers so they can get that higher adder attribution and in return we're able to get weblog data so we can start to create behavioral, like a behavioral graph. So we've got a lot of these different segments. We keep expanding, two things we expand the graph on. One, we get more data in that just comes in and we're like, we could build, think of it, use the analogy like we look in our pantry and there's more things to cook out of there. The other is it's demand from the co-op.
So the co-op sometimes has demanded or requested additional data sets to further their products. And then we go back to our massive co-op network and we might subsidize a certain data feed or someone's co-op membership for a certain data feed to bring it in. So yeah, we've got 15 or so data sets. They're refreshed on a monthly to real time basis. A lot are monthly, kind of the identity ones are monthly. The behavioral stuff's daily and then we have some
some true real time tag resolution stuff for ad tech. And yeah, and it all kind of connects.
John Kosturos (15:31)
brought that all
together on the updates and everything. So thank you. think you took us through that kind of entire process all the way even down to the updates. You know, I think where I want to go from there is, and I know this is a broad...
kind of universe, but because the data sets, there's some, many building blocks and those building blocks can be used for go to marketing, even for identity resolution, fraud detection and you know, stuff like that. But what are the most common platform use cases that you guys just like really.
you know, do very well at, you're powering those ecosystems in a big way now.
Nick Weldon (16:12)
Yeah, so I'll talk platform and then I'll talk maturity of those platforms because I think the two talk really well together for where we thrive. So I would say the sales MarTech ecosystems where we do probably our most of our works. These are CRM platforms, email sending platforms, dialing platforms. We do a decent amount with like DSPs and SSPs in the ad network.
A lot of these platforms, it's lot of enrichments. So CRM platforms, and you can think of some of the big CRM platforms, we probably work with most of the ones you can think of, the household names. Their users...
bring in data into those platforms and the better enriched, more recent, accurate that data is, the more value they get out of those platforms. Email sending platforms, they'll use our data to enrich, to create insights. So they might say, hey, your open rate is 42%, but in this cohort, it's 62%. You might want to spend more time in this cohort because they've now enriched that data. In the ad tech world, the whole big thing they're trying to get to in the ad tech world is users will upload data
I
think CRM data so there might be emails and stuff like that and those ad tech platforms are trying to connect those to devices the more devices they can connect to the more Ads that can be serviced the ad tech platforms make more revenue and the customers are more excited because they can fulfill more their their campaign targeting So it's really it's a lot of enrichment so that is something they're already getting There's some provisioning some of these platforms want to provision new data to users in there, but I would say enrichment is really worth
John Kosturos (17:32)
Yeah.
Nick Weldon (17:38)
the best bang for the buck is almost every one of the platforms where we do very well with already has an existing data pipeline. They know how to process large scale data sets. They're either wanting to use us to offset an existing pipeline where we might be a better cost provider. They might have an API waterfall where they want a flat rate of cost because the variable component might be hurting their gross margin.
that all of these companies know how to process large scale data sets. So they know how to take hundreds of millions of records so that they can process it and be pick the value out of it. So the analogy we use is like, we give our data to like oil refineries. Our data is not straight to the gas station. Those oil refineries refine it and make it ready for your car. But they want, you know, we're selling some of that more crude oil for large scale providers where we've seen our model
break, and I think John, this is where we do talk on SpringDB, I think this is where SpringDB comes in and very handy, is when we're the only provider in their data waterfall, and or they don't have the systems and processes already built to refine our large scale data set to make it ready for their end user, kind of gas station ready there. And that takes, as you know, there's processes, experience.
So that's where we've seen where we've created friction with individuals as they took our crude oil and put it right into a car and blow up the engine. ⁓ You and I both experienced that.
