Resources > Data Jam Sessions and Podcasts > Data Jam Session #3: How to Leverage Your Transaction Data in the Data Cloud

Data Jam Session #3: How to Leverage Your Transaction Data in the Data Cloud

March 24, 2022 by Segmint


Click above to watch the entire Data Jam Session!

Session Transcript Available Below:

Thu, 3/24 12PM • 58:55
Kent Blacksher, Mark Leher, Lance Cuthbert and John Thuma.

Kent Blacksher  00:03
Well, hello again, everyone. Ken Blacksher here hosting yet another Data Jam Session. And I couldn't be more excited for the panel we've assembled for you all today. Joining us to talk about, well, the ever shifting, kind of roll in the importance of the role of cleansing transactional data, which we've seen so much in the data landscape over the last few years and much of what we've covered as it pertains to our particular vertical. In the last few data jam sessions, we thought it'd be really important to talk about, well transaction cleansing, cleansing of merchant data and all sorts of that area as it pertains to early the AI space and things like that. So we've assembled a really a crack team of much smarter individuals than I on this session. So without any further ado, let me introduce them. Joining us from Snowflake is their senior solutions engineer John Thuma. John, a pleasure to have you today.

John Thuma 00:57
Hey, nice to meet you and see you. You know, I'm really excited to be here and talk about data and analytics, especially in the banking financial service sector. I did bring my work boots today. So we're ready to go. Um, you know, I've got about 30 years in this business. So I go way back, SQL Server four dot x, I'm way back when it was still a Sybase product really. And, you know, I've traveled throughout the whole data universe, when you when it comes to walking almanac of different historical uselessness, that really can really shine a light on where we are today. And it's exciting. Glad to be here. Glad to be working with you.

Kent Blacksher  01:31
I'm delighted. And for those of you who haven't met John, now, you know why I'm so fired up about this panel, because he's this awesome. Also joining us our own VP of data and analytics, the one the only Mark Leher, Mark always a pleasure.

Mark Leher  01:47
Thanks, Kent, I'm even more excited than you and John combined for this session, been looking at transaction data for 13 years. And so it's a topic that I love to talk about and looking forward to share with other people besides my wife and kids today.

Kent Blacksher  02:03
I love it. I love it and also joining us. And honestly, someone I've worked with for the better part of nine years now. So that's more of a testament to his inner strength than anything else, that he's still around our own VP of Enterprise Architecture. Lance Cuthbert Lance, what can I say? Just so excited to have you today.

Lance Cuthbert 02:20
Thank you, Ketn. And I'm really happy to be here. Yes, it's been nine years that you and I have been working together so far. So far back, when we started, I think we both had hair.

Kent Blacksher  02:32
Your words are coming off a little aggressive today. I'll take that and run with it. And you're accurate. Well, without further ado, let's get started. I want to kick it off by I think I'll point to you virtually John, if if you can, can you kind of set the scene for where your thoughts are on the industry right now. Mark and Lance, when he's done I'd certainly love your take as well. Just give us that that overview of what you think is, is kind of the lay of the land right now. And where do you think things will continue to go?

John Thuma 03:04
Yeah, absolutely. I mean, first of all, you know, the state of banking today and retail banking, and, you know, core banking, checking account savings accounts, you know, retirement accounts, all those things. They're up for massive disruption. Over the next 10 years, you won't even recognize what banks look like today, gone will be the physical side of banking, right? Things like, you know, you got big tech coming into. So these things are big, right? So you're going to be able to pay and be able to we already are able to pay in cash or cheque with these things. In fact, one major FinTech company just did a study said 41% of the of the people have changed the way they've interacted with their bank, they're no longer going into the bank, they're trusting the applications, the mobile apps, they're trusting the websites, right? And that's a multi-generational thing that's not just isolated to the younger generation, everybody did. And there's reasons for that COVID disrupted everything, and we're never going back. Right. And before COVID. So there's really three things before COVID - The data, the data, you know, revolution occurred, right, big data and data science and all this stuff, all this banter. But you know, what really caused the catalyst was COVID-19. Right? And now that we're coming out of COVID, and we're in a post COVID world, there's things that are really affecting us, these things are going to go away, I think, and not to mention, they're going away already with the inflation that we're dealing with. Right? So and these things too, right? These things, these credit cards and plastic things, the physical things, that brick and mortars are going away, and they were before COVID. Right. And that just accelerated the whole thing. And I think and this is my big, I'm gonna be very provocative here. So everybody put your seatbelts on and put your helmets on might upset some people with what I'm going to say. ATMs will be the payphones in the 1990s. They're going to go away. So you know, gone are physical things. And we need to, really, to one big thing, and if I have a concept of what we're going to talk about today is data is Your biggest asset period. And that's what we're about. And that's why Segmint and Snowflake working together are so critical and so important to the future of banking, because startups are coming. Big tech is here. Right. And, you know, it's going to be a different universe in 10 years. So that's my opinion.

Kent Blacksher  05:19
I love it. And that's, that's why you're here, and definitely thought provoking. But in line with what we're seeing right in line with what we're seeing, Lance, because you're next on my screen, I'm gonna roll to you for kind of your thoughts, as it pertains to what you're seeing, and what do you see in the industry right now?

