In this episode, Veronika Durgin, Vice President of Data at Saks, talks about utilizing the data vault, leveraging infinite compute in the cloud, and so much more.
In this episode, Veronika Durgin, Vice President of Data at Saks, talks about utilizing the data vault, leveraging infinite compute in the cloud, and so much more.
--------
How you approach data will define what’s possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.
Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.
Learn more and register at www.snowflake.com/summit
Steve Hamm: [00:00:00] well welcome Veronica. Welcome to the podcast.
Veroinka Durgin: Thank you so much Steve. So lovely to be here. Thank you for having.
Steve Hamm: Sax Fifth Avenue is a well known brand among luxury goods retailers, but the company last year spun off its e-commerce business as a standalone company called Sax. The two companies have overlapping ownership and seem to work closely together.
So please help us understand the relationship between the two companies and how they collaborate.
Veroinka Durgin: Absolutely. S is the premier digital platform for luxury fashion in North America. Sex. The e-commerce business has an exclusive partnership with the Sax Fifth Avenue stores business that allows us to deliver a similar shopping experience for customers. For example, with the stores, our e-commerce customers have access to services like buy online, pickup, on store.
They have the ability to return in store as well. They as, as they have access to specific, um, services like alterations. [00:01:00] Each business is able to have more focus on their individual experience while working together through service agreements to ensure that our customers are served seamlessly across the entire Sax Fifth Avenue ecosystem.
Steve Hamm: Yeah. Um, I think when we talked before, you talked about how the data belongs with sacs or as the data is shared. Could you just talk a little bit about that? How is kind of, how is data shared amongst these two overlapping companies?
Veroinka Durgin:. We own the customer data from the full shopping journey across the Entires Fifth Avenue ecosystem. We don't classify. Store data differently than any, any other data, and we make sure to align store data to our digital data.
From our perspective where customer engages, whether it's withs.com or SAX fifth Avenue stores makes no difference in how we treat data. We take a holistic approach to the customer journey by leveraging all of the data that we have. One of the things that [00:02:00] makes SAC stand out for me is our overall customer focus and high standards of customer.
By joining digital in store, we're able to provide our customers the best experience regardless of how of the way they shop with us. I mean, we have very rich first party data in our challenges and also our biggest opportunities is getting this data to value quickly so we can use it to power a truly memorable customer experience.
Steve Hamm: Okay, great, great. Now you come to this role after a number of years as a data scientist and a data engineer in a half dozen different companies. How have your experiences as a kind of in the trenches data expert prepared you for your current role?
Veroinka Durgin: In the trenches is such a wonderful phrase. It truly was in the trenches. Uh, I started my career actually as a database administrator, and I spent quite a few years focusing on performance optimization and [00:03:00] tuning of on-premise database systems. And I mentioned on-prem because there's some art. To architecture, uh, of a platform in designing a database so we could do more with the hardware that was pretty much limited.
We couldn't really easily scale our hardware when we were on premise. Almost infinite Scalability of the cloud makes it very easy to solve performance problems by simply adding more compute. So in that inevitable, it leads to very high cost of running data platforms in the cloud. I think my experience certainly taught, taught me to look at optimization first before increasing. the same time, I am super excited for all the advancements in the data space and new opportunities that they enable. Given the fact that in the past we didn't have all of that and we spent a lot of time trying to optimize our systems.
Steve Hamm: Yeah.
Veroinka Durgin: UHS data team's focus is to deliver value to our users fast, but we also have to be very careful and try not to accrue too much tech debt because this is what ultimately makes our [00:04:00] solutions very expensive.
So for me, having this practical experience of various implementation helps me guide my team through the trade offs of various solutions. Uh, having first has hand experience, um, of how difficult it was to accomplish many things in the past, uh, again, makes me very excited for the modern data stack, uh, and Snowflake and various other tools and platform removed a lot of constraints that we as engineers had to deal with in the past.
