This episode features an interview with Frank Slootman, CEO of Snowflake. In this episode, Frank gives us an update on the transition of Snowflake from a private startup to a public company, the impact of the Data Cloud over the past year for organizations across industries, and the future of data sharing.
This episode features an interview with Frank Slootman, CEO of Snowflake.
In this episode, Frank gives us an update on the transition of Snowflake from a private startup to a public company, the impact of the Data Cloud over the past year for organizations across industries, and the future of data sharing.
--------
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] So Frank, it's been about a year since we started off the podcast and yours was the initial one. So a lot has changed.
I wanted to talk to you about one of the most important things that's changed, which is really you you've. You and your leadership team has coined the term data cloud. You know, the idea that data is mass migrating to the cloud, where it can be better integrated, managed, shared, analyzed, all that good stuff.
And you positioned snowflake as a leader or the leader in this journey for businesses and organizations. So I wanted to know how has your vision of the data clouds evolved and developed over the past months?
Frank Slootman: That's a great question, Steve , um, in our, our, our basic observation about the industry is that , uh, infrastructure clouds are of course, highly , uh, scaled and consolidated in the likes of , uh, you know, Amazon web services, Microsoft Azure, and Google cloud. Uh, and so on, there are massive and getting bigger. Um, we also have very large [00:01:00] application clouds, right?
Alexa, Salesforce, and, you know, Workday and service now and SAP , uh, and so on. And that's all good. And, and, and customers, you know, broadly use these capabilities and then you go look for data and it turns out that data is fragmented and segmented in millions and millions of places, you know, behind all kinds of perimeters network systems, applications, and so on.
So you wonder why, so data is just taking a back seat to , uh, to the infrastructure as well as to the applications. And, you know, you're gonna get asked the question, why is this an issue? Why is this a problem? Um, you know, a lot of the analytics , uh, efforts that we've had as an industry, we're very much what we call in silo, meaning that they were targeting, you know, one specific data set and, you know, using things like Deb wall to, to dashboard that.
And uh, you know, people were reasonably happy. With those kinds of results, but the world is changing, right? Data science is really doesn't care. Number one, you know, about the boundaries that exist , uh, you know, between data silos. Um, they're very much into discovering and describing [00:02:00] relationships between data sets.
That might be a, you know, very unlike. Uh, unlikely if you will, in the sense that , uh, I might, you know, relate , uh, an internal data, separate an external one structure would unstructured , um, because it's really, it's really about understanding data relationships, because once I do, then I can use those relationships to uh, to train models on and then I can use them to predict.
Right. And , um, you know, obviously that that's the value that we get from discovering data relationships, but these data relationships can not be impeded and encumbered by data silos and data living in millions of places. So it's very important that we arrive at a notion of a data clouds where , uh, the workloads have unfettered access to data, regardless of, you know, what it is and where it lives and so on.
And we're going to blend it and overlay it and join it. To our heart's content, so that data science , uh, teams can discover these very, very high value relationships that are there to be discovered, you know, in the data. So that's sort of the , uh, the, the, the whole reason , uh, why the notion of a data cloud is [00:03:00] essential.
I mean, we're, we have the technological underpinnings to do it. We do it every day. Not it's much bigger than snowflake. It's something that has to happen in the industry. Given how important data science is, is shaping up. Um, you know, for all of us , uh, you know, not just in the world of tech, but then in the, in the business world in general.
Steve Hamm: Yeah. You know, I think you've done a good job of, you know, stating the vision and also positioning the company. Because I think when people out there in business think about the data cloud, they think of snowflake, which is a pretty important strategic spot to put yourself and so well done with that. Now you talk with hundreds of customers every year.
And I know that this has been a pretty strange year because of COVID, but I imagine you're doing a lot of phone calls and zoom calls. What are the biggest business trends related to data that you're hearing about from the C-suite.
Frank Slootman: Obviously people are still , uh, incredibly preoccupied with the , uh, the effects of the , uh, the pandemic. I had a call earlier [00:04:00] this week with the CEO, a very large uh, you know, pharmaceutical healthcare. Company. And , uh, you know, they were talking about how overnight, when the pandemic hit that , uh, that the demand coming from hospitals was all over than that, but it was up 30% here.
It was down 30% there. Um, so managing a supply chain and shipping things just became almost impossible. Um, it was just frantically calling around trying to get a grasp on a, on a rapidly evolving situation. That was sort of the first time where people started to realize this. We can't do things, you know, uh, you know, based on, you know, historical trends anymore because the trends are all broken.
We can't steer the ship by its wake. Right. We need to understand the relationships and then you use those relationships to predict. And , uh, you know, one of the datasets on our snowflake , uh, data marketplace is called star schema is essentially a real time COVID , uh, incident, the fatality. Uh, very detailed , uh, data set.
That's very, very widely used by our customers because [00:05:00] they're using it to figure out, you know, where the disruptions are both to the, to the positive and, and to , uh, to the negative. So there's this incredible awareness that anecdotal observation, you know, isn't going to cut it when you have this location of this , uh, this sort of magnitude.