John Kosturos (18:59)
Yeah.
they're not alone in that world because I just a long story from my past in my, a past life, I actually purchased 15 different providers, data sets, and we aggregated them and we were spending millions of dollars on the data, millions of dollars curating the data. And at the end of the day, as the leader of that organization, I was communicating to it.
cloud and data engineering team that I didn't have the skills to basically understand everything that they were going through. And we would find that data sets were slow to onboard.
only certain attributes showed up and it's always like a back and forth. So it is important to, you know, make sure that when you're consuming big data sets, you do have experience around. and luckily for you, Nick, think some of your customers are the top data experts in the world, but can you give us some idea around the scale of data across these different data sets? Like how big are they?
Nick Weldon (20:00)
Yeah, so it's a little bit of like an iceberg. So what we send out is really what we've refined, even we've refined down and then Don, I work with you and you guys refine it even more. Our B2B contact data set is...
about just north of 100 million US contacts, they get 700 million plus internationally on the B2B contact side. Our consumers about 250 million US individuals. companies, we're talking maybe 10 to 15 million companies, devices, we're probably close to a billion devices at this point across.
You know, think phones, computers, CTV, wearables. On the websites, we're in the tens of millions of unique web pages now and hundreds of thousands of top level domains. So it's a kind of individual publishers. I'm kind of the raw data.
petabyte a week range, maybe more. it's a lot, because we get a lot of streaming in data, then we dedupe it and we might see the same website ping 10,000 times in a day. And then we have to dedupe that down.
We find too, like, when you're trying to understand is this IP connected to you, John, we want to see hundreds of occurrences on that IP address. So is just you logging in at a Starbucks, or is it something, you know, an office or a home where you're constantly pinging? So we do want repetitive signal to help make sure is this a fluke or not. But yeah, we're probably in the petabyte plus range a week. It's pretty heavy at this point. ⁓
John Kosturos (21:23)
Yeah,
the way I would...
Nick Weldon (21:24)
Everybody
knows petabytes is it's like it's a different level of machinery. So it's not like you can just move it with like a small tractor. Now we're talking big industrial scale tractors.
John Kosturos (21:32)
Yeah, and with experience,
what we've seen is, hey, it could be 15, 20 million companies, 10 to 20 million companies that's updated every month. And it could be 350 million people updated every month. And it could be a billion devices. And it could be, you know,
35 billion website page views every single day, right? And the cycles of those are different. But when you get them all working together, it's some very powerful.
Nick Weldon (21:58)
Yeah, the value of our data set is
if you just need a B2B data set, I'll use that as the example. There might be five or six called comparable-like solutions to us out there. So we might have five or six competitors. Then you need a B2B data set connected to devices. There might be two competitors. And then you need a B2B data set connected to devices, connected to behavioral signals. Not like one-on-one. So our value really is we're probably top-rate quadrant on one data set.
John Kosturos (22:15)
Yeah.
Yeah.
Yeah.
Nick Weldon (22:26)
Once we start adding multiple together, really isn't, there's just not a market comparable. So you have to go and get five, six, seven companies, try to get all their data sets to talk to each other. The commercials don't really map and it gets really expensive and all that. So five by five really does really well. Either we're a cheaper alternative to something in your existing data supply chain or your product roadmap demands innovation. And you need more data sets to keep up with customer demand. And that's where our data talks really well to each other and prices really well to each other.
John Kosturos (22:47)
Yeah.
Yeah, the one thing I really like is, so if you're building a product, you know, when you get all the five by five data working together, you can build audiences, you can do emails, you can do telephone, you can send direct mail, you can build integrations into meta and...
LinkedIn and Google retargeting and every ad network you need and push audiences for digital advertising, you can resolve people that are on your website or viewing your ads. Like it just gives you a great foundation. then like what we've helped companies with is okay, the foundation set and they're still data junkies. And so are their customers. like, I want healthcare and I want education.
Nick Weldon (23:26)
Yes.
