The Hidden Revenue in Bank Data with Oban MacTavish

Spade has raised a $40m Series B, and I was lucky enough to sit down with Oban MacTavish, CEO and co-founder, to learn about how Spade makes merchant data usable, and to hear all about their fundraise. 

The data your bank gets when you swipe your card is shockingly bad - a jumbled descriptor of ~50 characters, a merchant category code that might be wrong, and a "city" field that's sometimes just a phone number. 

In this episode I learn about why payments data hasn't meaningfully changed in 30 years, how Spade built a body of merchant data covering 99.9% of payment-accepting businesses in the US and Canada, and what banks can actually do once their merchant data is good. 

We also discuss Spade's freshly announced $40M Series B led by Oak HC/FT with participation from Andreessen Horowitz, what nearly 500% YoY growth breaks inside a company, and why Oban is saying no to obvious opportunities to stay focused.

Transcript

SPEAKER_02 0:00

We have merchants who show up at our door. They say, hey, we're having trouble. Could you tell people that we're actually this business? Like we don't know why our descriptor is so messy. Can you please tell people we're John's pizza down the street? We have all these people calling, they're really pissed off.

SPEAKER_00 0:17

Hello and welcome to In Control, where we learn about the finance of everything. I'm Tash, your host, and I'm also the CEO and founder of Cable, which supports the show. If you enjoy this podcast, please subscribe on Apple Podcasts or Spotify and share it with your friends and colleagues. When I joined Monzo back in early 2016, one of my first tasks was to enrich merchant data. What that meant was every morning I would log into our homegrown BizOps tool and be confronted with a slew of merchant enrichment tasks showing the data from MasterCard about where customers had spent money. Usually something nonsensical, like EC23 S T A R 123. And I would also see a postcode or a zip code and a little Google Maps view of where the customer had actually spent money. Using this information, I would work out what the shop or merchant name was and I would add it in and upload the logo as well. So instead of a customer getting a push notification saying they had spent$5 at EC23 S T A R 123, it would say Starbucks with the logo. I guess I had assumed that over the last 10 years, this problem had been solved by banks everywhere. But it turns out that it has not been. And so today I'm speaking to Oban McTavish, CEO and co-founder of Spade, to understand what banks' merchant data looks like today, how they use it, and the hidden value that can be unlocked from within this data. I am also excited to find out more about Spade's newly announced$40 million Series B fundraise led by Oak HCFT with participation from Andrews in Horowitz, Flurish Experian, National Bank of Canada, and Y Combinator. Oban, thank you for joining me today and congratulations on the fundraise. And I want to start off by understanding what payment data actually looks like today. What is it that banks see when their customers spend money?

SPEAKER_02 2:12

Thanks, Natasha. I think uh payments data, as a sense today, is terrible. I mean, nothing has changed really in payments data in the last 10 plus years, and certainly not in the last 30. If you think about where we started, we had the networks setting up giant payment rails that were moving around bits and bytes of information. You can think of this like those really messy descriptors you've seen and those merchant category codes. I think we're all familiar with. Um, and it hasn't changed much. The networks came out with a new standard, 222, that was supposed to like revolutionize a lot of the information. They continue to add additional data about people. Is it going to be a gentic commerce? I think that's actually a field that'll be coming in within the next couple months, too. But nothing much has changed, which is kind of surprising, but actually makes a lot of sense if you think about how the system works. You have the networks moving data between acquirers and issuers, and everyone kind of has to agree. And we have issuers all over the planet who rely on these pieces of information who are parsing these bits and bytes to make really, really important decisions, whether it be fraud, whether it be what you're showing your customer, whether it be what reward you're giving, all of it's built on this underpinning of low quality data. And making any changes to that is really, really hard. And I think that's probably why we're, you know, we're living in a world where not much has actually changed since you were looking at that data back in Monzo.

SPEAKER_00 3:23

So what why is it that the data looks like this? Is it does the does a merchant have anything to do with this? Like when Starbucks sets up a new terminal, are they able to like put in their actual name? Like why is the data so bad?

SPEAKER_02 3:39

It's a great question because the merchant does in some ways. So when you sign up for a payment terminal with somebody like a Stripe or a Square, you do give them a lot of information about your business. And I think Stripe actually gives you the option to like adjust how you're going to appear in the descriptor. Um, many of the old school POS systems, it's a little bit more jumbled. But the unfortunate reality is that because of the rail that we're all existing on, there's only so many characters that are getting moved. Like the descriptor field, which is supposed to include the name of the business, where it's located, and other contact information, is only around somewhere between 40 and 50 characters long. So you can imagine if you're trying to fill all this information into a small amount of space, it actually doesn't matter. You could be like, hey, here's all of my information to my acquirer. They might not even be able to move it fast enough during that off-stream. And I think to make matters worse, the incentives are a little bit mixed, right? If I'm an issuer, my incentive is to protect my customers' funds. It's to make sure that I'm only letting in the safest transactions. It's to protect Natasha if someone steals her credit card and to create like a really seamless experience that also feels very safe. If I'm a merchant, my incentive is actually often to maximize how many sales I can make. It's to sell to people as soon as they show up at my door and make that really seamless. So if I'm a merchant signing up for a POS system, that means I don't want to tell a bank that I'm a crypto company. I'm gonna tell them I'm a large digital goods merchant. So oftentimes there's some incentives at play here that means when a merchant is signing up for to process payments, how they're describing themselves might be a little bit generous. You know, most businesses aren't trying to be manipulative, but I do think there's some reality that like the incentives here aren't necessarily super aligned for a bank to know exactly where you're spending your money. It's you know to make sure they're getting as much information that to reduce chargebacks, but not actually so much that they're gonna impact their payment performance.

SPEAKER_00 5:23

So for those who don't know who are listening, POS means point of sale. And so if you're a merchant, all good, just have to. My wife keeps telling me, you know, my wife keeps telling me to uh make sure that when when we're talking about all these financial terms that we explain what it is. Um so when a a merchant um wants to actually like take payments from others that you know, they have their own bank, but they need to get somebody else involved so that they can actually take those payments. That's their point of sale, uh, their point of sale vendor, like a they get a terminal from them, maybe it's Square, maybe it's Stripe, um, maybe it's one of the old school vendors. And then that information is kind of set up with that point of sale vendor, but there's so many people involved in that too. There's, you know, there's the MasterCard or Visa. And who is it that is actually setting the standards for what this data, what this merchant data looks like? Is it MasterCard and Visa? Is it the banks? Who is the sort of determining, um, the determining business in that respect?