Steve Hamm: Yeah. Now, what was the data management situation at Sac and Sac fifth Avenue when you arrived? I think it was mid 2021. That must have been right around the time that the two were separated. So kind of give us the sense of what the status was. You know, where, where were they in terms of, um, cloud? I mean, had they already migrated applications and data to the cloud or, or was that happening just then?
Veroinka Durgin: Well, ours Fifth Avenue brand may be renowned [00:05:00] in almost a century old, uh, but we operate through digital lens and move as an early stage startup. I actually get teased a lot by France, uh, that I joined Sax for fashion and. Why I really do love my shoes. Uh, the reason I first joined sex was for tech. Uh, Sex has a modern technology and data stack and we're well positioned to make positive tech advancements.
Typically at older companies, we tend to see older data stacks with practices from, you know, 10 to 20 years ago, uh, withs. eCommerce, separation and digital mindset. We have the power to do everything that a young startup can and more. But to kind of answer dive a little bit deeper into your question, before I started Sex was already using Snowflake for their data platform prior to separation when they were still part of the parent company initially.
And you probably hear the story often data was lift and shifted from on-prem systems to Snowflake. While that certainly [00:06:00] gave the parent company ability to grow, as you can imagine, doing lift and shift did not only bring all of the on-prem issues to the cloud, but in many ways it also prevented the team from using menu Snowflake features and capabilities.
Now as part of the separation sac decided to build our data stack from scratch so we can modernize our platform and approaches and really take advantage of all of the Cloud capabilities and all of soft Lake's capabilities.
Steve Hamm: Yeah. No, that's, that's great. So, um, you became a aware of Snowflake earlier in your career before you came to sex. What was the attraction? What was the situation there?
Veroinka Durgin: So is, as I mentioned earlier, I spend. A pretty large portion of my career working with on-prem databases, and that was good and challenging in many ways. Uh, when my team was tasked to migrate our kind of first on-prem database systems to the cloud, we decided to test both lift and shift and [00:07:00] rebuild solutions.
Typically, again, you see probably lift and shifts because they seem to be a little bit easier, but when we actually tested. Rebuild solutions. Rebuilding our service in the way that took advantage of various cloud options allowed us to get faster and cheaper systems. This is the moment. This was a little bit of an aha moment for me.
This is when I got really hooked up on the cloud because it offered me opportunities to be really smart and creative and enabled me to do a lot of things that I couldn't before. So that was about five years ago, and that's exactly the time where I first heard about Flake. So Snowflake message at the time was that it just works.
Uh, and you know, I kind of took it as a challenge and in my mind I was like, Okay, let me see if it actually truly just works. But once I got my hands on it, I really didn't wanna let it go. And you know, if anything is a proof of that, is the fact that I'm right now is Snowflake data superhero.
Steve Hamm: Right, Right, right. Yeah. Now, uh, [00:08:00] you said that thats adopted, Snowflake. Before you came there, uh, what were the most important initial applications for Snowflake ATS and, and one of the benefits that the company is gaining from them at this point?
Veroinka Durgin: Well, Sex utilizes Snowflake to power our data platform and host all of our data for easy access. So the, our platform sac, um, uh, Snowflake, um, holds all of our raw, raw as well as model data in one place, which truly improves user's ability to browse and discover company data. It allows us to reconcile data, validated.
We do various historical analysis and we also support our various machine learning efforts. Having all of S's data in Snowflake. Also positively impacted personalization capabilities and customer experience in marketing by basically providing access to auditable detail and understandable.
Steve Hamm: Yeah. So is, is, is [00:09:00] Snowflake really the key to Stacks and Sacs? Fifth Avenue being able to share data in such a rich and valuable way.
Veroinka Durgin: Snow Snowflake is the central data platform, so it's not necessarily just even S and Sax Fifth Avenue data. Every data from every source that we have, we utilize Snowflake data sharing to add additional third party data to Snowflake. So it's kind of sitting at the core of our data platform.
Steve Hamm: Yeah. Now it sounds like you're pretty enthusiastic about Snowflake, which is great to hear, obviously. Where do you see the relationship with Snowflake going from here?