John Kosturos (23:36)
So we like stitch additional data sets because you know it's not always just about being cheaper it's like how do you create a mode in this new world of competitive solutions popping up and AI and it's like create very functional foundational data sets and then verticalize and and and that's where we've been really successful with you know five by five
partners and stuff like that because you you guys really do offer a great foundation.
Nick Weldon (24:04)
I look at
that's exactly why I see the partnership on it is we give them the foundation to start with kind of that core identity, which is where you start. And then that's probably good for half the market right there. That's enough.
John Kosturos (24:15)
Yeah, maybe more,
yeah.
Nick Weldon (24:17)
Yeah, maybe
even more. And then there's the other half to third that need very targeted add-ons. And you guys know how to talk really well with our data. So you know where the gaps are. It might be a small business segment that they need more focused on, or health care. There's different attributes they need. We have some health care data, but definitely not the depth that you guys can. And you know to blend it together.
really well to our data set. yeah, that's absolutely. And then the other thing I'd say just through the partnerships you've created, John, is our data might be 80 % refined to what people need. And you have created partnerships and solutions where instead of putting the pressure on a company to figure that out, which some have already, these bigger companies have, there's a large cohort that.
John Kosturos (24:55)
Yeah.
Nick Weldon (24:57)
We all wouldn't recommend they build that the systems and there's companies out there that you've partnered with that can refine it. So you're either refining the five by five data, the extra 20, 30 % the end person needs to go into their product or service and or you guys are bolting onto it. So just adds more complimentary value and they can service their customer needs. And we can't at five by five go do that because it doesn't benefit enough of the overall average of the co-ops. I have to balance investment over overall benefit.
John Kosturos (25:24)
Yeah.
Nick Weldon (25:28)
We can't get very niche because we have so many use cases
out there, but our vision with 5x5 with the core and with our partners is for no product leaders roadmap to be limited by their access to data. That's access, both like literally how do we give it to them. That's licensing terms. So do they have the rights to do it? A lot of data providers might limit what you can do with the data sets. And then that's the types of sets. So as you just mentioned, like there's a lot you can build off 5x5. And I don't think there's many use cases.
John Kosturos (25:40)
Yeah.
Yeah.
Nick Weldon (25:58)
from a broad standpoint, we don't have data for. And then if we don't have it, we go to you, John, and you come and fill in the gaps. So that's really the vision we'd had. That's what wanted. We wanted data. And then we want to make it in a way that it's affordable. So you don't have to go overwhelmingly gray. So those were kind of our tenets. And that's where the two of us have just meshed together.
John Kosturos (26:01)
Yeah.
Very
cool. So Nick, where are you guys going next? Well, before you answer that question, let's go to domestic or global coverage. Where are you guys really focused in terms of bringing in data? Are you focused on US? Are you in any global environments?
Nick Weldon (26:36)
So yeah, up until about November, at least publicly, were just US based.
In November, we were able to release in a beta fashion that we've been expanding aggressively on the international side. It's a cohort that we've seen be wildly under invested in from the data ecosystem. think a lot of it happened during the kind of GDPR side where everybody just took a step back, us included, not knowing the risks and where it was trending and what the regulatory frameworks would be there. I think a lot of us are feeling more comfortable for specific use cases with where that's.
what's happened and where it's trending from a regulatory standpoint. We're seeing that it seems like the EU is being a little more.
called partnership minded, especially in a business standpoint, for how to use that data for business purposes. think it's very restrictive. So quick answer is we have a B2B contact data set that is like our US data set that's globally, from our point of view, of all the data sets I have right now, I think that's probably our most...
Probably one of our best data sets right now, just because there's not a lot of competition and alternatives there right now. So if you're looking for international B2B contact data, there's just not a lot of places to go look, especially in comparison to the US. from what we're seeing, I think we one of the best specifically there. And we were able to use our co-op model and build, go from like, think, last to, if not first, right near first in that cohort. So that's one. From there, we're taking our US playbook.
and just internationalizing it where it makes sense and where we can compliantly do it. So we have device graphs that are coming out like this month or in May, this month. So it's going to be maids, cookies, IP addresses, IP to hem, IP to company. So you can do international device targeting there, which is I think a lot of people are looking to do that.