SPEAKER_02 6:17

So the network set standards for the data they're gonna require you to send. Um, and then, you know, so that really everyone's listening to the network. The network says, hey, this is the protocol, this is all the information we need to move, this is what people need to sort of authorize these transactions. And they set a lot of the standards that need to be collected. So if you're a merchant, if you're if you're an acquirer, a merchant acquirer, uh, think like Square, you have to collect some certain information from them. And then obviously there's also regulation in place here, like uh, you know, certain types of businesses aren't allowed to process payments or for a very long time weren't in certain states in the United States. Um, you know, it's very difficult to be banked, things like that. So one of the things that happens here is that there's a requirement that you collect an MCC code. But validating what this MCC code, um, whether or not it's actually true, is like a hard problem. Uh, it was a lot simpler when if you think about if we go far enough back, the only people processing payments were literally in person. Like you were going to a restaurant and you were saying, Hey, uh, you know, do you want to accept these really newfangled cards that mean people don't have to carry cash? It's like, of course. And you you gave them what was called a merchant ID or a mid, which the idea was like this was an ID that represented that POS system, it represented that business. Because in the past, that one ID would represent your business, you're your corner store, your restaurant. You don't have 25 POS systems, e-commerce doesn't exist. So that mid made a lot of sense. And, you know, MCC codes, if you go look at them, I mean, there's NCC codes for like, you know, industrial equipment and stuff like that. Like this is just really a list that's grown and grown and grown. And I think it it just the tool is not built for today. And what that means is that, you know, even if we take away bad actors, I think fraud is just like a challenge everyone in FinTech and financial services faces, like those people are manipulated. But even if you're trying to do your best, you're just trying to find the category that makes the most sense. Maybe there isn't one that fits your business. Like there is no like a marketing software category. So how would you define yourself? Are you a large digital good? Are you a computer electronics retail store? Like some of these times when you're sending B2S systems, like you're just choosing what looks best and your payment processor does some level of validation, but how could they validate it, you know, at scale?

SPEAKER_00 8:18

And so MCC code stands for merchant category code, is that right?

SPEAKER_02 8:22

Yeah, exactly.

SPEAKER_00 8:23

So everybody is basically choosing the one that fits best with whatever it is they're selling. Okay.

SPEAKER_02 8:28

Exactly.

SPEAKER_00 8:29

And so and that is when a when a consumer gets their bank statement, they're kind of sh and and they see those long, nonsensical mixture of letters and numbers, which is supposed to represent where they spent money. That is the information that has been required by the MasterCard or Visa, the Rails, and the POS system has enabled the merchant to put in, and it's kind of gone through all this matter of movement through all these different vendors, and it's finally landing on the consumer statement. And that is literally all the data that the bank has. Is that right?

SPEAKER_02 9:00

Yeah. So you can think of it that the so all of this is happening in the ISO 8583 protocol or the ISO 222. And banks get that all that stuff you're seeing, that descriptor, is or that's the 40-ish characters that describes the business. That is in in 8583, that's like it's one of the data elements. So they call it DE42 or DE43. It's a lot of how people will talk about these things. But really, they're getting that about the merchant. They get that merchant category code, uh, and then they might get some other information about the merchant, maybe a little bit of geographic location data, like what state is it and if it's the United States, what's the website? What's so funny is that in that DE42, uh, in that DE42, you're actually gonna see cases in that descriptor, you're gonna see cases where you're supposed to put in the location, but for an e-commerce business, they realize that, like, okay, well, so is it all San Francisco? It's like, no, people put in websites, people put in phone numbers. So the city you bought something in is often a website or a phone number for e-commerce, which creates even more confusion. So the bank gets that information about the merchant, then of course they get other information, like, okay, did you swipe your card? Did you tap your card? Did you punch it in? I mean, there's even like, did you say it over the phone? So there's like, how did you actually input your pen? Was a person there? Did this happen uh virtually? So there's a lot of that information that gets moved around. And I think what I find so interesting is that when they were building this up, you can tell this standard started in a time with bits and bytes, because they were really trying to minimize how much information could be put in there, not for a nefarious reason, but because, like, okay, we got to build the system to be instantaneous, global, and also simple to ingest. So you needed literally, it's all fixed-length fields. There's only so many characters you can put in here so someone can read it instantaneously and stick into their system. And there's so many times nowadays when you're talking to old financial institutions who are storing so much data because you know, by regulation they have stored for pinch up seven years, they care a lot about how many bits and bytes of data you're moving around because they're like, oh my God, like, you know, if they're doing a billion transactions a year, suddenly the bits and bytes are getting pretty immense if you're adding in all this information. So it's a it's a real, it's a really hairy problem. So your bank really doesn't know, like, they do not know the address of all your transactions. They actually know a lot less about what you spend your money on than what you think they do.

SPEAKER_00 11:12

Interesting. And where are they storing all of this information? Are they storing it in their core banking system? Is that where it comes into? Like how how do they receive this information?

SPEAKER_02 11:22

Yeah, so historically, all of this information was trapped at the core level. And it's why, like technically, what you see as a consumer, that like messy description is coming from this, the the processor. So if we got really, really if we like really get in the weeds here, you swipe your credit card at the POS system, it goes to your the acquirer, they are passing it to the network. The network is sending it to the issuing processor. So in some cases, like Visa has Visa DPS, which is like they do the processing too. But in some quite uh cases, that would be like a Thesys uh or an I2C. Um and the processor takes that data and now gives it to the bank or gives it to the core. So historically, that data would go to the core, and the core then does some transformation on this. The banks historically didn't even have their own software. All they're doing is really buying widgets from this massive on-prem uh on-prem database. So, what's so funny sometimes is that they'll take that data and the core says, here you go, we'll just display it. They're making their own transformations and it really lived there. And for the longest time, like banks could not intercede in transactions in real time. So when you were swiping, you didn't even banks could set rules ahead of time, but they could not actually say, hold up, I don't actually want to authorize this transaction, which is crazy because they're the ones bearing the risk. So for the longest time, data was just stored at the core level, so in some on-prem database with like a you know a Pfizer or an FIS, et cetera.

SPEAKER_00 12:42

Yeah. That one of my favorite uh things to talk about in the financial crime space is like real-time transaction monitoring, um, being able to actually stop transactions before they before they uh before they are authorized. So yeah, that not able to do that if the data is with the core like that, yeah.