And so that definitely has been a catalyst in all our conversations. Uh, the notion of it. Data Claus kind of farm because now we always talk to, to a large enterprise is about them having their own data, cloud their own brand at institutional , uh, enterprise data cloud, where they're, you know, unique data architecture, their data sources, you know, how they provide data as well as how they consume data from the outside, how that comes together because you know, data clouds become the big.
The beating heart of the mother enterprise, right? So we need to start evolving towards , um, an infrastructure like that. So we enable, you know, really our future that way. Right. And we, we, we talked a lot about optionality, which means it's very hard to know, you know, what we're going to need, but we got to architect ourselves in such a way that we have an enormous amount of flexibility to move in every direction, you know, [00:06:00] as, as the, the world sort of unfolds in front of us,
Steve Hamm: Yeah, no, that's really interesting. Hey, I wonder, I mean, obviously there are huge changes in the way people see data and, and use it internally. And then with partners. Is it changing business in kind of beyond the data? I mean, for instance, are companies starting to, you talked about supply chains, you know, are companies really starting to think about, Hey, we should have redundant supply chains and maybe we should have suppliers from our continent and things like that.
Are they rethinking things pretty fundamentally or is it pretty much just we're going to manage it better the way it is?
Frank Slootman: Well, I mean, it's really changing , uh, in, in terms of , uh, latency in terms of, you know, how fast things arrive and how fast systems , uh, react to events that are discovered , uh, you know, in the data. I mean historically , uh, you know, what we've done is , uh, you know, we run batch processes. kind of shivered is hurting the word batch, try to just ditch back to the 1970s, but that's, that's still, you know, how things work , uh, in, in a lot of places, you know, where we run in these [00:07:00] massive processes and, you know, in a matter of weeks, you know, the dashboards are populated and, you know, we have nice visualizations and all that sort of.
But fundamentally data is being used to inform people and let these people do , uh, after that , uh, is, are known , uh, if anything, right? So, uh, I, I call it so last century, you know, that, that whole mentality , uh, you know, about doing things from reporting to dashboarding, of course, where it's all moving. Uh, you know, we've heard the term digital transformation , uh, probably once times, because everybody uses it to , uh, to describe this about anything that's going on.
But, but to us it means is running end to end , uh, digital processes , uh, fully disintermediated from, from human intervention and they run , uh, at Lightspeed. They are completely data driven. They're incredibly, highly scaled. They're very economically and they're incredibly precise. And , uh, you know, that, that obviously is, is how, you know, modern enterprises are really instrumenting and building themselves, right?
They, they start from the ground up by , uh, you know, by running themselves end to end digital. That is the big trend. That is the big change. And , uh, you know, [00:08:00] for people who have historically grown up, you know, with the more traditional , uh, reporting dashboarding , uh, attitude towards data editor , they're, they're trying to change and they're trying to build data science teams, and they're trying to.
Put in place data platforms like snowflake and really catch up and evolve, you know, to that level of sophistication. And it's a big change. It's a huge change.
Steve Hamm: Yeah. And it sounds like that change in the availability of data and insights really lets the whole business just be more flexible, you know, because we, I mean, there was a lot of inflexibility flexibility and in business supply chains and, and relationships and stuff like that in the past, but it sounds like this is the tool that opens it up to being a much more flexible environment.
Yeah.
Frank Slootman: Yeah, we have this term, Steve , uh, the time value of data, and everybody knows about the time value of money. Um, but we sort of created a derivative of that. Uh, the time value of data changes a lot when data arrives, you know, in seconds and minutes versus in days and [00:09:00] weeks, what does that change in terms of what now becomes possible?
Right? Because as data ages, It rapidly loses its value and its its impactfulness. Right. And then one of the big things that we're, we're, we're seeing it snowflake is that because people are able to provision, you know, so many concurrent workloads, but such enormous sized clusters, that data is showing up, you know, in seconds and minutes and all of a sudden sort of exploding people's heads.
Like what, what can we do with this? Right. And , uh, you know, right now we're in a, we're in a situation where, where we're, we're only limited by people's imagination. And their budgets because the technology is not going to hold you back anymore. And I said, that's an enormous change from where we historically yeah.
Steve Hamm: Yeah, no, that's great. Now snowflake went public over the summer. Remember that, you know, And it ended up being the most successful software industry IPO ever. I believe now how has being a public company changed the demands on the organization and on your leadership group? I mean, you talk about [00:10:00] mobilizing data.
Do you have to do that like in new or different ways than, than when you were a private company?
Frank Slootman: well, one of the, one of the big things you have to do as a, as a public company is that you have to guide to financial markets. And in other words, you know, you provide guidance on what you're going to do for the quarter and what you're going to do for the year. Um, and depending on what kind of business model you have, that that can be, you know, incredibly challenging and work can be relatively straightforward.
