Interview

Ravi Kumar S., President, Infosys, interviewing Frank Slootman, Chairman and CEO, Snowflake


00:00
00:00
  • Listen on
Share on
  • Ravi Kumar S.
    00:14
    Ravi Kumar S.

    Hello, everyone, welcome to this session for the next 20 minutes, with one of my favorite CEOs, Frank Slootman. Frank is the CEO and Chair of Snowflake, one of the most respected and accomplished CEOs in enterprise tech, very seasoned, best known name for taking large start-ups, incubating engines to well-oiled operations, scaled operations, building large go-to-market teams, building large installs of clients, and more importantly, generating a market-driven, value-driven narrative before taking companies to public. Frank has done this three times: as a CEO of Data Domain then as a CEO of ServiceNow from 2011 to 2017 and then as a CEO of Snowflake. He took these three companies into public with very successful IPOs. Snowflake in 2020 was the largest software IPO ever, and it was the largest company to ever double its value in the first opening day, reaching to seventy five billion dollars of market cap. Thank you, Frank, for talking to our teams, and it's an honor to host you at Infosys.

  • Frank Slootman
    01:28
    Frank Slootman

    Yeah, thanks, Ravi, we really appreciate you guys having us participate in the event. So good to be with you.

  • Ravi Kumar S.
    01:36
    Ravi Kumar S.

    Thank you, Frank. So, Frank, you know, I heard this from the social channels about your start at Snowflake. I believe you were bored at ServiceNow, you wanted to kind of semi retire, go back to sailing and yachts, which is your favorite hobby, I believe you wanted to get professional sailing into California. And then suddenly Snowflake happened and you got excited about it and you changed your path. So, what's the exciting thing at Snowflake, which really turned the trajectory for you to come back to the corporate world and create another successful venture?

  • Frank Slootman
    02:15
    Frank Slootman

    You know, that's actually a true story. I literally had no designs, no intentions on taking the helm at another company. I really didn't think how I could top the experience at ServiceNow. But I remember having meetings with Mike Speiser, who was the lead investor at Snowflake from time to time. We served another board together. And, you know, all of a sudden, they said, you know, what would it take for you to take the helm at Snowflake? And it totally surprised me because nothing had really led up to, you know, to that to that point, and it sort of gave me a rush of adrenaline and just thinking about the possibilities, you know, Snowflake is an important company. It's, they sort of threw away everything they knew about database management and they wanted to redesign completely for cloud scale computing, cloud architectures, really not carry anything forward. And as you know, our founders were long time Oracle architects and technologists, and they sort of had lived the problems of that generation of platform and really were eager to rethink, reimagine all those things. And it resulted in such a step function, if you will, in terms of technology platforms, that we see customers literally going, you know, orders of magnitude change, not incremental amounts of change, but orders of magnitude in terms of complex workloads and in just processes and so on. So, all of a sudden, your mind starts running. You know in terms of the possibilities and so you know, I'm a sucker for great products. That's pretty much what I get up for in the morning. And all of a sudden, you know, oops I was I was back there again, you know, and here we are almost exactly two years later right now.

  • Ravi Kumar S.
    04:17
    Ravi Kumar S.

    So, Frank thank you for that. Thank you for that context. In fact, I think in many ways you are the visionary behind data cloud or data exchanges. And the last time I spoke to you, you mentioned about these two steps. Landing workloads on the cloud by itself is a big virtue with untangling the legacy and everything else. And then data exchange part. Tell us a little bit about why you believe landing workloads on the cloud is such a big deal. A lot of our clients, joint clients talk about it. That by itself is is a big journey.