Those counts will be publishing not as big as our US counts, but again, it's like compared to what's commercially available. That's a big statement we use a lot on is how is the data compared to what's commercially available? We all want perfect data, but as we know at scale, it's hard. I think our device data will be right up there with the top commercially available solutions and it'll talk well to our B2B contact data, which will talk well to our US data. Everything talks well. So it's like, it'll be a lot of co-op members right now.
And it's our fastest growing segment, I would say, in five by five right now is international from a percent of growth standpoint, who are upgrading, who've been waiting into international data and testing it. So same thing, they're going through testing processes and the testing has been coming out quite positive.
So that's probably the hottest growing section right now, where I see us like over the next six to 24 months is to start expanding more into the fraud prevention and identity space. It's kind of the same type of data that's used in ad tech. Like for the fraud prevention space, that's for like e-commerce platforms. They want to understand, hey, is this device seem right, address, et cetera. Same kind of thing used for marketing, just higher standards.
John Kosturos (29:15)
Yep.
Yeah.
Nick Weldon (29:21)
well in fraud prevention, you kill it in advertising and marketing because it's a higher quality standard. So that's probably one of the next big areas that we were in, but we're going to be investing more into that segment.
John Kosturos (29:25)
That's cool.
Okay. So I
do have two more topics and then we'll get into where you headed next. And I know we're getting kind of late into the game, but you guys have so much to cover. One of the things that you mentioned earlier was terms. And you know, if you're a platform or you're an agency and you need to consume data at mass and you need to forward license it to an end user, it's not always a straightforward game when you're going out to data vendors. Like you really.
You understand how to negotiate terms with vendors to be able to use their data in your product or service.
Nick, tell me how you guys, because you're kind of unique. I mean, there's a handful of companies in this space that just sell the platforms and you're like, not after the end user consumer, right? So tell me how you guys make it seamless for a platform to just buy data, onboard it and start providing it to their customers.
Nick Weldon (30:28)
Yeah, so there's two buckets I would say. One is the licensing terms. One is the actual access to the data itself. So on the access side, for the majority of our data sets, we're giving you a full file in your environment so you can query it once, 10,000 times, 100,000 times, however you want to query it. So that gives you the flexibility from a product leader. The other side of this too is on the licensing terms.
there's key pillars that are important to when you're buying data. What are your rights? Are there derivative rights? How do you forward this on your end users? Do you have to delete the data if you leave? And so we have standard licensing terms. We also have flexible ones where people can pay more for certain rights that they want. So it's just like when you're looking at the data side that
Where's the liability stand? Like there's just a bunch of these, these different things. We're pretty good to, we don't get a lot of red lines on a lot of our stuff. And then we do have modular ways where somebody needs some extra rights. We have the ability to do that. There's just a cost component to that, whether that be because we take on more risk or they want more value out of that data set. And then we can use those funds to go either invest more in compliance internationally or build stuff with it.
John Kosturos (31:38)
Yep.
Nick Weldon (31:41)
So I put those are the two buckets if you're looking at it and they're different for your sit your where your phases in the business so your earlier stage Access might be a different demand than when you're late stage The thing too is when we have a lot of data sets which a lot of these companies were finding need more types of data Adding more data sets on is just a skew ⁓
John Kosturos (31:51)
Yeah.
Yeah.
Nick Weldon (31:59)
So you don't have to manage seven data sets with seven different licensing models with seven different access models. And that's where I think the complexity gets really tricky where you can use this in certain use case and that not a use case and all that. That's why I also think having one kind of MSA stack, one access stack simplifies that down quite a bit.