Because we have a very new business model, which is based on, on consumption, consumption, exploit equals revenue. And in our world, it's a one-to-one relationship. We have to guide markets on, on things that haven't happened yet. In other words, you know, how do we know what our customers are going to consume in terms of snowflake for the period that we are.
Just about to gift guidance for right now. That's because w what do we know about their priority, their intentions , uh, and so on. So we make heavy, heavy use of snowflake itself. Uh, and we, we built various sophisticated machine learning models , uh, to reference historical data and [00:11:00] patterns to allow us to tell the markets, this is where we believe we will end up for, for the period.
And , uh, you know, prior to going public , uh, we actually spent the year before actually, Giving ourselves mock guidance, just to see how we were doing , uh, on the we're running these very, very sophisticated , uh, models. And , uh, that's worked out really well for us. I know we're, we're quite precise , um, um, guiding the markets , uh, you know, on where we are in consumption and revenue, but that's , uh, that's a good example of something that is completely different from, you know, what we had to do when we were private.
Steve Hamm: So Frank, I understand that one of the top priorities post IPO is that snowflake is really reorienting itself and going vertical. Can you talk about how the data cloud is helping healthcare and finance and some of these other organizations or other industries , uh, changed the way they do business?
Frank Slootman: yeah, they did the drive towards , uh, you know, much stronger vertical market orientation was not so much as triggered by the IPO. Uh, it was really triggered by our orientation [00:12:00] towards very large institutions and enterprises. And, and the reason is, you know, when you get high up in large institutions, They don't want to hear about architectural distinction and you know, why you can run this workload so much faster.
Uh, they want to know how you change their lives , uh, in terms of business outcomes , uh, customer experience , uh, you know, whatever it is that matters to that organization. And those ended up being very industry specific conversations. I mean, whether you're talking to a bank or a hospital, Is a completely different conversation and customers really demand that we step into that world as opposed to we make them step into ours.
And that's been, that's been a big pivot. Um, it's very important. You know, when we have conversations with customers that they always have an industry context, they're not interested in sort of high level abstract conceptual conversations. We know we have to bring it down, you know, to, to their relevant industry context , uh, of the conversation.
So, you know, we're talking. To a big bank. Um, you know, and we talk about the snowflake data marketplace, you know, we specifically talk about listings and data [00:13:00] providers , um, you know, like New York stock exchange, like SNP, like FactSet, because the old that is the type of data that they use. And then of course what they want to know as well.
What are other people doing? Uh, what snowflake did I should be interested in? Um, so those conversations have a lot of traction , uh, and , uh, and that's really the evolution of, for us as an organization. That means that. We have to learn industries and we have to learn it fast. And I'm, I'm, I'm personally very engaged in , in, in, in, in really accelerating and being a catalyst for making sure that we become really good , uh, um, you know, industry partners to our customers that they feel like, yeah, you guys know a lot too, you know, a lot about my industry, you know, a lot about what's going on in other institutions , uh, that is helpful and that is productive for me.
So that's, that's the evolution, like I said, Driven, you know, by us being in larger institutions and higher up in the tree, if you will, because that's the type of conversation they want to have with us.
Steve Hamm: Yeah, it looks like you also have to up-level who you talk to in those organizations. You're going from the technology leaders to the business leaders, or maybe a combination of the two , but, but you're definitely talking to the business [00:14:00] leaders , uh, these days, which means you're getting more and more strategic.
I would think.
Frank Slootman: Totally true. I mean, we used to , uh, you know, sort of reach for CTOs, chief technology officers, chief data officers, CIO, chief information officers. Um, I can tell you that that more than half of my meetings are now with CEOs. Uh, that's a huge change. Um, and that's because CAS are becoming very, very interested in data and what data can do for their enterprises.
You know, as we discussed earlier, COVID has been the catalyst for this kind of thinking , uh, probably it was going to happen anyways, but like so many things go with this and there's been , been an accelerator of many things that were already happening, but now they're happening quicker, you know?
Steve Hamm: Yeah, we've had actually, we've talked about this in a couple of our other podcasts, that just the whole idea that the companies are starting to see themselves as technology companies. I mean, they have, that has to be part of their expertise and, and what made them realize that is when they started to realize the value of data.
So that's the switch that makes the CEO realized that he didn't, you can't just kind of put [00:15:00] technology off on the side and not think about it. He's got to engage.
Frank Slootman: yeah. Tech technology, obviously all companies, I don't care what industry you're in. I mean, everybody is a technology company or whether we like it or not , it's, it's being forced on us. Uh, you know, we announced a very strategic relationship with BlackRock , uh, you know, recently, and of course black crime is by far the largest asset manager.
Uh, you know, in, in, in the U S or, or, or in the world for that matter, you know, 22 trillion in assets that are being managed by hundreds of , uh, of institutions. But, you know, aside from being an asset manager, they also have the industries , uh, you know, most prominent software platform , uh, for managing assets , uh, as well.