  • Frank Slootman
    04:53
    Frank Slootman

    Yeah, we often talk about that as the transition from Snowflake 1.0 to Snowflake 2.0, because it's very much a generational change. The first five years of Snowflake, once we started to really put product in the marketplace was exactly what you said. It's a modernization move, where we're taking on premise workloads to the cloud to give them obviously enormous scale advantages, massive performance advantages, utility models on and on and on. And as you well know, you know, customers for them, time is not their friend. They have to go sooner or later and sooner is better than later. So, there's a huge movement afoot to move a. data to the cloud and b. then start enacting these data-based migrations. So at least we're on the cloud, we are now running our workloads in a completely different fashion than before. But that is not the end of the story. You know, Snowflake 2.0 was really our vision of what else needs to happen as part of this transition. I mean, if you look at the landscape of cloud today, we see massive infrastructure clouds, right between AWS and Azure and GCP and so on. We have massive application clouds between Salesforce and SAP and ServiceNow and Workday and so on. And then we look at data and we don't have data clouds we have data massively fragmented and proliferated all over the place. Sometimes referred to them as silos, but our founder, Benoit Dageville refers to them as bunkers because it's so hard to penetrate, you know, they're hiding behind networks and system and application parameters and that's a real problem and the reason that it's a problem is, you know, historically, most of our analytics has been what we call in silos. So, we ran analytics on Salesforce and SAP and ServiceNow and so on. But data science doesn't look at the world that way, right? It's trying to understand data relationships that completely ignore data silos that could be completely diverse data sets that establish patterns as well as different types of data. And we have to enable that or we never are going to see the real problems of data science, machine learning and artificial intelligence. So, my word of caution to a lot of large customers is don't start building the silos of the future. You really have to think about erasing the silos, then you will completely unleash the potential and power of data science. And your data science teams will be grateful that you have the vision and the foresight to do that rather than just bring your legacy workloads to the cloud. That's just not enough.

  • Ravi Kumar S.
    07:35
    Ravi Kumar S.

    Thank you, Frank. In fact, every time I speak to clients you know recently I spoke to two retail big box companies who have 70 percent growth on e-commerce in the last one year for obvious reasons, and suddenly their e-commerce profitability is much lower than their on-prem physical stores because their shipping costs are higher, their returns are higher, and they don't do trade promotion on e-commerce as much as they do it in physical estates. So, they are telling us, take this data estates, monetize it and give me back money. Do you see that there will be a point when putting data warehouses on the cloud is no longer going to be a cost? It's actually going to be a revenue generating work stream, especially in the digital front ends?

  • Frank Slootman
    08:23
    Frank Slootman

    Yeah, I mean, there's definitely an awakening or a realization out there that a lot of businesses are sitting on incredibly valuable data. For example, you know, there's enterprises out there that are running tens of billions of fast food business right? What do we think the data is worth that these people are sitting on? I'm here to sell hamburgers or whatever. But it's actually the data that could be extraordinarily valuable. And retail, of course, running on incredibly low margins. But of course, sitting on incredibly potent data, especially when we can combine that data with other sources. I mean, the more we combine data, the more potent that data becomes. This is why the notion of a data cloud and data sharing is so instrumental in unlocking that value. Companies like you know like Facebook and Google they have a single cloud where the data is obviously incredibly powerful and has yielded enormous advertising efficiencies. But for the rest of us combining the data will unlock tremendous amounts of value. While in order to combine it, we need to have an architecture that lets us do that. So, all of a sudden, everybody is waking up like, hey, there's a whole other revenue and profit center here that we could have. Same is true in big finance, right? I mean, large financial institutions, the credit card businesses. So, we're incredibly data rich but our ability to monetize that and combine data is still in the early stages. So that's really what this is all about, right, is just unlocking the value of data and what you can do with it.

  • Ravi Kumar S.
    09:57
    Ravi Kumar S.

    In fact, at Infosys, we call it the network effect of the cloud, and we are hoping that the data exchanges of Snowflake give us that opportunity to straddle the value chain and ring learned AI models, not just give away data in raw form, but learned AI models which will help them do better operations. In fact, Frank, I wanted to elude a little bit about the industry transitions which are happening now. You know the energy sector is going through a transition where you have bidirectional flow of energy because of consumers becoming producers, cars are getting connected in fact, we did a very large deal in Europe where we are going to see connected cars by 2030, it’s half the car companies across the world are saying that they're going to go into connected cars and electrification. Manufacturing is going through servitization and platformization. 5G is moving Telcos from B2C to B2B. In fact, 5G will disrupt a lot of industries with compute at the point of consumption and storage of the cloud. We spoke about retailers, they're all straddling the digital estates and trying to monetize, so all the traditions of industries and if I may have a pivot, data has a pivotal role in it. Can you tell us a little bit about how you see this industry transitions, especially as you get to the other side of the crisis, we’ll see an accelerated path towards that transition?