Veroinka Durgin: So in general, when we look at really any platform or any tool, we consider three things, right? First one is whether it's cloud native, and if anything, snowflake is cloud native. Um, Functionality. So our primary goal is to deliver value to our cus [00:10:00] to our users and customers fast, right? So we wanna take that data to insight as quickly as possible in as reliably as possible.
And Snowflake also checks this box. We are also looking at companies that we can partner with. The ones that, you know, help us grow and also grow with us. And we consider the maturity of the company, their culture, the rate they're innovating at. And Snowflake checks off this box as well. So we have this really great partnership happening with Snowflake.
The, the other thing I kind of wanted to dive a little bit deeper into is functionality that Snowflake provides. Snowflake markets itself as data Cloud. And what it really means is that, Knows as a company we can store multipurpose data in the same platform, right? Snowflake is our data lake. It's where offload some of our operational reporting.
This is where we build our data vault so that we can have access to modeled auditable data. This is what our machine learning teams are [00:11:00] using, so it's giving us all of. Variations and flexibilities on the same platform. And there are obviously, you know, future things that are coming to Snowflake, like Union Store that we're incredibly excited about as well as ice.
Steve Hamm: Yeah. Cool. Cool. Hey, let's go a little deeper on data modeling, You know, I mean, this is a big important concept in data analytics and data management, all this kind of stuff. So what are the kind of the, the most cutting edge data model techniques that you're starting to use now
Veroinka Durgin: Such an amazing question. I think all of the modeling techniques that have been. Proven are actually very old and have been around for a while. Um, what is really incredibly interesting about that is many teams and companies are just discovering them. So, um, I mentioned, you know, traditionally you probably see from the data warehousing perspective, you traditionally see United States Kimball type of [00:12:00] modeling with your dimensions of fact.
I mentioned data vault. Um, data vault is something that we use that I, I feel it's. Really, truly a powerful methodology. Um, it's been around since 2000, so there's really nothing new about it, but I feel that building data vault on Snowflake is what gives it almost like a turbo button where it truly.
Together, they make each other shine. Uh, p people are also now talking about, um, streaming data. There is kind of like, that's still newer conversations. Um, so I don't know where data modeling will go with that specific, um, kind of like area. So we, we are yet to see that. But I think it's only new because it's old.
It's like, what's all this new?
Steve Hamm: Oh, that's really interesting. Hey, tell us a little bit more about data Vault.
Veroinka Durgin: You know, I, I learned about Data Vault from Can Graciano. Uh, I literally read a blog on Snowflake that said you can run data vault on [00:13:00] Snowflake. And I went, What the hell is data Vault? That was three years ago. Never looked back.
Steve Hamm: Yeah. Well that's really cool. Well, the amazing thing is this is incredible. So within this field, I mean this field of of, and I'll call it data management and data analytics, there's so much to know that even experts are still discovering things that they didn't know before and finding new tools that they can use.
So I think that is really another exciting thing about the.
Veroinka Durgin: I mean, for me just to add. I, I, I find myself actually very lucky that I kind of grew up, you know, for the first two decades, learning the old ways of doing stuff because the new stuff now makes me so excited. And what's really awesome is that when we put the old patterns that are proven together with new capabilities, this is when the magic happens.
Steve Hamm: What exactly does data Vault. do, and how does it [00:14:00] kind of interact in a positive way with Snowflake?
Veroinka Durgin: So data, data vault methodology includes, you know, architecture, it has data modeled, it has, you know, methods of delivery. It's kind of like this end to end methodology, but I think there, there are few things that important to me from, from specifically from that. Um, data vault methodology, and I think the biggest one is modeling business first.
We are so used to data professionals as engineers diving into specific source systems that we forget to talk to our business users and understand how our business works. Once we model our business, then source systems kind of don't matter anymore. We can append. Source of data to, to fit into our business model.