So they are a full-out technology company. In conjunction with, but also being a financial law institution, they go hand in hand. Um, so it won't be long before in many conversations, we just can't separate the technology conversation anymore from the business conversation. Yeah.
Steve Hamm: well, BlackRock is, is so powerful. In fact that it basically tells businesses the business world, how it wants them to [00:16:00] behave. I mean, you've, you've probably seen some of the stuff where they said, Hey, climate changes company, we're going to invest in companies that deal with it. Not that that turn their basis away from it.
So I, you know, as you go higher and higher in these larger organizations, you are going to be part of the conversation with the, with the most powerful, you know, leaders in industry. And that's, that should be pretty pretty thing.
Frank Slootman: Uh, that that's absolutely true. And , uh, you know, technology skills are becoming a mainstream. They're not just for technical people. Uh, you know, everybody has to become technology savvy and conversant and all those kinds of things. I mean, people have been trending that way obviously, but , uh, again , it's, it's it's accelerated year.
Yeah.
Steve Hamm: You know, a lot of people are talking these days about the value of sharing data. Um, you know, I think. People talk about it more these days, because it's easier because of the cloud. I mean, it used to be one of the big bottlenecks sharing data internally and especially externally used to be just hard as can be, but the data cloud has changed [00:17:00] that.
So if you can talk about what, one of the most interesting things that you, that your customers are doing internally and with their business partners, or even with their customers. Creating, I think what you've talked about, this data network effects what's going on there.
Frank Slootman: the reason that that data sharing is , uh, is such a big deal is, is because data science is doesn't really care about , uh, silos and boundaries between data they're interested in data relationships, and they typically are, are discovered in a exists, you know, across data sets. So if you can bring data together, you know, you, you are not in the.
Positioned to discover and really mobilized those relationships. So that's, what's driving data sharing. Uh, we have to bring data together. That is just a big shift from , uh, you know, we're where we have storage. We have been where we just did analytics inside of a single silo. We call it a day. We populated a dashboard.
Uh, we went home at night. That's just not good enough. And data sharing was really hard historically because we either used API, which is like sucking data out through a [00:18:00] straw and had all kinds of other , uh, you know, management issues, latency issues, and cost issues, and so on. Or we use a massive, heavy lifting, you know, through a file transfer , uh, protocols.
And that of course had, you know, we had to surrender custody of the data, to the all kinds of governance problems related to security and privacy. So after a while we were like, you know what? This is just too hard. To do it anymore, or we're going to do as little as we possibly can. Um, because of the snowflake data cloud , uh, you know, we, we can just now turn on data sharing relationships and literally in minutes, I mean, uh, you know, w we, we shown that as part of operation workspace , Uh, in terms of managing vaccine data out there, how important it was to create a visibility across the supply chain, not just in the silos of a ups or a Pfizer or a Johnson and Johnson or a CVS, right?
Because data starts to make sense when you, when you can see the entire supply chain, as opposed to just one piece of it, right. It's sort of like you have a flashlight, you know, and you're walking into a dark cave, you know, w w wherever you shine that you have perfect visibility, but. What [00:19:00] you need to do is light up the whole cave.
And now, now you really get the overview of what's going on, right? And there's also this delay of data problem that people need to get data in much faster. Um, you know, where they simply can't act , uh, as an organization on , uh, on what's going on. Now, the idea of data network effects , uh, is, is really important because, you know, once you're a, you're on a data cloud , um, you have access.
Um, you know, through the data cloud to external data and you provide access to other parties to, to your data. Now , you're, you're now engaged in all kinds of data, networking relationships , uh, with other parties. And it's sorta like being on Facebook all over again, but not with your data, the relationships , uh, really become , um, you know, the, the fabric of your , uh, of your data operations is pretty hard to a tear yourself away.
But you become more and more heavily , uh, invested in , uh, in building more of these data, networking relationships. Uh, we, we now visualize our data cloud and , uh, you know, it looks like one of those star Trek, death stars out there because it's just massive networking, [00:20:00] constellation, representation, visualization, or what's going on.
And you can drill down by industry and you can drill down by individual institutions. You can take any flavor you want from it. But I'd really start to hit on that data. Networking is a huge, huge thing. And it's growing in leaps and bounds. And of course that's what snowflake is. You know, we're, we're fanning those flames as, as hardness fast as , uh, as we know how we do that in any number of ways.
Steve Hamm: Yeah, that is really interesting. Just. Kind of set up a light bulb in my head. I mean, as, as snowflakes data cloud gets bigger and bigger and it pokes into more industries and into smaller companies, big companies, you're essentially gonna have , um, a map of the business world, a schema that is visualized that you can kind of use to, and other people who are members of the data cloud can use to kind of figure out , well, what the heck is going on in the world?
Frank Slootman: Yeah, it's a, it's a great point. Um, you know, we always , uh, you know, try to show prospects and [00:21:00] customers, look, this is what your data cloud could look like. Forget about the snowflake data cloud. It's ginormous. It has many players, but your version of the data cloud, what is your, what are your data now?