SPEAKER_02 12:58

No, especially when it was only batches. So they would have to, you'd have to literally like request a batch. So it happens at the end of the day, and they'll dump massive files, CSVs, to like a bank customer that they then store in a database.

SPEAKER_00 13:11

And so if this problem has been around for ages, this data hasn't really changed much, why has no individual bank, why has no technology company, why has a core not previously actually like solved this problem and like built some clever uh transformation layer on top to provide this value back to the banks?

SPEAKER_02 13:31

Yeah. I mean, they've tried. I think what's so interesting about this is like it is a problem that has been in plain sight and has been people have tried to solve internally, they've invested. Many large banks have a service they've built internally, but it's just it's a problem that will never stop, which makes it a really hard problem to invest in. If you're like, hey, the number of merchants that open and close could be millions in any single month across the United States and Canada. That's a huge hairy problem. It's not just like, hey, we're gonna find the top 10 and then we're set. No, this is something that you'll have to deal with for the rest of your life. For the your business will have to be spend money to sort of wrangle this problem and tackle this problem forever. And like that doesn't even include what happened was modernization hit POS systems, where suddenly you have this merchant ID that was representative of a merchant. Now Square has potentially hundreds of thousands of merchants under a single mid. So if you had your database set up such that the mid was like the core identifier, mid equals business, that just broke. Now every single thing.

SPEAKER_00 14:28

It would be for square. So it would just always show square. Yeah.

SPEAKER_02 14:31

Exactly. And now there's like it just like, you know, everything breaks. So now you're at the mercy of this constantly changing data and just becomes a massive pain in the butt. And I think what is so interesting about it as a problem space is that it's not that people have never tried, it's that most people who have tried have done internally and then they invest heavily in the system. They spend millions and millions of dollars. I think of it sometimes in the same way I think of like AWS. Like before AWS, every single business was required to have people who manage their server racks, to have expertise in building out this infrastructure. And then AWAWS said, We're gonna do it cheaper than that, because you don't have to hire people anymore. We're gonna do it better than that because it's infinitely scalable, and we're gonna make it available to everyone. And I think that's one of the things we try to think about at Spade is like, how can we make this instantly available to people better than any type of system they could make themselves and also do it at a cheaper cost because you can do it at scale, things like that.

SPEAKER_00 15:20

Interesting. Okay, so if it if all you're getting is this like truly terrible data and it's always changing and there are always new merchants coming and going, like what can actually be done with the data? Like, what is the solved version of this?

SPEAKER_02 15:36

Yeah. The solved, I mean, I guess spade, this is where I have to say, like, that's what spade does. Um I think the way to think about this is that we've done something very special and very different than how most people would approach this. So most people who tried to do this, they said, wait, machine learning, amazing. We're gonna take all these transactions and we're gonna clean them because there's some like nugget of truth within this thing. Like somebody signed up for a POS system, there has to be some truth in this. And I think that was sort of the the the wrong way of looking at it. Because if that had worked, Visa MasterCard would make this a solved problem. If that had worked, JP Morgan Chase or Platt or any of these people who consolidate immense amount of transaction data, but it didn't work. Um, because there's also just quirks. Sometimes it's just the person who signed up for the POS and did it 20 years ago, and now they sold it to their friend, or they moved it around and the name changes and things like that. So the canonical source of truth cannot be the descriptor. So we did the opposite thing. What we did is we went out and sourced and built and constructed and created this massive corpus of merchant data that covers about 99.9% of payment accepting businesses in the United States and Canada today. And we are a data company. We wrangle this data, structure it, verify it, enrich it. And then in real time, when you send me a really, really crappy payment, I will figure out, think of it like a search problem. I'll figure out which business that belongs to and then give you all that information back. So we actually made this into a search problem instead of turning it into like a cleaning problem, if that makes sense, or a classification issue.

SPEAKER_00 17:01

Oh, that's fascinating. Okay, let me repeat that back to you to make sure I get it. So so most people to solve this issue did what an automated version of what I was doing at Monzo, which is every time a new transaction comes in, look at the descriptor, look at the MCC code, look at the zip code, and try to identify exactly what business it is, enrich the data in real time and send it back. The problem with that is humans make mistakes. And so a human setting up a terminal might put in something wrong, or humans are just unpredictable, and I might have sold my point of sale terminal to someone else. I might have given it to someone like there are issues, people are not perfect. And so what you did instead was firstly build this massive set of merchant data. And so you basically went to try to get all of the data that you could on all merchants in America, North America, and Canada, and enrich that with the logos, with the descriptors, with the addresses. And so then anytime you get some merchant data from a bank or a fintech, instead of transforming it, you're just saying which is it most like? And presumably there's like some kind of confidence score there. And then you can like immediately just send back the correct information.

SPEAKER_02 18:13

Exactly.

SPEAKER_00 18:14

That's brilliant. Okay, okay. So tell me how you got this data. Where does this data come from?

SPEAKER_02 18:20

Well, we can't. Some of it's I mean, some of it's public, some of it's, you know, we have partnerships with businesses, we have POI data, state registry information, um, we have first-party merchant data, third-party merchant data, we have data from other places within payments. Um, a lot of it we generate ourselves today. I mean, when Spade was born, I guess, you know, we had no transaction data, which was also part of the reason we did this. Like, if you're a bank doing this, you think you want to use your assets. It feels very cost-effective. You're like, I have all this transaction data, how do I solve this problem using that? We had nothing. So a lot of it was like hand labeled, like literally building, you know, literally going into a list of every Walmart in the United States and hand creating the Walmart brand concept in every single location and curating this data ourselves. And then now a lot of that's obviously automated at scale. Um, and yeah, and we stick it all together and do a lot of data science work. Like sometimes we joke that like our product is data. Like we data scientists are like engineers at Spade. Like, if you're a data scientist, oftentimes you're sort of relegated to this idea, like you do analytics, or maybe you're writing a model, but like at Spade, so much of it is like, how do you wrangle this massive asset? And then we also get a lot of feedback. So we're very unique in that our customer is the issuer and the issuer is incentivized to have the truth. So the bank or the fintech will send us very, very unique data, truck GPS location data for a lot of transactions, uh, you know, uh other types of GPS data, feedback. Like if you get it wrong, they'll literally say, like, hey, this is wrong, because you know, they complain about it. So we get that feedback too. So we have this very tight feedback loop that's creates like curated and validated data. And we're even, you know, not yet, but we have plans to start reaching out to merchants directly. Because we have we have merchants who show up at our door. They say, Hey, we're having trouble. Could you tell people that we're actually this business? Like, we don't know why our descriptor is so messy. Can you please tell people we're John's pizza down the street? We have all these people calling and they're really pissed off. Um, so it's very interesting. Like, that's our whole job.