Lance Cuthbert 05:36
Well, I think, you know, John touched on it a little bit is the the trust that customers are starting to have with having their data out there. For these types of analytics for this type of data analysis. You know, being a, as an information security officer in a past life, these kinds of things always scared us, because we didn't want access people accessing our customer data, we were always concerned about making sure we had high walls around all of that data. And it was great from a data security standpoint, but it really limited the usage and the usefulness of that data. And I think as we are progressing with technologies and the ability of the technologies to provide the accessibility to the data, but also keep the security of that data intact, while you're doing that is is really huge, is really huge. And I think, you know, just that trust that people are developing. And I think part of that trust is coming through regulatory acts as well to things like the CCPA GDPR, over in Europe, you know, this is showing that and making institutions conform to security and not just be the wild wild west with customers data, they're no good.

Kent Blacksher  07:05
Certainly makes perfect sense. Mark, your thoughts? You know, you're totally on the data side as well. Where do you see the industry kind of right now.

Mark Leher  07:18
And I would echo almost everything John said he had more props than I do. But we see exactly the same thing. And the data that competitive pressure from from the fintechs that banks and credit unions are facing right now is incredible. I've got a team of six library scientists, we're looking at transaction data on a daily basis, and they're pulling out new Fintechs all the time, by now pay later is new. But that's like a year old example, at this point. There, that's just one of many that are coming in waves that banks and credit unions have to respond to. And so in, in an atmosphere where decision making just things change too quickly. Today you can't sit around and and pontificate and you have to use data to make decisions. I like to use professional sports as an analogy, probably an industry that was a little bit ahead of the curve. I was an economics major candle player shortsword and 2003 working on in an in an honors thesis or something. And my professor gave me the worst advice ever. He said Don't, don't no one do any more baseball paper, because we're done with baseball. That's all been done. This is in 2003 Turn on a baseball game today, and you have four infielders, on the right hand side. And, you know, the other side's totally empty, because they've embraced analytics. And they know that, that, you know, pretty likely that that batter against this pitcher, if you put the ball in a certain spot, it's going to ground onto the right ground out to the right side. And so that's where they're going to.

Kent Blacksher  08:54
Exactly right.

Mark Leher  08:56
And the teams that are embraced that you can see it in the results, they're doing better. And so other industries, and I'll bring that to fintechs, need to and they're starting to embrace the same thing, you got to look at the data.

Kent Blacksher  09:11
And you that's a perfect analogy. I love that. That baseball example. You were talking about cleansing transaction data. And I wanted to point to Lance, because this is an area Lance spends a lot of time as well. When you think of that and you think of like merchant payments and so forth. Can you kind of just level set? What goes into that process of cleansing? And then Mark, I'll come to you in and around why you think it's so important and what are we seeing others do with cleanse data? But Lance? Can you set the scene for the process what it takes for those teams to get into that?

Lance Cuthbert 09:48
Sure, sure. And it's a it's not an easy area to start dipping your toes into that water. It's it's very difficult. And you know, it all starts with and it seems almost too simple, but just getting the data, how do we get the data? And what are the mechanisms of the of receiving that data? Is it via an API? Is it via a batch type of process? Okay, so you need somebody, a provider that can, you know, fit your needs as to how you can provide the data to them for processing. The other thing is also the type of data, you know what, what kind of transactions are we going to be processing here? Is it just going to be debit and credit card swipes, or are we going to be doing ACH bill pay recurring payments, you know, so trying to cover that whole scope of the different types of transactions that we could glean important information from if only those transactions were cleaned and categorized, then you kind of get into the actual meat of the matter, which is starting with the data cleansing. We've all seen those strings that are on our debit card or credit card statements, and they're, they're very verbose, sometimes they're very ugly, they have a lot of strange characters and numbers. And, you know, you almost need a doctorate from MIT to understand what they are. And that's where we start cleaning things out of those transactions, trying to get down to the to the root of the transaction, removing a lot of the, the noise from that transaction. Once we get down to that, then then you can actually start doing some magic, right you can go against, and what Segmint uses is a human supervised, and human developed taxonomy of all of these different merchants. So we, you know, we currently maintain a taxonomy that has over 170 million variants, and that allows us to do that matching. And matching is great, but you also have to maintain the granularity of that matching as well, too. So if we see a transaction for Uber, four or five years ago, that would have been fine, because it was Uber ride share. But now you have Uber ride share, you have Uber Eats, you have Uber freight, so you want to be able to get down into, you know, the, the granularity of that of that particular merchant, and then also being able to provide categories, you know, the analysis that someone may want to do, and that data isn't going to be just at the level of Uber, they may be interested in people that are using ride sharing services, or whatever. But using those categorizations allows you to expand your analysis of that data. All of this is great, um, but it has to be accurate. So how do you make sure that it's the results that you're giving back are accurate. And, you know, Segmint accomplishes that by we're constantly monitoring those accuracy rates, we're constantly monitoring to make sure that we're maintaining that the level of accuracy that our customers demand. And, you know, we do that with testing with large datasets, statistical samples, and it's just an ongoing process to make sure that we're always you know, hitting it at the highest accuracy. And then it kind of gets down to also the speed, scale, and security. You know, the old joke used to be it can be fast, it can scale, and it can be secure. Now pick to write because you couldn't have all three in the old world, right. But with some of these new platforms, these new technologies with the advent of Snowflake, you now don't have to make that decision. You can be scalable, it can be fast, and it can be secure.

Kent Blacksher  13:56
That you nailed it. Right. You nailed it. Mark, I want to point to you is particularly on the taxonomy side. And then John, I'm going to get you lined up just behind that. But, Mark, can you talk a little bit more about what Lance was framing up? As it pertains to the more of the use cases we're seeing? What are institutions doing with this? Are they taking this cleanse data for just analytics? Are they turning it into warehouses? How what's the play for the typical institution?