Working relationship. How do you consume data? How do you provide data? And a first of all, you know, people have sometimes a hard time articulating that or, or really , uh, you know, giving good guidance on that. But when they start thinking about it, all of a sudden, all the ideas start coming , well, we could do this, we could do that.
Uh, so it's a very organic thing , uh, that, that just after a while, it starts to feed on itself. And sometimes it just feeds on itself because of the presence of partnerships and data. Uh, that they can get access to that they'd never realized that they could write. It was very important for us to prime that pump right.
To, to really make it rich and available. And then, you know, people's imagination takes them from there, you know?
Steve Hamm: right now, you've got the stope, like the other marketplace and you've got the private exchange capabilities platforms. Have you seen any of your big enterprise customers? Kind of put the whole thing out there, their entire supply chain, they're all those relationships, their [00:22:00] entire distribution chain, or maybe even their marketing relationships.
Are they really putting the whole network out there?
Frank Slootman: Now people typically don't approach this, you know, with a, with a, with a giant scope and a giant switch , um, it is much more organic and piecemeal. You know, a lot of data sharing relationships are one Oh one. So you know, one on many, but you know, if you, it's very dynamic, if you look day to day, week to week, month to month, you see things changing so rapidly.
And the reason is it's really easy to do , uh, you know, setting up data, sharing relationships as a matter of minutes, right? I mean, you, you externalize the data in terms of. The the access , uh, you know, we developed the credentials for it and there it is, boom, you're off to the races, right? So there's really nothing there.
This was the problem, historically what API APIs and, and, and file transfers. It was just a horrendous undertaking to do the simplest things. And now it's just so easy that you can turn on and off at will, you know,
Steve Hamm: So, but in the future, do you see, you anticipate the companies will kind of [00:23:00] create these meshes, you know, of all of their relationships, is that.
Frank Slootman: Yeah. And that's really the whole notion of having your own data cloud , um, you know, grab any, any large, you know, multinational, global entity. What does their data cloud looks like, right? What are their, you know, their important data relationships that they're maintaining and, you know, in large institutions, you know, they have multiple divisions, multiple businesses.
So, you know, we see their data cloud done breaking down and data costs for different businesses, right? So it has many, many different layers , uh, to it, right. It's just not one thing. Um, and you can navigate it to your heart's content. You're going to follow all these relationships, but at the heart of the data cloud are.
Data relationships. It's a consumer and a provider. Right. And that's, that's really the building block off the data cloud and nourish a gazillion of bells. Yeah.
Steve Hamm: It seems like, you know, market analyst have been talking to for decades, even about companies being able to monetize their own data.
But it seems like the snowflake [00:24:00] data marketplace is really opening the flood Gates for that kind of activity. If you could talk about some of these new revenue streams that companies are developing on the marketplaces, I think that would be interesting to people.
Frank Slootman: In a marketplace, you know, so far has been a vehicle for people to , uh, to really discover and explore and test , uh, all kinds of different data. And it's been very successful with the monetization has been what we call out of band. So. That, that that happens through, you know, a separate set of conversations and especially for , uh, professional data providers, people that sell data for a living and obviously have business models, they have conduits, they have contracts, they have, they have models of, of, of, you know, what the currency is and, and how they price it.
Uh, and so on. Um, w w which is. Fine, but we are moving towards a, and we are literally in the middle of developing this towards a series of monetization models , uh, for people that are , um, that are on our snowflake, they can Mark place so they can transact in band instead of out of band. Um, we've signed up a bunch of , uh, what we call design partners.
You know, people that can really help [00:25:00] us do this correctly for their type of business. There will not be one model. And the thing about data is that it's not all created equal. And , um, so the sexually not a trivial thing. But we think if we do this and we do this correctly, I think it will unleash a whole new data economy if you will, a whole new data industry , um, because it becomes so effortless now, you know, to, to transact on data, which historically has been incredibly hard, not just because of the physical movement in the custody of the data.
But also about you, how do you buy this? How do you pay for that? You know, all those questions. Right? So the more , uh, you know, it becomes like an Apple , uh, you know, app store , uh, in many ways, right? I mean, we're absolutely used to that metaphor in a way we use that example a lot because they did it incredibly well.
Right. And the monetization model is incredibly well, but distribution model, the operational model is incredibly well-established. They behave really well. They use all the common services. Uh, so there's a lot of sorta analogs and corollaries , uh, that are useful for us to examine.
Steve Hamm: I get that. That'll make it much more efficient, but are you, are you seeing any companies that kind of like aren't , they, they don't, [00:26:00] they're not data companies, but there are companies that are figuring out how to monetize their data. Because they've now organized it and managed it. And they have these Sharon capabilities.
Is that starting to happen too, or? No?
Frank Slootman: Yeah, absolutely. Um, you know, our whole industries , uh, where people scratching their head, then they know their data is valuable, but they never knew, you know, how to even have the first conversation around monetization. I mean, we, we know people that own, you know, tens of billions of dollars in terms of fast food , uh, industry data, because they were on the restaurants.