John Kosturos (32:18)
Yeah, I think just experience working with you and then other providers is number one, that's what you guys do. So your paperwork is structured toward enabling these types of platform use. Whereas like others, like got to tell them that it's an end user agreement. Like you got to get to the right department. It's a whole game.
Nick Weldon (32:30)
It was all built around that, yes.
Well, also have a lot of ones. Well, big thing too is a lot of the companies
out there, they look at it as if you're using more data, they could have charged you more for the data. lost a lot of you.
John Kosturos (32:48)
Yeah.
Nick Weldon (32:49)
In our case, we get the usage back. We want you to send 10 billion emails, because then we get 10 billion email signals back. So we have an incentive system that's really built around making it easy for you to get the data, use it as much as possible. And we actually capture the upside value, not through charging you more or restricting you, but through the contribution. So that's, say, one of the big differences why our model didn't have to change, where a lot of companies might have had to.
John Kosturos (32:56)
Yeah.
Yeah.
Nick Weldon (33:14)
or they have this kind of usage-based model, because if you grow a bunch, they want to charge you a lot more. Versus us, we'll either give you more data sets, we keep coming up with more, or we get the usage back. So that's helped us stay very true to that model.
John Kosturos (33:16)
Yeah.
Yeah.
Yes.
All right, last question before I get into the future. Intent is a big kind of keyword in our space. And I've gotten used to with the traditional providers seeing a business type intent where a business is surging through multiple individuals on a topic and...
Now I'm going in and building buying groups and just doing AVM advertising and just, we call it whitewashing that company and just trying to get in front of everybody who's relevant. Tell me how you guys are kind of enabling this idea of categorized intent that breaks down to a potentially individual.
Nick Weldon (34:12)
Yeah, and I would actually say we've got 90 % of the ingredients, or 100 % of the ingredients, but we take it 90 % of the way to intent. So we're not trying to say this person is in market for this solution. What we are is we're an observation network. Almost everything we see is deterministic there. So a couple things. One, we look at almost everything in our graph person out. So the devices, their attributes, even their behavior. And then if you want to get it to a company, you just roll it up.
John Kosturos (34:21)
Yes.
Yep.
Nick Weldon (34:36)
into a company. our behavioral graph is person out, which lets you get person level intent. But what we're saying is we're not saying this person is in market for this solution. What we're saying is we've observed a tremendous amount of web pages that are connected to this topic and connected to this person. So we're saying this person has observed a tremendous amount of content related to this topic.
it's the co-op member gets to decide what is intent. So is it, I'll use buying a home on the consumer side. Buying a home on the consumer side could be, looked at homes, mortgage rates, moving companies, neighborhood safety.
That might be true intense. If it's on the company side, might be multiple people at that company over a certain seniority level have observed it over a certain amount of time. So we're getting like what five by five, but we try to do everything. It's a massive observation network. saying we've observed even when we say an email is good. We're not saying this email is good or bad. We're saying we observed an open or click on this date.
John Kosturos (35:34)
Yeah.
Nick Weldon (35:35)
You can decide is that a high enough level activity or recent enough for what good, because what might be good to you might be bad to somebody else. It depends on what they're, they have a Flamin' Yone kind of steakhouse or is it more their wide scale and they want to go more maybe McDonald's. Both can make a lot of money. So there's different versions of what good is. So first is it's person-based, which is very unique and hard to do. Second, we have a massive network of publishers. So we're seeing a tremendous amount of behavior and content.
Third
is we have a very good, this is one of the probabilistic but high confidence probabilistic things, it's a categorization engine there. So we say, hey, this page was about this piece of content. And then we just link it together. So if you're building and you get to build your own intent model off it, but we're taking a 90 % of the way there, the hardest part of that, the two hard parts of pulling off a person-based behavioral product, one is you need a massive network of publishers.
John Kosturos (36:14)
Mm-hmm.
Nick Weldon (36:29)
And that's an earned network there. So that's one. The second is you have to be able to connect that massive amount of web behavior to identity. And then that identity to other useful attributes. Be it consumer, B2B attributes. So...