Mark Leher  14:25
Sure. So everything Landsat is exactly spot on. And once you go through that process, then it really is about the use cases, what are you going to? What are you going to do with all this great cleanse, categorize contextualized data? So we were seeing these change on new ones pop up all the time, but if either we're talking about their particular interest to banks and credit unions, there's a lot around financial wellness. So how do you understand the challenges or the new opportunities? Another way to think Given that customers or members are going through, cryptocurrency is a great example. That's something that it's been in the headlines for the last six, seven years, that bitcoins worth $40,000. Today, still kind of fringy. But we were seeing in the data when we cleanse, we're cleansing and identifying cryptocurrency exchanges, where people are putting dollars into buy those cryptocurrencies, we're saying two to 3% of customers at any given institution are participating in those exchanges. And it correlates quite well. With with the price of Bitcoin, which everyone knows, is pretty volatile. So that's interesting as an alternative investment space are, from a financial wellness perspective, are those dollars that a bank can bring under management? Or is there just advice around taxes or risk for diversification that an investment arm of an institution can bring in those cryptocurrency investors. So that that's one great use case, buy now pay later we already spoke about, that's something that we continue continually seeing growing over the last 18 months, pretty reliably. That's another kind of financial wellness. Free Up to now I think this is going to change those are, that's debt, that's consumer debt, really, that's not showing up on credit reports. And it's really easy for people to get I can, you know, I'll use a prop here, can I'm gonna own this mug, and three more months after I make the payments. And that's how easy it is anything you buy online, you can split into four payments, and it's almost mindless to do. But consumers can get themselves into trouble. There's too much of that debt. So financial wellness is one theme. And then you can dig into the data as well to figure out how to make How To Make Money is one one use case we're pretty excited about what's resonating with our customers is looking at subscription payments that customers are making. That's actually that's two prong that ties back to financial wellness, or people do they have a lot of subscription payments that they've just forgotten about, and can be notified to the bank, the cleansing can help identify those. But also when we look at subscriptions and look at the payment type, people are paying paying those subscriptions via ACH, that's an opportunity for an institution to reach out and say, Hey, switch those payments over to a debit card and get interchange revenue based on that. Because it's the subscription that represents an annuity for that situation. So not only financial wellness there, but again, an opportunity to make more money. And then the last thing I kinda like to tie in there is we can look at things like philanthropic giving sound, what's the opportunity there, let's really let's let's know what's important to my customers. So when I go try to get them to change their fitness membership from ACH to a to a debit card, maybe I offer I'm gonna make a $5 donation to a charity that's important to them. So it can be tactical, but it also gives again, tells you what's important about your customers. That's just scratching the surface. And we're looking at millions of transactions flowing through an institution. It's it is an extremely rich data set on which institutions can pull those use cases.

Kent Blacksher  18:17
Yep, I believe me, I know, I see this in my day to day when I'm talking to institutions who want to talk about better ways to leverage data. Absolutely. John, I want to I want to switch gears and get over to you. I know, we had seen the announcement around, you know, Snowflake and Segmint and working together. But I want to lean on your expertise here. What were you kind of looking to accomplish or solve with this partnership? And I have a sub question for you. If I recall, and I haven't heard this, but I was told about a little birdie have an interesting story about why you thought specifically of of our organization around that concept, but the floor is yours.

John Thuma 18:57
No, absolutely. I mean, you know, Snowflake is is all about data and unlimited scale of data as well as unlimited compute of data and being able to segregate compute from Compute, that means I can load data while the Segmint data processes that and cleanses the merchant, the merchant disruptions. And it doesn't disrupt my business intelligence users. It doesn't disrupt my data scientist, we have an environment that scales just the other day, I was working with a customer and we moved him over from Hadoop. I won't, you know, disparage any my competition. But we moved on from a flavor of a dupe into into Snowflake, and we released it out to the public, to the user community. And the feedback was just wow, where's this band? Right the these queries used to take eight hours to run now they run an eight minutes they run eight seconds. They even said this is hilarious. Unhappy Carl isn't unhappy anymore. He's He's happy right now. So when you have a happy Carl, that's a happy data user. That's amazing. So the point of this is, is how do I exploit data? And part of that is exactly what Segmint does. They cleanse, they, they help cleanse data, using a very difficult process that I've built in other organizations and other lifetimes. You know, we've done this merchant cleansing business, and it's not something that's one and done, you don't finish and everything goes away. They have years and years and years, probably a lot of people years invested in doing this. And that's what it takes, you are never done with what Segmint offers, you know, merchant cleansing lifestyle, he lifestyle indication, things like that take a constant set of data scientists and library scientists to look at and monitor. And from my perspective, this partnership, you bring together the Cloud Platform, the cloud data platform that, you know, supports data engineering, data, Lake data warehouse, data science, and of course, our app marketplace, which is coming out. And I'm excited about this partnership, because they're going to be one of the first Segmint is going to be one of the first or near first, you know, partners to bring a solution to the app marketplace. And we already have a data marketplace. And so this is, you know, this is an exciting time. And what it does is as we come out of the COVID-19, and what I call the post COVID era, we are entering into a new era, and that's inflation, high gas prices, knowing your customer knowing how their behaviors change, is going to manifest in your ability to have a better conversation with your customer. And that's what Segmint does. So as gas prices and inflationary pressures start to affect all of us. How does that affect the way we spend money? Are we buttoning down? Are we spending smarter? Are we not? You know, vacationing? Are we not buying new cars, or we're buying used cars? What are we doing to manifest these behaviors more intelligently, and it's it's the unlimited scale, the unlimited set of compute, and in solutions like Segmint that allow us to really understand what the voice of the customer now is, because the customer is not coming into the bank anymore. So you know, gone are the days where I walked in the bank, I opened the door, I walked in the bank with my wife, and maybe my little son, I walk in to see to the right there's Betty, she's my banker, I've known her for 10 years, I go up to Betty a and she did my mortgage, she did my house, she set up my 529 for my kids, education, she knows me, you know, and I walk up to the to the teller I have that one on one interaction, those things are gone. And their luck. Now today, I use these and I already pointed at one of these. And I'm using this and all that. Now my conversation with the bank is all electronic. It's all data. And so I have to mine that data in order to have better conversations, marketing conversations, know your customer conversations. And it's a better conversation, because you actually are manifesting it based on tastes and qualities that you can refine from their data pattern. And that is really, what Segmint offers me at least my understanding I've known Segmint for quite some time. Now. I've known them for five plus years now. And they've you know, they're a great group of people. And they're very smart. And I've challenged them, I pushed them, but they've always come through. And, you know, I've used them both as a customer and as a partner. And I'm just so excited to see that they're coming and they're going to be one of the first apps on our app marketplace.