And I think this is super important because it makes our data model very flexible. Um, the other great [00:15:00] features of data vault, you know, it's very much pattern based. It's, it's very automatable. It's, it's meant to be agile where you can incrementally deliver fast to your business users. It's historically, it's auditable, so we don't actually lose any data and transformations.
Take all the data as it is, we can always go back and see how things were as they were at that point in time. What makes Data Vault kind of even shine brighter with Snowflake is Snowflake ability just to store an enormous amount of data and kind of scale compute so we can run multiple process and parallel.
We can store a lot of data. It's compressed. It doesn't actually cost us that much.
Steve Hamm: Yeah. It's interesting what you said about business modeling and how that's where you have to start. Because when I think about, you know, sophisticated companies today, it's almost like the business leaders are in one eye, every tower over here, and the data scientists and the data engineers are kind of these wizards in another tower.
How does your company kind [00:16:00] of get them together to meet in the middle, to really make sure that the data management system and the data analytics is really. You know, doing the right thing for the business.
Veroinka Durgin: Well, as I tell my team, we have to talk to other humans who are not engineers. So we . I know I'm, I'm kind of being funny, but, um, we, we meet with our users. We try. Understand what they're telling us. We, we try to come up, uh, with common business definitions. We went through this incredible exercise of defining all of our KPIs in very simple language.
So we cannot collect, refer to that document and understand what everybody's talking about. And by start. Conversations with business users who can explain to us how business works. It also helps us identify gaps in our systems, right? This is general what it comes out of. And I also value, um, data education a [00:17:00] lot, right?
So while we try to learn how to speak, you know, non-technical, non data language, at the same time we're trying to teach our users how to speak with databases, how, how to understand data. They're looking, you know, looking at how to derive insights.
Steve Hamm: Yeah, it's interesting cuz you know, the whole history of BI has, or the promise has been, oh, you're gonna be able to have the executive or the business leader kind of communicate directly and, and really understand what's going on now. I think that's, that's always been the promise. Just out of reach. Do you see that being, coming more reachable, more touchable?
Veroinka Durgin: I mean, to be completely honest, um, s is probably the most data savvy company I worked for. There is a very large number of, you know, business users that I've spoken with that know how to use data, which is, you know, one of the analogs that I usually share, they're like these amazing runners. All [00:18:00] I have to do is give them new gear so they can run faster, but they already know how to run.
Steve Hamm: Okay, well that's really cool. Okay, so let's look into the future a little bit here. So looking out over the next year or so, what are the most important data management and data analytics trends that you see emerging?
Veroinka Durgin: I, So we, we have just a very, Okay, so I think companies realize the value of. High quality data. I think it's, it's pretty clear to just about every single company out there. So, and they're starting to stop treating data as software exhaust. I think what I, I'm hoping to see maybe, This is more of wishful thinking, more investment into capturing data in the way that is also meaningful for analytics.
So we don't have to scramble, do reconciliations after the fact. [00:19:00] Instead, analytics is upfront right there as software. Capabilities are designed. So I'm hoping to see that. I also, uh, I'm actually positive that what's gonna happen is every company is gonna look at their cloud bills and they're gonna start paying attention at how they're running their platforms.
Steve Hamm: Oh, well that's interesting. It almost goes back to your old, you know, kind of where you started as a DBA, where really tuning and efficiency are important even in the cloud. Yeah.
Veroinka Durgin: And you know, in the past I had to do that. I had no choice. But with the cloud, almost infinite compute, nobody's paying attention to it until it shows up on the bill with a, you know, very high amount number.
Steve Hamm: Yeah. Yeah. I gotcha. Hey, I'm gonna ask you to put on your visionary cap now for a minute. Looking out five years or more, where do you see analytics going for businesses and even for society?
Veroinka Durgin: It's such a great [00:20:00] question. Five years. In data years is a very long time. You know, I kinda look back and I think that it was probably five years ago when I did my very first migration from on-prem to the cloud. And look where we are right now. Um, you know, five years ago Snowflake was just starting and, and look where it's now.