Well, geez, what does it, what is that data worth , uh, in retail? Uh, they are incredibly rich in data. What is that worth? Um, obviously in banking, they know that our data, I mean, if you're on a card business, whether you are, you know, JP Morgan or you discover or capital one, the word was a big snowflake customer.
Uh, obviously they, they know they have extraordinarily valuable , uh, data sometimes, you know, Uh, they're interested in monetization and sometimes data gets leveraged for, for advertising purposes. So there's, there's, there's other models , uh, in play as well. It's not always [00:27:00] strictly data , uh, either. So, so yes, I mean a data monetization.
It can become a whole new line of business for , uh, you know, for, for industries that , uh, you know, that, that have not existed before because the platforms and the infrastructure just didn't exist. It was just too hard.
Steve Hamm: right. Frank, I've heard you talk about the fact that not all data is created equal.
What do you mean by that? Exactly. And, and how does the data cloud help separate the wheat from the chaff?
Frank Slootman: Well, um, I'll give you an example , uh, you know, you'll you look on our snowflake data marketplace , um, you know, one. One data set. And I mean, we were listing almost 400 data listings right now, which is quite a lot. Um, but there's one that stands out head and shoulders, and that happens to be a data set from star schema.
It's, it's very detailed, real, near real time , uh, incident fatality, data covert , um, you know, almost everybody in our customer base. Um, uses that data. Uh, that's what I mean by, you know, some data sets are incredibly impactful , uh, very globally applicable and everybody uses them and others are, are much more incidental industry specific or [00:28:00] even.
Uh, you know, specific to, to, to just one institution. So, you know, having the data size that the, the, the data sources , uh, that are, are incredibly impactful and globally applicable is a, is a huge deal. I mean, we can spend all day. Uh, loading our data cloud with data , um, but, and not move the dial and you can find just one and , uh, you know, that that completely changed the stinks.
Um, we also see institutions coming together and combining data and creating new data products, which is a very interesting phenomenon also because people then don't have to join the data themselves. They basically, you know, can, can access the data and it's already , uh, joined and overlaid and blended. So.
We're going to see all kinds of things that , uh, you know, that we may not have foreseen, you know, when we first got started on , uh, on this journey, just because it's possible.
Steve Hamm: interesting. It's kind of like the oil industry, you know, you sell crude oil and then somebody comes along and refines it into gasoline or jet fuel, and then somebody else uses it the same. Same stuff and , uh, makes him the plastic or something else. But that same, that same thing is happening in, in, in the data.
[00:29:00] Frank Slootman: Yeah , we, we actually have much more profound examples of that because, you know, we have this whole notion of data services or w w what our technical people referred to as installable applications, where we're really combining coat with data. And what that does is then, then data exists at multiple levels.
It exists in a raw form, and then it exists in, in, in successful levels of value added processing. Right. Uh, and this is how software developers can really add value to data and make the data much more useful, impactful, valuable to potential consumers. An example of that is for say, I have a , um, an MRI scan, which is an unstructured.
Document , uh, obviously, you know, today that just gets sent to a radiologist and they read it and, you know, they, they, they basically , uh, you know, w however they , uh, they evaluate it and write that up. Um, but suppose a bunch of software developers take that raw MRI scan , uh, you know, can compare that scan against industry data, historical data, other patient data , uh, combine that with genomic data clinical data.
[00:30:00] Demographic data, you know, whatever. However you want to go at this and basically read that scan, but give a much richer, comprehensive analysis of that skin. Then the radiologist could ever come up with on his own or her own that's value added processing. It's very, very precise value added processing.
That would be incredibly valuable. Wouldn't it? And , uh, you know, we're, we think we're gonna unleash an entire software industry around people that have specific expertise around specific data that can bring a higher level of value to these different data sources. And because you have a marketplace where people can discover the service and be monetized to service, right.
It can really be fueled , uh, in terms of, you know, it lowers the bar for people to build things and market things and sell things, you know, as well as transaction. No,
Steve Hamm: do you foresee a time when the data marketplace actually has some of those. Solutions providers kind of cataloged on it as well as an addition to the data providers.
Frank Slootman: Yeah, this is, this is, this is on its way. I mean, I will tell you that [00:31:00] in the middle of this year, when we do our big summit , uh, event, we're going to show 'em some of the leading edge examples in a lot of our connector technology, that's built by us as well as by partners. Uh, you know, we're going to use these models , uh, uh, to do exactly that.
Um, so we're, we're excited about where it is. This can all go. So it's not just data, right? You've got to think. You think about data in a much bigger context than just raw data, right? Because the way, if I can bring a value, incremental processing to the data , um, that starts to become really interesting because as a customer , I, I now don't have, I don't have to do that.
Right. I I don't have to be smart. You get to be smart for me on these, using your service. Right. And I'm producing the results of your service. I'm consuming the results of your service. And , uh, instead of me having to do that, so you haven't need subsequent levels of value , uh, you know, brought to the data , um, and, and, and having a good marketplace and, and model around it.