SPEAKER_00 20:14

That's really cool. Okay, I got a couple of questions there that that that came to mind as you were talking through that. The first is around like confidence. Um, do you do some kind of confidence scoring and like how confident can you be? And like what does that look like?

SPEAKER_02 20:28

Yeah. So we we have very, very talented engineers. Um, they built, we have a neural net that we run in real time on the transaction that does a comparison and takes in an immense number of factors to decide how likely is it, given the input they gave us, that this merchant is the correct merchant. So essentially, like if you think about it, if we manage like 36 plus million identities, we'll be able to triangulate down to a couple, you score them all, and then you return the most likely counterparty, as we call it. So we have our counterparty match score. Um and we're very confident. Um, we measure, I mean, we are a very unique business in that we are completely measurable. And what I mean by that is when someone is looking at a vendor like Spade, they could count how many transactions were right on. You could theoretically, with infinite time, look at every single response we gave you and tell you if it's right or wrong if you had the correct answer to the question. And so we're we're very unique in that we can measure this and we have accuracy like far north of 99.5% at scale. Um, so we do a lot of dip tests, we pull all of our data out. Uh, you know, we have people to go label the data, we can track how many error reports we get, and you can sort of triangulate like how good are we actually at this? And we have the by far the most accurate product in the market. Um it's a constant battle. It's very fun.

SPEAKER_00 21:41

Do you do you give your customers, the banks, do you give them a confidence score with the with the data, the the transform data? Okay.

SPEAKER_02 21:48

We do, we do. And uh Yeah, that's an interesting experience.

SPEAKER_00 21:54

And if people if merchants come to you and say, you know, I we want you to Clean up this data. We're having this problem. No one knows what it is we do. How do you think about fraud in that, in that regard?

SPEAKER_02 22:06

Yeah. So that's one of the reasons we haven't opened that up at scale. Like, I have big dreams of having like a self-serve portal where a merchant could just show up and do this. But then I'm like, do we really want to open up our data to those sort of external factors? I think we've, it's why we've resisted it. A lot of what we do though today is validating our merchant data. So like, how can you verify? Like, if you think of a problem, right? If you have five different ways to represent target, how do you know they're all the same business? Because there's nothing stopping another business from calling themselves targets. And that looks pretty similar. And could that be a mistake? And they're in the same postal code and they're actually on the same block and they're next to each other. Like, how do you know? Or like abbreviations, you know, ABC pizza is like Anthony, Bob, Charles pizza, but it could be the same thing. And how do you make sure these things are similar identities? And because we're so good at entity recognition, we do a lot of things where we're like, okay, we see a lot of website content. And now we get feedback from customers. So if a customer tells me there's a lot of disputes at a specific business, can you take the legal entity information and take that to your knowledge graph and say, wait, this random PO box in Miami that sells vitamins also has five other businesses at the same place. We could connect all these things and tell our customers that's fraudulent. So we actually do a lot of that work where we're doing a lot of scanning of websites and correlating all these factors to say, like, hey, this is all of these businesses look really similar and behave really similarly. Like if we see fraud at one, you can sort of permeate that through your network. So like thankfully, we don't, we're really like we've most of our data is focused on factual information. What do they do? What do they say they do? What's the MCC code? But what do we think the business does themselves based on how they present themselves?

SPEAKER_01 23:45

Yeah.

SPEAKER_02 23:45

We haven't yet really like I'm not giving you a stamp of confidence to say this is a good business. We don't do that. We do flag some businesses as high risk, but we kind of stay away from the idea of being like, yes, these people like are who they say they are or are not compromised or something like that.

SPEAKER_00 23:59

Yeah, interesting. When I mean, obviously with everything that's happening with AI in the world right now, something that always comes to mind when I think about a particular business's um operations is does every release of an of the new model make their business like better, significantly better? And it sounds to me like one of the areas that you are probably using AI internally and get significantly better with every model release is that information gathering. Because like at the beginning, maybe it was literally you Googling something. And now it's like all of the agents can go out and do that information gathering for you and compile all of this information to change your confidence levels with the with regards to a particular merchant's data. Is that right?

SPEAKER_02 24:41

Absolutely. Um we spend an inordinate amount of time thinking through like how can we collect and validate our data? How do you verify that this latitude and longitude is correct? Like a little known fact is that like, so latitude and longitude, there's like a varying degree of certainty, and like the worst is like inter-interpolated, which means you just took like an address and then you just converted that to like some point um on earth. And the problem with that is that addresses are actually terrible. Like if once you start looking at a lot of addresses, you realize that like there's actually so many different ways to represent the same business. Like postal codes are actually different if it's right on the edge of a two states. You know, people argue about their addresses. If you look at like um highways, sometimes they'll just have north or south on it. It will be describing a plot of highway, north or south. So like we spend so much of our time saying, how can we validate and improve our data? And there are many cases where we have better data than Google Maps for things like gas stations because of all this truck data we collect. But to answer your question directly, like that is a huge portion of our business, is just improving and validating our data using AI. It's been, it's actually been super transformational for us. Giving access to our agents to scour the internet, collect more information, and then bring it back to us, and then use traditional ML techniques to make sure you're doing deterministic matching is like it's been a superpower. And it's how you can scale and say we can feel really confident that like we're gonna give you data on 36 million plus businesses, it's really high quality data. And it's why we've been able to maintain our accuracy. Because before it was me, literally me, going to a store locator and typing in addresses or my co-founders. Like that was what we were doing. And agents have been pretty transformational in that. But I think we've always, because we work with banks, we always sort of layer it in with deterministic traditional machine learning, like math-based approaches. And then you get this messy middle, and then humans step in and humans will validate it, humans will spend the time.

SPEAKER_00 26:35

Yeah. Okay, super interesting. So we know what it looks like when it comes to a bank. We know now how you transform it into something much more useful: a real name that's readable by a human, logo, address, et cetera. Um, I want to talk a little bit now about what is the value to banks of having this data transformed in such a way. Um, tell me some of the examples or or the ways that your customers then use this data that you're giving them to provide value back to their customers.