Kent Blacksher  23:42
Excited to hear that I do have to ask, what was the what was the story? What was the story about how you thought a Segmint and by the way, appreciate the the kind words more than you know?

John Thuma 23:53
Yeah, I mean, really, the story is, is exactly what I said you walk into your bank, do I really know what my customer wants? Do does. Does my customer really want a HELOC loan? Or do they really want a mortgage? No, they want a better house. They want to they want to add on to their house. They want to go to Home Depot, they want to buy some stuff, maybe they want to contract out I mean, being able to understand that allows me to do that. But you know the other thing is my next best action. My next best action may not be a new product. It may not be a new mortgage, it may not be a new credit card. It may not be that it may be we know John Thuma enjoys steak and wine. So let's give him in his interaction with us. Maybe it comes through the mobile app or comes through this web page when he logs in. It comes in the form of Hey John, we know you love steak, try this recipe and pair it with this one. Right and off you go so you can still manifest relationships because of the things that Segmint offers and key lifestyle index. as well as merchant cleansing, we know a lot about the next best action that we can give. And that improves our marketing outcomes or ROI. And now with inflation and potentially stagflation coming up, how is that going to manifest and changes in behaviors? And how do we react to that as a as an institution?

Mark Leher  25:19
I love that. All right, and I'm gonna jump in, but mistake examples, great. What consumers I think, really crave today's experience, and banks and credit unions have the advantage of all this data. And it's no accident that the other thing, a lot of Fintec out there are coming with free services, like Credit Karma, get a free credit score, but link your accounts and it's because they want the data so that they can deliver these types of things. But John, you're the nail on the head institutions have to be having these, these digital convert engaging in the digital conversation. Every time a consumer swipes their card or pays a mortgage payment? Or does any transaction that you're getting a record? They're telling you something? And are you going to listen. And it's you sure it's

Kent Blacksher  26:09
A real world example. And this ties back to kind of what Lance was talking to when we're thinking about things like, why gets information value, right? Sometimes it's merchant that is important to an institution. But other times it's the category. For example, if if Lance has routinely been paying his mobile phone bill at Verizon, through the ACH or whatever, at an institution, and suddenly see, they see that payment is gone, they may be alarmed, oh, we're losing primary foi status, or Lance is going through some sort of financial hardship, when the category of just mobile phone payments would have picked up the fact that he just switched from Verizon, to at&t. Right. So again, just piggybacking on what you're all saying, sometimes, the cleansing piece, it's the merchant, that becomes important, but sometimes information values, a little more broad than that, kind of piggybacking on what John was saying about, you know, knowing the actual the institution needing to know the individual, much better than than they often currently do. I do want to shift shift gears to a term that we've heard a lot about, and I'll start with you, John, about the data cloud. You know, obviously, with Snowflake, we, you know, what is this? Why is this different than say an AWS or Azure or Google Cloud? What are some of the misconceptions people have about this space? And what to kind of look like in the future?