Uh, software developer has always been hard because, you know, where does the data come from, right? I mean, who wants to develop on dummy data? You really got to be very intimate , your understanding of the data to do something useful with it right now, this is the opportunity, you know,
Steve Hamm: Yeah. Hey, [00:32:00] that leads right into my next question, which is, it seems like one of the most powerful aspects of the data cloud is how it democratizes data and data analytics. And puts all that power in the hands of the real users, the business users, and you're giving people throughout organizations, whether it be a line manager all the way up to the CEO and the C-suite, you're giving them the tools to do the kind of sophisticated analytics that in the past only a data scientist could manage.
How is this changing organizations?
Frank Slootman: Well, it allows organizations to become data driven versus data informed. Uh, you know, we spoke earlier about, you know, the fact that, you know, in the last century we were running a report. And repopulating dashboards. That's, that's being data informed, right? This should, data goes to people and then we hope like hell that people do something useful.
What is , um, you know, when you go to dead ribbon , uh, it's end to end programmatic and digital, right? It's it's real time. It's highly scaled. Um, it's economical and it's incredibly precise in terms of [00:33:00] its , uh, predictive ability and , uh, It's it that's really what modern enterprises are going to be made up of.
They are just a series of , uh, of digital processes and the people working in enterprises are the people that built these processes and support them and, and, and resource in provision them and so on. But fundamentally the, the business process becomes end-to-end , uh, digital. So. Yeah. It all starts with blowing up these silos that we have built over over decades.
Right? My biggest caution to customers is don't build the silos of the future. Don't stare, steer the ship by its way, you know, think data cloud, you know, allow the data to come together because if we don't, you know, we're going to be very happy. You know, when we wake up , uh, some years into the future, you know,
Steve Hamm: I love to tear down those silos again. That's right. That's right. Hey, I don't want to get too far down into the weeds with technology in this conversation, but I think it's important to talk about snow park. And I understand this is a new capability in your data cloud that enables developers to write code in their language of choice for data integration, data [00:34:00] analytics, all those kinds of things.
Now, snowflake started off with a real focus on SQL. As it's a native language, very popular, easy to use all that kind of stuff. So why are you making this shift and why is it a significant change for snowflake and for the users?
Frank Slootman: Yeah, it's not so much a shift as it is an evolution and an expansion of our , uh, our scope area. Absolutely. Correct. Uh, I mean, w SQL is a great common denominator because almost anybody that's been around data has some level of proficiency. And , uh, you know, w we've seen that, right? I mean, uh, snowflake has been able to address the largest data States in the world, but also the smallest.
You know, where, where you have, you know, one person would a handful of files, ingesting data and running queries in a matter of, of, of minutes. Um, so that's been great, but there is another audience and that audience are actually developers. These are highly technical , uh, people, and they want to manipulate data.
Inside the preferred programming and scripting languages. Um, there is a [00:35:00] whole audience for that because that's really, you know, where the whole Hadoop crowd and the people that sort of came from that whole big data era, where they came from. Um, that's how they grew up and the shorter day are evolving forward.
To just kind of a model say they are going to be using Java and Python and other languages , uh, to manipulate data. Right. And there's a lots of artifacts in the language itself that are there for, for data manipulation. So we feel that as a data platform, as a data cloud, we. Absolutely positively have to enable that.
So what snow park really is , uh, is a strategy for the language runtimes to be inside our platforms and that really combines code with data. So in other words, it's not just a data platform is the combination of coding data on a single platform. And it really turns it into, you know, an applications , uh, platform rather than just a data platform, all in a back to the earlier conversation about higher levels of value added.
Processing , uh, uh, against data. This is very significant because our whole spectrum of [00:36:00] workloads , uh, expanse as a function of this , um, it's not the masses. It's certainly a smaller audience, but it's a very impactful and important audience. So, uh, very important, very meaningful , uh, not just to address those, those workloads, but a combination of code and data really changes , uh, the texture of the data cloud to be a much bigger thing, because we're not just going to be circulating data there and we're going to be circulating data applications there.
And , uh, that's obviously a huge universe. Yeah.
Steve Hamm: no, that's great. No, it we've been. Over the past year. It seems like the year that's lasted forever, but you know, we've been dealing with the pandemic. The economic crisis in the us and around the world. It seems like these things are on the path to lifting though. You don't want to jinx yourself. So how has snowflake been using data to operate in the midst of these uncertainties and, you know, keep your rapid growth going in spite of uncertainty of an in spite of uncertain times.
Frank Slootman: Well, we're, we're , uh, we're [00:37:00] a snowflake on snowflake , uh, company. Um, our, our CIO , uh, sunny BD , um, he has a whole presentation. Then we, we put him in front of , uh, customers all the time to talk about , uh, snowflake runs on , uh, on snowflake because all the data that, that, that. Comes from operational , uh, environments, whether it's Salesforce or, or on systems , uh, it moves into our own internal snowflake , uh, system and it's heavily, heavily analyzed , uh, uh, nonstop.