SPEAKER_02 27:05

Yeah. I think maybe also a good place to start with that is like Spain has grown a lot. We were a pure play data company for a really long time. Like, I think, you know, when we started the company in late 2021, when we raised our pre-seed, we were really focused on being like, okay, we're gonna be like just pure play data. And I think as we've grown, we've realized that like there's a lot of last mile problems of using data. If all I'm giving you is data and I'm unopinion about what you do with it, it can be very challenging to extract value. And that's why there's been over the last year, we've spent a lot of time expanding our product suite and getting deeper with our customers about saying, like, hey, if you if you took our data, what are you doing with that? And how can we better help you with that? And there's really been a transformation of from like a peer play data company into what we're sort of calling like a data and AI platform, like genuinely a place that customers who have massive amounts of payments data can take that asset and create value with it. And generally, we work across four primary verticals with our customers. We call it authorization. You can think like fraud and compliance. This is like includes things like people using our data in a fraud model to decide whether or not to authorize a transaction. It includes a corporate card who's saying, I only want to let my customer, like my end customer, spend at a specific business. Uh, it includes banks who have limitations about where their customers can spend because their customer is a child and they don't want their kids spending at a liquor store. Um, we also do a lot of work in attribution, like rewards and personalization. We help make sure that our customers' rewards are accurate. What's so funny is that most consumers don't know, but when you have that reward credit card, like you're required by law to get the rewards your bank is promising you. So if they don't, you should call them. We spend a lot of time with our banks to make sure if they're promising someone 5% cash back at Walmart, they're giving 5% cash back at Walmart. And that is like a shockingly sticky problem. And it's why if you go to your credit card, I'm not sure who you bank with, but like an Amex and they have that like digital coupon book where they click in card-linked offers. If you read the fine prints, like you have to shop one of these 17 Lululemons, you have to use this link. And that's not because they're trying to scam you. Like your bank genuinely probably wants to give that to you. It's because they can't ensure that if you did something other than this happy path, they're gonna give you the money back. And like that is that that's a huge area of what we do today. And then we're spending an immense amount of time in like analytics and AI. Like, I mean, that's just like a huge bucket of work for us today. And like I was skeptical when we were first building spade. I was like, this doesn't feel tangible, but obviously things have changed with the release of all the big AI models and stuff like that. But those are the big practices of work that we do. I'm happy to talk about any of them.

SPEAKER_00 29:37

And did you purposefully set them all up to begin with A? So you've got Auth attribution analytics.

SPEAKER_01 29:43

Um hundred percent.

SPEAKER_00 29:46

Okay, easy to remember. Um when you mentioned auth. So let's let's start with that one. So it makes a ton of sense. Your banks can then take this data and put it into fraud models, maybe feed it back into their transaction monitoring systems. It can enable them to do more targeted fraud or um money laundering detection. Um with all of this merchant data, you could probably provide a pretty good KYB product to know your business product. So as a bank wants to onboard a customer, is this customer who they say they are? Is that something that you do today, have thought about, have purposefully avoided?

SPEAKER_02 30:24

We've definitely thought about it. I mean, we get requests for it all the time. We have a very unique data asset. Uh, we work at the intersection of payments and businesses in a way that like you I usually have these two things very separated, which is very special. I think as you know, as a founder, you're you know that like you can't do everything and we've generally tried to stay away from KYB. I think we we absolutely have some products in the pipeline that will, if you are a person dealing with businesses uh and are interested in understanding how they engage with payments, like please come talk to us. We do have a lot of data about that and we're we're coming up with stuff like that, but it hasn't been a primary focus. We've really sort of tried to stay in the transaction lane for as long as we've could.

SPEAKER_00 31:02

Yeah, that's interesting. Almost feels like you should be selling that data back to the KYB providers, even yeah. Interesting.

SPEAKER_01 31:09

Exactly.

SPEAKER_00 31:10

Um and so then the attribution one is interesting too. So rewards, um, making sure that that credit card companies are able to determine exactly where it is that you spent money so they can give you your cash back. Um talk a little bit more about the rewards angle there. Is there also an element of you being able to provide the data on which these banks can actually build their propositions, not just have we done this thing correct, or we want to give you all the rewards for shopping at Walmart, but like what should your credit card offering actually contain because of where your customer base is shopping?

SPEAKER_02 31:44

Exactly. Um you're just selling spade for me. It's amazing. Um so yeah, I think one of the things is that we don't just like to work on the sort of that like I'm let's call it just more traditional program, like you know, 3% back on travel or a cash back into big merchant. We spend a lot of time with customers like Built, um, who have very unique value propositions, localized rewards, very small businesses. And really, what Spades data does is often open up that as a new vector of attack. Like banks know their customers really well. They know where you live, you spend a lot of time talking, they know, they know like where your job is if you're getting money from people sometimes. Like, are you employed? What kind of financial products do you have? But what they don't know is those details. And we really fill in those gaps so they can say, hey, I want to offer a really personalized reward. I want to offer a really personalized uh you know experience to someone. And I think you can see this in fintech companies like Cash App and others who have invested immensely in building out rewards programs that are informed by how you spend your money. Natasha has a persona as associated with the bank. She's this age, she lives in this city. What does that say about her? Where does she spend her money? Spade's dream is really to say we should own that end-to-end and say, great, like I can help you make a value proposition to Natasha that looks great. That says, like, she likes to do these activities because this is where she spends her money. If you want to convince her to spend more money on your card, you should be giving her rewards on that. You shouldn't be pushing her like 5% back at you know, Starbucks if she only ever drinks coffee at blue bottle. Like these types of things aren't possible without better data. And we can do things like only shopping within specific geographies. What part of your neighborhood, if you know where someone lives, can you say, hey, I'm gonna give you a reward for that coffee shop down the street that I know you love? Like there's these very special moments that I think financial institutions are interested in doing. Like, I think that's one of the things my favorite parts about working with banks is that I think we have this assumption in fintech that like there are these big stodgy institutions who like, you know, they what do they really want? They just want to like take our money and like we should be working with these fintechs, like amazing. I love our fintech customers. But whenever I talk to banks, it's not that they don't want to be doing these things, it's that they haven't been able to. And I think there's a real renaissance in banking technology that I'm so excited to be a part of where we're able to partner with our financial institutions and say, hey, like, how where should you be rewarding and how can spade be that delivery mechanism to unlock your ability to do it? And we have really, we have projects, we have AI products in that space that are really about helping them enact those rewards at scale without having to like manage all these merchants and things like that. So we do, we actively support that for our customers today.