John Kosturos (36:38)
Mm-hmm.
Nick Weldon (36:45)
For like what we're seeing is it gives individuals the flex or companies the flexibility to build their own little secret sauce connected to it versus a black box, which is like some of our counterparts. A, they're not person based, which is limiting. And B, it's a black box. what they may say this company is in market for intent. You don't know what that means. Was it a hundred observations? One, C-suite, number of employees. It's a secret sauce.
John Kosturos (36:52)
Very cool.
Yeah.
Nick Weldon (37:11)
So it's hard for you as a product leader to build off that because you don't know what went into it. Worse, this person observed a lot of content related to this. You can roll it up and decide what you think a good score is.
John Kosturos (37:15)
Yeah, if you're a product leader, can almost just...
Yeah. Like if you're
a product leader, you could just build that into your UI and like let the customer create their own like, yeah, that's pretty cool.
Nick Weldon (37:27)
It could be a feedback loop based off campaign performance. Hey, it's like,
we've seen people build regression models off campaign performance that it might be multiple topics unrelated, like the, you know, that you would have expected. And that's where the true intent comes from. It's people looking for homes with blue jeans. Great. I don't know. So that, lets you, you don't have to, you could build a training model that trains off of campaign performance to then create the intent. And that's probably where the best people are.
John Kosturos (37:38)
Yeah.
Yeah.
Yeah.
Nick Weldon (37:55)
now in the car.
John Kosturos (37:56)
One of the things we've actually done for that is, I mean, one holding a week's worth of intent files is one thing, but holding six months or a year is a whole other ball game.
Nick Weldon (38:06)
This was terabytes of data today.
⁓
John Kosturos (38:08)
Yeah. And then
like processing against it. So like we built look back periods that allow them to process quickly and no cost per query type of access. And yeah, I mean, I like the way you explain that though, is that you again, providing the building blocks for a platform provider. Yeah.
Nick Weldon (38:25)
It's a massive observational network. We're observing
this behavior and then we use our graph to link the behavior back to identity. And then from there you can traverse. Where like where you come in, John, in a really positive way from my experience is one, just like you can make the data exponentially more usable. So way more companies can access to it because it's a heavy data set. So I'd say the amount of companies that can process that amount of data daily is very few. We work with those very few. So you make that data accessible to a much wider range of
companies that normally wouldn't get access to it. And the second is some of those companies might need additional data sets layered onto it. So because it's just behavior plus.
John Kosturos (39:00)
Yeah. Or they don't want to pay AWS 50. Yeah,
they don't want to pay AWS 50 grand a month to process all of it. Yeah.
Nick Weldon (39:06)
Exactly. Well, we
use this analogy all the time. It's like, if you look at behavior, a lot of people are in market for Ferraris, but who can actually afford them? And that's where there might be supplemental data sets to qualify the intent beyond just the behavior. So that's the other side of that keto intent is looking at the behavior and then qualifying. Does that individual or company actually have the means to acquire what they're looking at? And that's...
John Kosturos (39:17)
No.
Yeah, one of the cool things
that we've done is...
you get, you you circumvent the different intent categories and then you create an audience and then we merge it with a social affinity audience. So we can see, okay, what are these people have affinity for? And then we took it a step further and said, who influences those people? So you could do like celebrity endorsements down to these people that were showing research on a topic, but you have to stitch like a social graph
to the other signal graph, to the identity graph, but when you do those things, it's super powerful, like transformational.
Nick Weldon (40:06)
And that's where almost every company we're seeing, whether it be with our data or other data sets, they're layering multiple data sets. If anyone could just buy a data set off the shelf and build a seven, eight figure company, there's no competitive edge.
John Kosturos (40:16)
Yeah.