John Thuma 27:36
No, absolutely. I mean, what is the data cloud? I mean, Snowflake is is the data cloud, right? The easy button, what I call the easy on ramp to manifesting data exploitation for your organization in your institution, right, continue to use the tools you love. We don't want to disrupt everyone keep using Tableau with Snowflake. Keep using talent, keep using elation, keep using Informatica keep using Power BI, we have a set of JDBC ODBC drivers that you don't have to abandon the products you love. The other big thing, use the language as you know, right? Now, of course, we support NC SQL. So every almost every institution has an NC SQL person have small Tibet, right? But now we have through our snowpark implementation, Python abilities, right? The the ability to your everybody moves their data to the analytics or the data science platform, Gone are those days with snow Park, what you can do is move the analytic the algorithm down to the data and do it at scale, right and do it a mass scale. And that that means you don't have to sample right, and you don't have to move the data. So it's governed, it's secure. It's there. We also support Scala and Java. So use the languages you know, in house and Python and Scala are two of the most popular data engineering and data science languages out there. And this enables you to modernize your stack, right? We support we have massive, massive partnership with Anaconda. So that means you can use Jupyter notebooks to interface with with Python and Kafka connectors. We also have relationships with like data robot data, IKU, h2o and, and sage maker. And they actually push their their processing down into Snowflake where they can manifest out, you know, Advanced Data Science techniques down into the environment. And of course, we work across the same experience across AWS, Azure and Google Cloud, right. And you can also you can also failover you know, you can move you can replicate your data from AWS into Azure, if you want to and failover to Azure, and, and also redirect all your client interfaces. And finally, across all of your data, back in November, we launched our unstructured data capabilities. And we have some excellent, excellent semi structured capabilities as well through JSON and XML and whatnot, but our JSON parsing and our array processing is fantastic. So, and of course, are semi or fully structured data data that nicely fits into rows and columns like Excel. And then I think the big one of the biggest things that people don't know about Snowflake is that we have a data marketplace and an excellent data sharing ecosystem. And it's exploding, it's exploding. So you know how you have to move a lot of data, like, let's say you wanted to, you know, move terabyte of data, it's gonna take a while to download, you have to FTP it, if you do all that. And then you have to, you have to know how to load it into your database tables, it takes a special kind of skill set and infrastructure that can support that, let alone a network that can support that, that download, well, all that goes away with our data exchange in our data marketplace. And that's exciting. And then now, with Segmint being one of the first apps that's going to be available in our app marketplace, that really, really, really defines what the data cloud is. For me, the data cloud is a single pane of glass, which is Snowflake that supports data engineering, data, Lake warehousing, data science, and applications and of course, our data sharing aspect. Does that help demystify some of that for you?

Kent Blacksher  31:08
Absolutely. What? What would you say? Uh, for the viewers right now, who maybe aren't in this space? What are the misconceptions? Like, you did a great job laying out what it is, but what are the misconceptions people have before they take that step? If you will?

John Thuma 31:26
Yeah, I think, you know, fear, their fear, they're afraid, right? They, you know, if I move to Snowflake, what kind of transition is it? Well, if you're familiar with NC SQL, and SQL is a defined standard that is supported by you know, you know, SQL Server and Oracle and mateesah and Teradata and things like that. If you know that language, you're going to be right at home with us. There's always constant. There's a capability to manage your cost of spending compute, you know, there's there's guys there's guardrails, we can throw up to make sure that you know, compute doesn't run forever, that spend doesn't run forever. There's there's things that we can do to manifest and control costs, right? And I'll be more than happy to walk anyone through those things. I mean, there's, there's all kinds of guardrails. But I think the biggest thing is, is that, you know, is the flexibility. The neat thing about Snowflake is that every week or every couple of weeks, we're launching new tech, new capabilities into the into the product. You know, Scala just got GA, unstructured data got GA, the cool thing is that just naturally flows into the environment. And the customers can take advantage of that immediately. And there's no bill, there's no extra bill for that. And it's very, very simple to use us not to mention, no more guessing at hardware, or even virtual hardware virtual servers with us. If I want to expand a cluster, I can, it can do it automatically. And it can shrink automatically. I can also write something that can blow a cluster up if I want to and horizontally and vertically scale that cluster. So and it's really, it's really a powerful demo. And I welcome anyone who wants to sit down and go through it with me.

Kent Blacksher  33:13
Awesome. Absolutely. I'm hopeful people will be taking you up on that because it is very, very slick. I want to shift gears kind of go back a little ways to something you were talking about Lance, you had mentioned briefly, things like GDPR, and aspects around that as it pertains to, well, things like data privacy, we've been talking a lot about the cleansing of data, talking a little bit about some of the use cases and why the solutions are so powerful. And then John and Mark and YouTube, Lance have done a great job of setting the scene for what's coming next and why everyone needs to be paying attention to this, particularly in banking, if I space. But can you spend some time talking about data privacy? Collectively, how it works with something like Snowflake, right? I mean, I know there's a lot to cover. I'm not asking you to take, you know, 30 minutes, I promise, but can you give us kind of a? What are you seeing? What are the best practices? And how does this all kind of meld together? And Mark, I'll cue you up kind of right behind that would love your feedback as well.

Lance Cuthbert 34:17
Sure. And I think the really important thing to note is everything that John just laid out in relationship to the features and functionality of Snowflake is all done under an umbrella of security. One of the things that drew us to Snowflake is their attention to the different regulatory and you know the different levels of security that you have to have to to protect your customers data, you know, with Snowflake being PCI compliant sock two compliant, high trust compliant, ISO compliant, and I know there's a couple other ones John, so forgive me, I can't remember. But you got it. You got a lot Yeah, and so you know, that, again, helped us make that decision to move to Snowflake, because we knew that our customers data would be secure, inherently secure. You know, with our environment in at at Segmint, we've taken great pains to, to make sure everything is secure from end to end, with Snowflake, you pretty much get that right out of the box, which is a wonderful thing. And then on top of, you know, that regulatory compliance, adherence, you also have additional features from a security standpoint where you can actually put automatic masking of data. So as you're ingesting data, and if the system identifies that, hey, this could be a social security number, or this could be a, an account number, it will automatically mask it, as you're bringing that data in. So certain people can see unmasked, certain people can only see it has mass data. So that definitely, you know, helps with the overall security of the data that you're storing the data at rest and, and data movement, it's all done over a secure channel. So the data again, and security with that, which is really huge for especially for a company like us, who's dealing with customer data. Now, granted, we don't take PII, but this ensures the security of that data. Another big piece of the that John had mentioned, that also, you know, lends to the security is the whole data sharing ecosystem that Snowflake gives to you. So what's really exciting about this, and how it could tie into merchant payment cleansing for us is that, you know, as John was saying, and you know, in current situations, you have to generate a CSV extract, and put that on disk somewhere. And then you have to send that via some sort of network connection, whether it be FTP or secure FTP, there's a whole lot of opportunity for something bad to happen with that entire workflow, you know, between data being corrupted, or data being seen, a data breach, you know, data, artifacts, get left all over the place, like little breadcrumbs. And you don't want that, right. So with the data sharing ecosystem, the, our whoever we're partnering with to do the NPC, they can share that data with us from Snowflake, to our account in Snowflake, entirely secure. They don't have to generate any extracts, they don't have to do any ETL work, we don't have to worry about the wrong people seeing the wrong data inadvertently. So it just allows us for this whole process to end up being very secure again.