SPEAKER_00 34:11

This is bringing a lot of uh different like uh threads that I've been hearing about and talking about recently together, which is really interesting. And um, I want to try and bring some of those together and see if you agree with it and if it makes sense to you. So I previously interviewed Joe Mancini um on a podcast on this podcast about bank MA and he was talking about how um really banks should be able to give customers the exact products that they need right then, right where they need them. So if you're spending a ton of money at a car shop because your car keeps breaking down, they should be able to come immediately to you and say, Hey, do you want a loan for a new car because it looks like you're spending a lot of money on your car and maybe it's time to replace it? And the other thread that I'm thinking about is um what the Stripe founders, the Collison brothers, recently said, which was that software previously has been built as this, you know, this thing that they build and everybody uses the same version of it and it's deployed to the cloud and everyone can access it, and that's great. But with AI, now we should be thinking about software development differently, which is everybody could have their own pizza. We should be making this for them at the time of purchase or at the time of need, and it can look exactly um like what you, customer, need versus what we think everybody in this generic space needs. And so stuff can be much more personalized using AI. So if we think about those two threads, the final piece of this puzzle to me is you know, we work with banks at cable, we see bank data, we understand the challenges that banks have in implementing new vendors and specifically in implementing any AI tools because the data is just so bad. And so it seems to me like you kind of come in underneath these other ideas, these threads, and you say, Hey bank, you've got this shitty data. We can make it brilliant. But not only is it good for you because it can help you understand how your customers are spending, but it also puts the data in such a format that you can now use AI. You can your data is so good that AI now works, and you can be providing the pizza-like experience for all of your customers, this personalized, as they need it, individual products to consumers. Is that a big part of the value prop at spade? Like we make your data AI ready.

SPEAKER_02 36:36

Absolutely. I think it's why I was when I talked about those four use cases. Like if you would ask me a couple of years ago, like, where would we, where would the biggest chunk of our business be coming from? I would have definitely told you authorization. I would have been like, that's like fraud is a big problem. There's always a problem with fraud. Like everyone wants to solve it. And it's true, it is always a problem. But we're had we're living through a moment right now where financial institutions are changing how they buy and what they buy. And I think spade is and will be a requirement for you to be a modernized financial institution. As you've gone, you've done the hard part, you've onboarded these millions of customers, you've told them to buy, you've convinced them to trust you with their money. Like you deserve a massive round of applause for that. But the future of that is saying, okay, now that we have these customers, how do we, how do we engage them? How do we give them a credible service? And so much of that is built on what you know about them. They've pulled their data out of the core. It is now sitting in a Snowflake database, and now it's terrible. And like they get to that moment, and you're saying, like, we should be partnering with an organization like Spade. We come in and we say, we're your partner, like we are a piece of your infrastructure. We're just like AWS. Don't think of us like a point solution. I'm not trying to sell you widgets, like, you know, weird, like your customer said 3% more on Netflix last month. Like that isn't helpful. And I think that was how a lot of people sold to banks before. We're coming in and saying, we offer you a centralized platform and foundation to build your anything your that touches your payments data should be better payments data and you should be building it on top of us. And for some banks, they want to just take that data and run with it. I mean, I mean, bank investment in IT, I think is like breaching$650 billion or something like that over the 2026. These financial institutions have massive teams. And in many cases, we act as a thought partner. We come in and say, hey, like this is how you should be using this. We're very opinionated about how do you display it, what types of models should you be training? And we can work with you there. And in other cases, the bank says, I want, I want to own my user experience, but I'm really struggling to take my data and turn it into a rewards program. I'm struggling to take my data and turn it into targeted underwriting or uh targeted marketing for loans. So we can be an actual uh sort of end-to-end platform where you say, Great, you have some stuff you want to build yourself on top of our data, amazing. This other piece that Spade is really good at, you can buy it off the shelf. It is cheaper, it is more effective, and it is instantly available to you and your customers because banks often have more money than time. Um and if they have an initiative, why wait two years for it and for their IT team to build it when they can buy it from Spade for a fraction of the cost and instantly deploy it across millions of customers. And like the vision of Spade is saying, you know, we're moving into a new era for financial institutions and they need these pieces of infrastructure that unlock their ability to execute and compete. Because I think the marketplace of financial institutions globally is only going to get more competitive with all of the AI we're seeing and the shifts we're seeing from even just like getting into the cloud. All it leads to is more competition.

SPEAKER_00 39:22

Yeah. Okay. So without necessarily getting into like exact numbers, how do you charge banks for this product? Is it the per transaction fee? Is there a platform fee? Like, how do you think about that?

SPEAKER_02 39:35

Yeah, we try to keep our pricing simple. Um I think we generally price across two vectors. One is the total number of payments volume we see, like total transactions. How many times are you calling out to us with new data? So that's one factor. And the other piece is like, what products of ours are you consuming? So generally, as your business scales, if you're just using us for enrichment, you'll probably just pay us for that. Like there will be like maybe a small platform fee associated with it. If you want like support and white labeled onboarding, and uh, that's great. And then as you scale, the platform fees will often get larger because you're using more of our products. Maybe you're reporting things to us, you're giving us feedback or you're having us investigate incorrect transactions. Maybe you're running your entire rewards program on us. And then we have actual more like traditional platform SaaS fees that really the goal is that it's like, hey, this atomic unit is your transactions. Like as you scale, we scale with you. But if you're gonna be buying more of our software, we want to sort of like sell you those widgets in an easy to consume way. We don't want you to be like, you know, like we don't want to be nickel and diming you. Just pay us once per transaction. And then here's the software that will sort of allow you to deploy these things much faster.

SPEAKER_00 40:39

Interesting. Okay. And so from a bank's perspective, when they think about like the financials of this and how it makes sense, they're giving you a little bit of their transaction revenue there. And in return, they're able to find better ways to target people who might need loans. They're getting better ways to build rewards or um or value for their customers. They're finding out more about the fraud that might exist in their data. So really, it's just like there's a tiny fee for this, the not only additional revenue streams that they can find, but also potentially like lowering their costs elsewhere.