Nick Weldon (40:19)
there. So all the ones we're seeing, they're building stuff off of our data and or combining multiple data sets there. So like we think there's a lot of value. If I have research behavior that says you're into golfing and then there's survey, a completely different location based or survey based stuff that says you're into golfing, that's completely independent. The odds are very low that you're not into golfing because it's just too, so like starting to do like almost like a Venn diagram of overlying data.
John Kosturos (40:38)
Yeah.
Yeah.
Nick Weldon (40:46)
Almost any company we're seeing that's not doing something on top of regardless of the use case or really even size, those are the ones we see struggle. The ones that we see skyrocket quickly have a strategy to take our data and or other data sets and build some sort of recipe on top of them. To give feedback loop, there could be multiple ways to pull that off.
John Kosturos (40:55)
Yeah.
Yeah. Even if it's their own first party information, right? Like censored or whatever it is. Yeah.
Nick Weldon (41:11)
But if you just take
it and then just build a little UI in front of it, that should be a lifestyle business ⁓ there. Which is okay, but you're not gonna probably build a meaningful seven, eight figure business without building a model on top of it. And that's where you guys come in. You make that way more accessible to early stage companies. And then for companies that need niche stuff, they can come to you and say, can't service. We refer them to you all the time.
John Kosturos (41:16)
Yeah.
Yeah.
Yep.
So we only have a few more minutes left, but where is 5x5 headed?
Nick Weldon (41:38)
Yeah, I put two buckets that we're focused on, or three buckets, okay. One is just bringing out more co-op members to enhance and improve our current graph and data sets. So we're always looking for more of what we have to improve. So that's one. Okay, the second is a TAM expansion. So taking our graph and bringing it, it's still platform-based use cases, but bringing it into fraud tech and the kind of identity tech space.
It's similar data sets, but they have their own nuance, things they need for what they consider quality, usability, et cetera. So I'd say the fraud space is probably gonna be one of our largest expansions over the next 12 to 36 months. It's also a harder space to be in. So I think it'll just overall improve everything we're working on. The third is, so five by five, we have a very horizontal data strategy. We keep adding on more data sets off our ground.
So we probably, I think definitively, we have the widest data offering in our kind of wholesale.
data access category. Most of our competitors have very siloed. They build up one, maybe two or three data sets. We have 15. So we have a horizontal data strategy. One of the things we're working on right now, and we've just acquired our first company and we're going to be acquiring more, is acquiring companies in each of those maybe data categories that are specialists. So these might be smaller companies that did a really good job at one data set, but their revenue has maybe stalled just because they won, but have gotten really good at that.
quality or sourcing or something very niche or a model or something that like would take us a lot of time and effort to build. And because we have so many data sets, we can't invest maybe that time or capital into that. So we have private equity partners. We can leverage different vehicles to acquire these. So.
We've acquired one. I wouldn't be surprised over the next 12 to 36 months we acquire a handful more companies that specialize in maybe device data or consumer data or maybe some validation vehicle or international. So we're actively looking. So there's companies that fit kind of that.
Maybe like three to 10 million ARR, they're more bootstrapped, a little bit of profitability, and they have one of those specialties. Like we're an active looker. So yeah, I'd say improve what we have, a new target market segment or an expanded target market segment. We're still in there, but we're expanding in there. And then acquisitions, those are probably the three big buckets that should keep us incredibly busy.
John Kosturos (43:48)
Very cool.
Absolutely. And that's also a question if you're buying data, you should ask your data providers, like what is your strategy to continue to grow your data or to continue and invest in your data. it's exciting to hear that you guys are making that a top priority and we will be keeping an eye on you. And hopefully next time we get together is in Austin or in Miami in person. But Nick, it's been incredible having you on today and we look
Nick Weldon (44:22)
That's right.
John Kosturos (44:26)
forward to kind of continuing to watch you guys's journey. Absolutely.
Nick Weldon (44:30)
Thanks, John, and we appreciate the partnership. It's been a fun journey
together.