Kent Blacksher  38:04
And that's crucial. I mean, I would imagine And, Mark, want to ask you your thoughts in and around this, when you think from from your side of the Oregon and maybe even client facing? Can you can you give us your lay of the land on the privacy aspects, not just of this, but collectively as it pertains to data?

Mark Leher  38:23
Yeah, I mean, I, I probably would be foolish to think I can add any more value to the privacy conversation than Lance. Yeah, did. But what strikes me just listening to Lance and John the last few minutes. John, you said earlier, nobody wants a HELOC. They want a new basement is exactly the same situation with data. Nobody wants a better network or security or all all this stuff. These are all roadblocks to what they actually want is insights about their customers and the ability to execute on better relationships and deliver a better experience. And an everything John and Lance had just been talking. I wasn't counting but 2030 different robots that have previous previously existed to this. And now Carl, who's, you know, pissed because it takes hours to run a query. And it's running in eight minutes. He's happy that his query is running faster, but he's also a hero to the business side of the institution. And if you think they were grumpy before, they were probably grumpy before it too. And now they're delighted because they've got they've got the tools to win against all these competitive pressures that, that institutions are facing. So that's what I think. That's That's what I think is exciting.

Kent Blacksher  39:40
And it's all about keeping Carl happy. That's one thing we've learned today. And something I'm definitely going to take from this.

John Thuma 39:46
Well, you know, it comes down to time value of data, right? I mean, some data, you know, there's a range of data right now that's worth more. It comes down to time value money, same kind of same concept. data right now is worth more than it was or will be a week from now. And that's a very important concept, the faster I can process all of my data, all of my data, that's just some of my data, all of my data using tools like Segmint their merchant cleansing solutions, and, and they really have taken something that's very complicated. And they've made it something that is reachable by smaller institutions to so now these really advanced data science why I fell in love with the second I met Segmints product. I even know who these guys are, who these people were, I didn't know, Lance was, I didn't know.

Kent Blacksher  40:34
That's by design. That's by design.

John Thuma 40:38
Your face and I'm like, I love this, right? Why? Because anyone can use it. You don't have to be a taxonomist. You don't have to be a taxidermist, you don't have to be crazy person. And a crazy scientist with wicked hair. And, and and you know a lot of voodoo and you don't have to be a unicorn, you can be a regular person, look at their application, look at what it does, and understand how this how this can be used by your customer, or by your by your, by your people working in the institution. And you glean a lot of information about it. So the point I'm trying to make here is don't short. So don't short yourself on limiting yourself to your data. Don't worry about the hardware, don't worry about the cloud hardware, because there is cloud hardware, a lot of these solutions, you still have to think about, well, how many nodes do I need how many with us, you just flex that out or shrink it down? It's really simple. But my point is the opportunity costs of time. You know, if something takes eight hours, I can only I have to wait for that to come back. What am I doing for those eight hours? Well, I have to go do something else. I can't use my computer, because it's busy doing something. But if it comes back in eight minutes, I'm able to go, Oh, I did the wrong thing. I fail fast, or I succeed fast. And I move on to my next problem. And I can do that over and over and over and I can iterate over and over and over. That's the value of being able to clean your data and cleanse your data and merge cleansing solutions from Segmint, but also the scale and rapid speed that Snowflake provides.

Mark Leher  42:05
John, I'll come on come on up there with us this example that we see in transactions, I think a really interesting is micro deposits to test deposits. So we're seeing usually a transaction to Robin Hood or E trade. If it's $1,000, that's probably an investment. If it's three cents into your account, that's an indication that someone has just signed up, you're one of your customers has just signed up for E trade or Robin Hood or any number of 100 different FINDEX out there. That's something you want to know about pretty quickly. Because I'm hearing --

John Thuma 42:38
-- About the competitive aspect of that, right. I mean, you know, as a banking institution, I can learn a whole bunch of in please follow up with this. Marcus is more of your purse, your thing but thing all thing. Hey, who's not banking with me? Where's my money going? Right? It's my money going to Ford Motor Credit? Is it going to Toyota? Whatever? I can I can get a lot of that too, right?

Mark Leher  43:00
Yeah, those I would say are some of our table stakes use cases where you're seeing where were the competitive institutions, competitive mortgages, auto loans, student loans, really, how is your customer member interacting with the greater financial institution outs or financial ecosystem outside of the walls of your institution? Transaction data is going to tell you all of that.