SPEAKER_02 41:15

Absolutely. I think we we spend a lot of time with our financial institution customers to build an ROI case that makes a lot of sense. Like we need this to be value accretive. And if we're not helping you drive real business outcomes with better data, like we're failing, and that's not a happy customer. So our goal here is to say, like, okay, how are we gonna help you generate more revenue? Maybe it's because we're gonna help you target your loans better and then you're gonna, you know, expand your loan book. Or maybe it's because we're helping you reduce fraud losses, or maybe we're saving you engineering time because instead of having to clean all this version data yourself, you're deploying like an industry-leading app that drives better consumer behavior and things like that, interceding and just uh giving you one of the most obvious things is like, I'm gonna give you better data and fewer people will dispute transactions. Like that's MasterCard and Visa have these big studies about this. So, like, it's not just me saying this. Um, where there is evidence that like a huge portion of call volume for most banks is literally people confused.

SPEAKER_00 42:03

Yes. Being that's done that. Yep.

SPEAKER_02 42:05

Yeah, exactly. And it what's so funny about it is that people really tried to solve this using some of the older school products, but the problem is that these old school competitors of ours like worked really well on the big box stuff, but really failed in the long tail. And most confusion came from the long tail. So it's like, great, yeah, unnamed competitor does a really amazing job with target. Okay, that's fine. It also said target right there. The small mom and pop shock whose descriptor literally doesn't say their business name, but just has an address, is probably driving a un, like a much higher proportion of these call disputes and confusion. So if you're not really good at those edge cases, you're just missing the entire ROI case. And that's why, like when we invest, when banks invest in spade and invest in our partnership, like we see a lot of success when like we're measuring how many of our products you're using, how many verticals within your financial institution are actually touching this data, and how are we creating value for all of them? So you can make a very, very compelling ROI case versus like wedging in and saying we're just a fraud solution or we're just a reward solution. Like, no, we're infrastructure. Are you just gonna deploy data breaks in just one division or are you gonna deploy it across your entire organization?

SPEAKER_00 43:09

Interesting. Okay. Um, let's talk about this uh this awesome series B that you've just raised. So$40 million is a lot of money, a lot of capital you can deploy. Um tell me, I mean, everyone always says, you know, we're gonna hire some great people and we're gonna grow, blah, blah, blah. But tell me specifically with regards to your products, what are you excited about investing in now? And and given those areas, you know, the Auth Attribution Analytics and AI that your additional SaaS products kind of exist in. What are you gonna be using this money to really push into? And how do you think that the products that your customers use today are gonna be changing over the next two to three years?

SPEAKER_02 43:45

Yeah, I think like we are on a journey. We are going from a pure play data product into a software and data platform, and that is the biggest investment we're making. Our goal by the end of the year is to ensure that our customers aren't just engaging with our data via APIs, they have a Place they're going that gives them observability and access to their data that's never been possible before. So many of our customers, well, we have like we have QBRs with them where we take them, we're like show them all this data, and then we obviously really fun charts. You're like, hey, look at all these interesting places your customers spend their money. Here's the categories, here are the businesses, here's some patterns we're noticing. Just like, you know, our people are just like pulling this stuff together via like, you know, Looker or another BI tool. Our the banks are and our fintech customers, like, what? This is crazy. How are you seeing this? And what we realized was like, at the core of what our customers are looking for is not just sort of like actionable, like better data equals better product, better lending, et cetera, is observability and insight into how their customers engage with their financial system. And we can provide that incredibly easily. No more like, okay, you've onboarded this new like thing. You have to plug Redshift into your database from Snowflake and push it out to your BI tool. And it's gonna take three days to have it. It's like the beautiful part about our business is that by integrating with Spade, we store your data and we can make it available to you and generate those insights and drive that value for you in our own cloud, saving you immense amounts of money and also like increasing the value you can drive from our data. So I think a big push for us this year is actually moving from like a bunch of APIs into a place that people engage with. And we have a we have really big dreams of saying, okay, now we can really layer on our expertise in AI and say, great, do you want how do our customers want to engage with their data? Do they want to slice it in different ways? Do they want to pick out customer segments or spend patterns leveraging AI in natural language and make that available to their BI teams and replace some of these old school tools? Or like one of the things we see is like our customers are like, we want to create a rewards program. Why can't Spade just say, hey, you put it in natural language? Where do you want your customer to get rewarded? Describe what their behaviors are. Let us draw that box around their spend and give it back to you in the form of like, you just need a reward here. We've built a lot of these pieces individually. Like we have a, you know, we have a rewards attribution product that's live. It processes billions of dollars of rewards. We have internal analytics that we could make available. Now it's about taking all these disparate pieces and just plugging it into uh one platform our customers can engage with. And that's really like from a product perspective, is is where we're going. And then, you know, what better way to sort of cap it all off than with AI?

SPEAKER_00 46:13

What a one of the things that I think is a constant challenge as a founder is figuring out where to say no, especially with regards to like product breath. Um, as you think about that, you've got all you know,$40 million in the bank that like one way to think about that is cool, let's do everything. We now have the money to do that. Um, but that is also a pretty risky strategy. So what is it that you are specifically saying no to now?

SPEAKER_02 46:39

I think it's these ancillary use cases. Like we like KYB. I love that. I think that's amazing. Like, wouldn't it be such a cool way to use our data to find this ancillary use case of a new subtype of business I can sell to? Like, tell me that doesn't sound amazing on paper. It does. I have absolutely no idea how to sell or build the most compelling data product for KYB businesses. So I think where we're saying no to today is saying we're we're really focused in that transactional moment. All of those ancillary workflows that come off of it at a financial institution and a fintech are fair game for us. Obviously, there's a stack rank, like we know where we're gonna focus. But what we're saying no to today is like, I think there would have been a time where if a KYB company showed up at our door, I would be like, okay, how can we, how can we really sprint on this? How can we build an endpoint that lets them do this? Like, we're trying to say no to those types of opportunities and really stay focused. And like, do they generate first party data? Like, do they have a financial product that generates this data themselves? Great. That is our bread and butter, and that's who we need to focus on. And we're really focused on saying, if they have a first, if they have first party data and they're generating it, how can we give them that full horizontal view and help solve them, solve their problems?

SPEAKER_00 47:52

Ah, so there's actually like a flywheel here in that the products that you want to build and give back to banks are the products that help them generate more transactions to increase the transaction data that you hold. Whereas KYB would enable them to like onboard a customer, but it doesn't immediately say, and now there's more transactions. Whereas if it's like, here's a rewards card that your customers go use, they're spending money to get the rewards, which generates more transaction data.

SPEAKER_01 48:20

Exactly.