Um, I think, so my, my guess is because we're generating data at such a high rate, we might actually run out of compute. So I wonder if we'll have much more data than we have, you know, compute capabilities. So we'll have to start thinking about optimization. We'll be forced to think about optimizations and running, you know, our processes in a very smart way, and potentially not saving all data, but only data that is meaningful.
Steve Hamm: Right,
Veroinka Durgin: I, I'm also. You know, obviously data teams will become smarter and more sophisticated to do that. I'm [00:21:00] also expecting for many users to become data inform, you know, data savvy, data informed to know how to use data to make decisions. Um, The other thing I'm thinking is that as we get more efficient in how we process our data and how we store it, we will actually reduce carbon footprint of our data centers, right?
The less data we store, the less compute we use, the less power we need to spend to, you know, power those data centers.
Steve Hamm: Yeah.
Veroinka Durgin: The other thing again that I'm like hoping for is to see cloud platform agnostic data sharing. I think that would be an amazing thing to.
Steve Hamm: What do you mean by that?
Veroinka Durgin: So right now data sharing, kind of like you can easily share data inside Snowflake ecosystems.
You know, it's easy for us to purchase data on the marketplace. It's easy for us to share data with, you know, our partners who are also using Snowflake. But I would love to see that happening across, you know, Google [00:22:00] Cloud, big query, Azure, Azure, Synapse. So kind of like abstract, make that data sharing platform and cloud.
Steve Hamm: Oh, interesting. Yeah. Hey, so, um, something said a moment ago struck me about how fast things are changing in this area. And five years is an incredible amount of time. I'm actually writing a book with a, with a tech executive about the data, data analytics, data management space, and we've made the decision, there's no way to publish this in.
And with the traditional publisher, because the world changes so fast in a year that it takes to publish a book that you might look like a fool. So things are just racing ahead and, uh, I, I think this is, it's been going on for, well, at least a decade now, at this pace. And it'd be interesting to see how, how, how long it can do it.
But there's just so much innovation still coming. It's, it's really amazing.
Veroinka Durgin: It truly is incredible, and I think we're [00:23:00] kind of pushing ourselves to innovate. Again. We capture a lot of data. Everybody understands and realizes the value of data, which leads us to capture more data, which leads us to actually figure out how to store all of that data and process it.
Steve Hamm: yeah. No, that's really driving it, isn't it? Yeah. Hey, so we're about to wrap up and we like to finish on a lighter and more personal note. I understand that you see yourself as something of an evangelist for data for the modern data stack technologies, in part in particular, uh, very gratifyingly for Snowflake.
Tell us about that. What's going on?
Veroinka Durgin: know, it's funny, I oftentimes get comments like, Oh, you have this public persona. I, I don't consider myself an evangelist, or, you know, kind of trying to have that public persona. I really enjoy using Snowflake. I, I love Modern Data Stack and, and the capabilities that a lot, you know, it gives [00:24:00] me. I love data vault.
I work with it every day. So speaking and blogging allows me to really share my enjoyment with others. And I also really love learning with the goal of truly understanding things. I'm kind of curious and I get interested in things. And being involved in the community allows me to meet some really cool, incredible people like you, Steve.
So, um, so just I do it for fun because I.
Steve Hamm: No, no, that's fantastic. I, I really like your spirit though. This has been a fascinating conversation. You know, you talked about, The Sax Fifth Avenue brand. It's been around for many, many years. And you know, sometimes in the tech world we, we look at these old, old brands where we think, Oh, that's an old, old company that's kind of ossified and it's thinking, or it must be stuck in the mud or something.
But you said that in technology. You knows, really operates like a startup. And I think, you know, that's a great example. That's something that a [00:25:00] lot of other companies can learn from and to realize you don't have to be stuck in these old systems or these old ways of thinking. You can really reinvent yourself.
I mean, it may, that term may be used too much, but it, but it's, in this case it's true. And I think it could be true in, in a lot of cases. So I wanna thank you so much for the conversation that I think it's been fascinat.
Veroinka Durgin: Thank you for having.