John Thuma 43:23
Yeah, whereas before, I'd know if I'm, I'd have to walk into the bank, and talk to the bank and my bank or my personal banker, maybe if I'm fortunate enough to have one of those. And they would be able to look at that only when I walked in, right? Now you look at the whole, you look at the whole enterprise of accounts, you look at all of your accounts, and you can you can now start to look at trends and differences in but not at the individual level, as well as at the macro level, right to be able to glean the voice of the customer. And look at things like inflation, and stagflation and war and, you know, food prices and gas prices, and what what effects those are going to have on spending and spending behaviors of your customers, because we're all gonna have to tighten our belts because of all this.

Kent Blacksher  44:07
Absolutely, that's a that's a perfect example. And some of those practical use cases you guys were just talking about, I think, or I would imagine are of great interest to our viewers, particularly thinking about the competitive landscape and ways it can ascertain things like possible attrition or some branching out to products and services that we offer internally that maybe we want to make sure we try to keep our hands on. But I want to shift to Lance, because we've been a great dialogue. I know we're kind of reaching the last little bit here. But I'm curious as to your day to day and your time with Segmint. You talked about at the onset, the types of data we're typically using for cleansing we talked about maybe it's just Ach, maybe we're taking debit, you know, Bill Pay credit card data, a variety of different sources of data. If you're a bank or credit union or or FinTech or someone who's watching watching this today, in your expertise, what are the common issues or concerns these these organizations may have getting started? And or what's the best way it can crawl walk wrong? Walk us through kind of, for someone who's really looking to get rolling the best way to go about it?

Lance Cuthbert 45:20
Yeah, so kind of going back to our friend, Carl, that we've been using for this presentation. You know, once upon a time, somebody probably came and tapped Carl on the shoulder and said, Hey, we need all of this data. We want you to go collect all of this data, go harvest it, and Carl went and did that. And then came back and said, Okay, what's next boss? And boss said, well, Carl, go do analytics. Right. And, but I can't I don't have the resources. I don't have the know how I don't have that expertise. What do I do? Right? So I think, you know, where a company or financial institution needs to start is start working on that data strategy, or that vision of, you know, what data do we want to handle? How are we going to handle certain things like data, inconsistencies, data duplication, the variety of data, you know, any ePHI out there, I'm sure could point out that they have a loan servicing system, and a deposit system. And the two are completely different. They don't talk to each other. But we want to take the data out of both and use them for some sort of common analytics, and how do we do things like that? And, you know, we've talked a lot today about data governance and data security. So we need to understand those things. So I have a plan, you know, develop your plan, develop your strategy, and then work that strategy. work towards that. And, you know, you mentioned could it be a crawl, walk? Run? Yeah, I absolutely. You know, you can't, you know, one of the phrases we always use that Segmint is you can't eat the elephant in one bite. Right, you got to take small chunks and and make it so that you're getting value out of immediately out of the, the, the resources that you're putting into it. And I think, speaking of resources, the other issues that they come up with are, you know, the technical resources. You know, we've talked about how Snowflake allows you to scale. It's on demand, the security, right? And these are things that are without Snowflake, in the picture are very daunting tasks to somebody because they have to try and figure out, Okay, how much storage am I going to need? How much processing am I going to need. And these things are not easy to forecast, right? Especially for a small institution who may never have done anything like this before, right? They're going from Excel spreadsheets to an actual data warehouse. That's a pretty big leap, right? So you've got that, and then you actually have to have the people behind it as well to you know, you have to have the analytical know how you can have all of the data in the world. But if you don't have the right people to kind of help you tell the story from that data, it doesn't have the value that you want to have. Right. And then I think the other things that we've talked about as well, too, is the data security, you know, once you start collecting that data, and that becomes a very real thing that institutions have to be careful and definitely have to be cognizant of. And then we've also touched on this is, you know, okay, great, I've got the analytics, I've got the data, I'm generating information. Now, what the hell do I do with it? Right? What are my use cases? How am I going to make that data actionable? You know, many of the times we talk to financial institutions, and they may, they may have analytical processes that they're using right now. But, you know, by what, by the time one side of the house gets those models built, and does the analysis and gets the results to the other side of the house, the moment has passed them by, you know, they weren't able to react to what the data was telling them. So, again, how do we make this this data actionable? So I think those are some of the biggest hurdles. Yeah, for sure.

John Thuma 49:26
I think, you know, what the, excuse me, the, let me get onto that, if you don't mind, you know, small to medium sized institutions. The chasm was very big, they couldn't cross the chasm. It was the barriers to entry to these things to data science to things like Segmint to even, you know, Snowflake, they were wide and hard to do. They took resources they couldn't afford, they, you know, now those chasms are shrinking and they're it's much faster and easier to do, because the small institutions and medium sizes just want to do the things that the big institutions do. They want to create better audiences, they want to create greater engagement. They want to do marketing automation, they want to do closed loop analysis. They want to compare that to a holdout group. They want to have that next, you know, version of their Segmintation. They want to get to the micro Segmintation, almost down to the individualized next best action so they can personalize messages to their, to their, their account haters. And they want to get beyond the Net Promoter Score. And they want to they want to manage defection, they want to see who's who's closing down, who's How do I get the how do I get the credit card out of the sock drawer? Those are the things that the small to medium sized students want to do. With the partnership with Snowflake and Segmint, we enable that relationship to happen.

Kent Blacksher  50:48
Can you expound on that? Just a little bit more? How? How easy as it for someone to get started with Snowflake? Is this? Are you? Anybody in there, their brother can come and start as exclusive as it didn't? What's it look like for someone to get in there? And I also want to sink back on something I thought you said something about the ability for Snowflake to actually do advanced data science within the platform? Can you? I'm just trying to bring it full circle.