You know, I spoke earlier about , um, you know, how we guide , uh, wall street in the financial markets. Um, all that happens in there, but you know, obviously, you know, when you run a consumption model , uh, the analysis , uh, part of our business , uh, is huge, but , uh, our marketing people use it. Every part of our financial people use it, ourselves sales, operations, people use it.
There's not a single part. Of course, the engineering folks are all over this data. They actually support the platform. These are, these are , uh, you know, where we are. You know, we have data at our core. Um, you know, we are a snowflake centric , uh, enterprise in every way you can imagine it's just a natural reflex , uh, for us.
Um, so it's not like we really have to force [00:38:00] ourselves down that path.
Steve Hamm: Yeah, I want to drill deeper though. I want to understand how you and your management team use data to make decisions every day. I mean, you you've got these huge challenges of volatility of being a rapidly growing company, all the, all the uncertainties situation. How do you use data to help overcome those things?
Frank Slootman: Well, the, it depends on , uh, on which part of the company you're talking about. So, you know, the finance organization obviously is, is, is obsessed with understanding , uh, the consumption of our service , uh, will translate , uh, you know, into costs and revenue and all these kinds of things so that, so that we can guide ourselves as well as external.
Parties. And that that's a very significant thanks, but you know, for a sales organization, you know, they need to forecast their business, you know, based on, you know, what we're able to discern from the data, our marketing organizations have a huge demand capability , uh, and that is incredibly data driven.
Because obviously we're, we're constantly enriching our data and our, all our targeting and our marketing operations are informed and, and driven [00:39:00] by what we discerned from , uh, from the data, our engineering people , uh, that that's a whole universe unto itself because we are a service, you know, not, not a software product that we are a software, a service.
Maybe we're now processing close to a billion queries a day. Right? You can just imagine, you know, how much data we have , uh, you know, about our service and then it's across different cloud regions. And, you know, we have different, you know, operational load factors , uh, you know, performance, I mean, There's just an enormous amount of data.
And by the way, that's grown a hundred percent year on year. So, you know, we're, we're taking on Google ESC type scale here in terms of the number of queries that we are, that we are , uh, processing in the next , uh, couple of years. So it touches every part of our , uh, our business. There's sorta not one silver bullet here that we can point to that , uh, that is the share , uh, magic.
Uh, it just permeates itself , uh, through every, every function that we , uh, that we run.
Steve Hamm: I get it. You know, I talked before about how you and I did the first podcast about a year ago. Uh we've we've changed the format just [00:40:00] a little bit and right at the end, we'd like to ask a personal question, a little friendlier and softer. Now I know about your racing, but we talked about that. I think most people know about that.
I'm told that you also have some pretty cool dogs, including a brand new one who I think we've been hearing bark in the background a little bit. So your work-life is pretty intense. Do the dogs help you unwind or, you know, what, what, how do they add to your life?
Frank Slootman: Yeah. You know, I'm, I'm , uh, I shouldn't, I should say we, because my wife and I are both into this , uh, you know, we're a big animal welfare , uh, people , uh, in a lot of our charitable operations are. Both for domestic animals. So, you know, we've built shelters , uh, you know, in different places, you know, to really help different areas of the country become, you know, no kill zones for, you know, for, for animals and things of that sort.
And , uh, so we, we do a lot of that kind of work, but we also, we also do habitat for wild animals. So that's sort of, you know, been a focus , uh, for our , uh, Our charitable activities and, you know, I know number one, we have a lot of [00:41:00] affinity with it , um, because we were just like animals. Um, but secondly, it's also something that the government does not fund at all.
So it all needs to come from private sources. So yeah. Might as well be people , uh, like us , uh, in terms of our own animals. I, yeah, I mean, one of them is like one year old. The golden retriever is like 80 pounds, a big furball. Uh, and the other one is an eight month old female , uh, black lap under their best pals run around like crazy.
I always say to humanize me might be necessary from time to time. Um, but they're , uh, they're, they're a great company to have. They go everywhere with us. We never , uh, never leave him anywhere. So.
Steve Hamm: That sounds great. Now you live in the East Bay, so you have, you have some land around you, right?
Frank Slootman: Yes. Yes , we, we, we got room , uh, in, in the various places that we go. So they , uh, they , uh, they really liked that. So,
Steve Hamm: sounds great. Well, Frank has been great talking to you. It's, you know, it really, you know, talking again a year later, like this really makes me realize how far things have come. I mean, certainly for snowflake, but just the awareness of the data [00:42:00] club, the importance of data. How aware people in business and in the media are of how important this is.
I, I feel like I got on the ground floor with you guys and I'm seeing something really grow big. So it's kind of fun to watch from a bit of a distance here. So congratulations on this.
Frank Slootman: Thanks for your kind comments, Steven, during the discussion here.