SPEAKER_00 48:21

That makes a lot of sense. Okay, that's really interesting. Um one of the things that I see like pretty much every day is banks like really struggling to figure out where to use AI, how to use AI, um, how to talk about AI internally, even like what is the return on investment? What is the ROI that they can get from AI? Um do you see yourselves being pulled into kind of like consulting with these banks? Um, are they coming to you and saying, I've now got all this amazing data from you? How the hell do I use it?

SPEAKER_02 48:59

Yes. Uh absolutely I think the way I like to think about it is like who better to buy AI services from than the person who already stores your data. Like I think there's a real compelling story as why like AWS and, you know, um, you know, Amazon, Google, et cetera, started deploying models in cloud, is it makes total sense, right? Like we already have all of our data sitting in AWS. It's very easy for us to throw Clot or you know another one of these coding tools there. I think there's a similar pattern with banks. I think banks are super interesting. I think there's always going to be a portion of them who want to do all this stuff themselves. They will sign deals, huge deals with the big, big model providers and they're gonna look a lot like a tech company. In those cases, we can be, we are experts in payments data. If you're gonna be plugging payments data into a model, like please come to us. Like, we'll make sure your models are more performant. We're making sure that if you have a chat bot that's calling it, that's grabbing payments data and someone's like, So where did I spend my money last month? It's not considering that Amazon is a bookstore. It's deterministic. The data it's getting is consistent. We improve those performance. We save you money because it's not consuming tokens and like searching out to try to figure out what is Amazon, like what is the concept of Amazon? Um and we definitely work in a consultative manner. And then there's people who don't, there's people who want the power of AI without the headache of AI, which is fair. Like these are expensive at scale, these are very expensive pieces of software. And it's only probably going to get maybe a little bit more expensive in the future. And I think we do a lot of work where it's saying, like, how can you take the magic of AI, like natural language to X, like semantic information to a rewards program? How can you deploy that in a place that is secure, it's compliant, and it's ingestable by a financial institution in a way that's valuable? And like we like to walk that line because we, you know, I think I want to empower our customers to solve these problems themselves. And in cases they can't, we want to step in. So we do a lot of very direct consultative services to our financial institution customers to say, like, how can you best make the use of this data? If that's with us, great. If you want to buy a product, amazing. But if not, like of course we'll be there to talk through your strategy around this. So it's a long-winded way of saying yes. We try not to, I don't, I don't, it's funny, you don't want to get into the business of like billing people for these things. So I think you know, you want to build great partnerships, but there's definitely a there's a whole host of people I think running around trying to help people do this today.

SPEAKER_00 51:14

Yeah, for sure. Okay. Interesting. The final question I have then is you had immense growth last year, nearly 500% year over year growth, which is truly amazing. Um when you grow that fast, stuff just has to break. And so is there a good example you can tell us about something that broke over the last year that you fixed and and sort of how it's improved now subsequently?

SPEAKER_02 51:39

Like everything kind of breaks, I think. It's all very, very painful. Um I think one of the things that we've experienced is we went from processing tens of millions of API calls to billions. And thankfully, we didn't have any sort of moment of doubt. Like literally, I think we had 100% uptime last year. I think I might get in trouble, but I think it's somewhere around that. Like just absolutely five nines. It's north of five nines. We just like that's been something we've always had a very strong hold on is like uptime and availability because we are in the critical path for so many of our customers. Like, we don't go down. That's like our promise. We are infrastructure, it doesn't happen. Um but one of the things we experienced was like, how do we how do we maintain like when you have 10 customers, you know, all of them and you have these really tight relationships and you can maintain that customer success. And I think as we scaled, it became harder and harder to maintain that strong customer relationship and your champions who love you as we scaled, and it absolutely broke. Like you, you go from having 10 customers who get equal attention to a stack rank of customers based on the size of their relationship with you and all these things. And I think there was some pain in there. There was moments in time where like I regret I, you know, people are working too many hours, they're chasing too many things, you're not getting back to people as quickly as possible. And I think that that was a huge learning for us is like, how do we make this scalable? Like we did no, it was all managed via Slack and emails. And so we bought and bought software, we onboarded to vendors, we like hired people who are gonna just do that all the time. And I think that was one of the things that I don't know if I I realized how unscalable it was to have like such a white glove relationship with our customers. And when you're infrastructure and you don't do that, it's incredibly challenging because that slows down time to revenue. It slows down adoption, it makes it harder to get penetration across use cases. So, like that was an area where we had to invest heavily, and like we're still growing up in that sense. Like we just hired our first uh customer success person, literally, like I think he started on Monday.

SPEAKER_00 53:33

So um he's never and before that it was just a mixture of you and other founders and everyone just oh wow, okay.

SPEAKER_01 53:41

Just sprinting on it.

SPEAKER_00 53:42

That feels like a particularly um a particularly painful point when you're selling to banks because banks do sort of have this like they just have this feeling of like you should be breaking bread with your bank customers. Bank is like they want to go out for lunch, they want to meet in person, they want to shake hands. And obviously, like there are 5,000 something just under banks in America. So you obviously can't do that when you start growing significantly. But yeah, there's this like real bankers want to eat lunch together type thing that you have to try to maintain. Yeah.

SPEAKER_02 54:16

100%. I mean, I'm lucky to have a co-founder. She was a McKinsey consultant. And you know, the level of like the execution on an onboarding process, the Gantt charts, the fli every like just like the amount of work it was to get a customer into the door. Like you sign the deal, now the real work starts. And like thankfully, I mean, she's just an absolute like insane operator in those types of environments. And like we're really lucky because it meant that we could all play different roles. Like sometimes we joke that like last year, our head of GoToMarket Lauren, um, I was like her solutions engineer, like answering the product questions, Tess was her account manager, and we were all just kind of like, but that was it. Like there was no more. And I think it's just like it you can't do it anymore. And I think you we're trying to find ways to help at scale with software, with AI, but then also just like hiring headcount to make it possible. Because you're totally right. Like they want to break bread. And I think that's how you build enduring relationships. These are institutions that are built on trust. They they hopefully will be there forever. Now that's their goal. So if we want to be a vendor for them, we need to give them the sense that like we are inevitable and we're gonna be around even though things are like, you know, we're a little small today. We'll be around though.

SPEAKER_00 55:21

Yeah. Awesome. Well, but this was fascinating. Congratulations again on the Series B. And uh thank you for coming and chatting with me about it. Um, everyone listening, if you enjoyed this, please share it with your friends and colleagues. Subscribe on Apple Podcasts or Spotify and see you next week for another episode of In Control.