John Thuma 51:17
No, I appreciate that. So how do you get started? First of all, you can go on Snowflake comm it's very good. You can sign up for a free trial, right? Anybody can do it, you go in, there's a free trial, you'll see the button on our webpage. And we'll give you $400 worth of credits to run. You can load data, you can run queries, you can do all kinds of great things connect ODBC Connect, you know, Tableau, or Power BI or whatever, you know, reporting your ETL tool you want to connect to it. And never and that's open to everyone, right? We encourage that. If you know that we have different sales teams and account teams that can work with you as well. You know, so it's very easy to get started plus our documentation. If you go to Doc Scott, Snowflake Comm, you can see all of our documentation, and it's all laid out step one, step two, step three, we have a variety of also learning artifacts and, you know, little short 510 15 minute videos that show you how to do things very, very easy to use, and very easy to manifest. And it's available to all of you. And anyone can do it. You asked the next question was about data science. This is a very big area, we're putting a lot of investment here. And, you know, we're now in it's called snowpark. And what this means is, is that we can use Python and Scala you know, two of the most, you know, popular languages pythons, probably the the most popular language in data science and data engineering, to be able to, you know, load data engineer data without having to move your data to the analytic platform or to the ETL platform, you can perform those things using most popular libraries that are out there at scale. So you know, we we've pushed down the algorithm, the transformation down into the data, so the data doesn't have to leave the environment. So what that means is that Lance was touching upon this is all of our data at rest is encrypted, all data in motion is encrypted. And it's a single source of the governed truth. So that's why we're popular for the data lake as well, when you have all of that guarded by the encryption, the compression, when you have all of that guarded by role based access controls, and things that you know about SQL, but now you can exploit through Python and Scala and Java. You can you can build models that scale across your entire data ecosystem.

Kent Blacksher  53:43
Excellent. That's awesome. Guys, the conversation has been mind blowing for me. And I can only imagine for our viewers and in the vertical and elsewhere. final few minutes here, we'll we'll start with you, Mark, and then Lance, and then John. Any final thoughts, any views of industry or any tips for those paying attention to things that maybe impact right now they could start looking at? But Mark, I'll turn it to you for your final thoughts.

Mark Leher  54:15
Yeah, I'll just circle back to the use cases. Again, that's what I think's really exciting. Just I've been in the data and the taxonomy and classification space for 20 years, and probably anyone in tech can sympathize with. There's kind of nothing more deflating than having a great idea. Like, would it be great if we could do this? And then area budget? Yeah, but we'll see you in 18 months. And it's like, so it's just great. It's exciting to be able to have, again, from the data side and have the data clean, and be able to get in there and play and identify and look at use cases and see trends really quickly. Things that are gonna be able to drive real business value. So that's it think is fun. And I'd encourage every institution from the from the cheap seats, I think the number one tip is have someone who's going to look at the data. You got to get to know the data, immerse yourself in the data, and the solutions were to help them do that. And make sense of it more quickly, so you can drive that business value.

Kent Blacksher  55:21
Awesome. Thank you. Lance, same question. Do you have any final knowledge nuggets or tidbits?

Lance Cuthbert 55:28
Yeah, well, I guess, you know, as Mark talked about the use cases, I guess I feel compelled to talk about the technology. And with that, I would strongly suggest that everybody reach out to John, to talk to them about, you know, what they can do with the Snowflake platform and what it offers them. And you know, how they can turn all of those unhappy curls into happy curls. And I think Snowflake is a really good fit for any size, business small, medium, large, it has really helped Segmint go further faster than you know, a lot of the platforms we've worked on.

Kent Blacksher  56:07
Fair enough. John, I end up with you and your knowledge and wisdom.

John Thuma 56:15
I think that's great. Happy Carl's is important. You know, you know what, what Snowflake and Segmint do is allow you to have a conversation with your data. To mine your data, when you need to. You gone are that we talked about time value of data time value of action on your data. If I'm a an enterprise, institution, I need any scale, any size, I've got to do those things. And I also want to know, because I'm not having those one on one interactions face to face anymore. If Jason loves golf, the Cleveland Browns and stake right, if Jason loves the Cleveland Browns, he loves painting. So I love the Cleveland Browns. We also want to know if Jason if we can make a better offer to Jason maybe his cod is starting to mature his certificate of deposit is starting to mature. And if we can make a better offer to Jason and also offer him a car loan because he he has his car loan with another bank or another financial institution. We can make those two things together and increase Jason's portfolio with the bank. And that's the kind of things I want to enable. It's not just speeds and feeds. And while we can do all this cool gizmo stuff, let's we've commoditize the gizmos, right both of us. That's what that's what Segmint and Snowflake have in common. We've commoditized very hard things. Let's focus on your business problems. Let's focus on solving your biggest, nastiest thing that keeps you up at night and help your businesses grow revenue and grow activation and decrease costs. And if we can do all those three things we're winning, and that's what we want to enable here.

Kent Blacksher  57:59
Did not say it any better. What a perfect way to end this. Gentlemen, I can't thank you enough so much. On behalf of John and Mark and Lance, this is Kent Blacksher saying thank you for joining us for this Data Jam Session. Really appreciate the dialogue. We'll certainly be doing this again. Make sure you stay tuned to our socials and look for another session. But until then everybody stay safe and thanks for joining us. Have a good one.