  • Frank Slootman
    11:27
    Frank Slootman

    Yeah. Now you're touching on something that is incredibly disruptive, both in terms of the opportunity that it provides, but it's also highly threatening to traditional enterprise. Right when, when you talk to large banks these days, they talk about companies like Square and Stripe and so on, who are their competitors. Amazon is threatening to be, you know, not just in retail, obviously, but also in health care and logistics and other lines of business. So, we just don't know how traditional enterprises and industries, you know, what they're going to look like, because under the influence of digital transformation, that might be a whole new set of players on deck. And that's that obviously, it all starts with, you know, what are your systems, right? And what is your data? Data has become the beating heart of the digital enterprise. And, you know, we as a company have to, and you as well obviously, you have to meet people where they are, some of them are very advanced. They're born in the cloud. They have all these disciplines and that expertise, they were really founded on cloud and digital principles and there's others who don't even have the beginnings of a data platform or the beginnings of a data science discipline, they are enterprises, they're still running reports and populating dashboards. Nothing wrong with that. But obviously, this is far more disruptive and invasive than we can possibly imagine. So, we really need to sort of have our peripheral vision here and really understand what's all coming at us and how these changes are happening. It's so easy now for consumers to change from one service provider to another because it's just a digital experience, right? And just go from one login to another and all of a sudden, oops, you know, I'm in a completely different realm. So, health care, that's another area where so much of the value of health care itself is going to be delivered through data, not through science, per say, I mean science obviously is incredibly important but what can data bring to the science is just enormous. And you can talk to pharmaceutical companies, they notice right I'll give you an example, right if we have an MRI scan in our data cloud what can be done in terms of enhanced data processing to MRI cloud by combining it with demographic data, with historical data, with other patient data, clinical data. So that we get very, very precise, very, very, very fast, very value-added reads on the scans. So, in other words, it's going to be life-enhancing for patients because they're going to get much better diagnosis, much more precise, much quicker and much better therapies and treatments as a result. That's all because of data. That's not because of new science right. And we see these opportunities absolutely everywhere. And it also leads to a whole new realm in software development, because there will be people building these kinds of services, they are absolute experts at these types of applications right, more and more micro service oriented and they'll be used by lots and lots of different applications. So, we might be on the cusp of just an enormous renaissance in software development here, which we're very excited about, because the software development, it's all about scaling it down, where people with very minimal resources can create extraordinary value and then market that and transact on that, you know without needing massive companies and organizations to do it. The software has been our business to get into, you know, it takes a tremendous amount of resources.

  • Ravi Kumar S.
    15:00
    Ravi Kumar S.

    Absolutely Frank and those are great examples. In fact, I keep telling our clients that infrastructure is going to go big and software is going to be a modular. And you rightly pointed out. In fact, I had a slightly provocative question for you. The hyperscalers who are investing heavily on infrastructure, in fact between Google, Microsoft and AWS, they spend almost hundred fifty billion dollars in the last three years just on building public infrastructure on the cloud, which companies like Snowflake can ride on, how much do you see - and they don't make the economics of that is to actually build layers of value, above it and build those marketplaces where all of us can straddle - but where do you see they actually moving up the chain over a period of time and start to compete with best in class, data platforms like Snowflake. They already have, they already are competing with you on one side, but they are actually on other side they are partnering with you. What is your view about they seeing this is a serious business opportunity for themselves?

  • Frank Slootman
    16:06
    Frank Slootman

    Yeah, it's a question, obviously, that we get all the time. The reality certainly for AWS and Microsoft is that their posture is we need to foster innovation and choice on our platforms because if we're just going to try and build everything ourselves and the only option that's available to customers is the option from the cloud platform itself, they may end up in a not so best of breed situation because, you know, how can one company build the best of everything that just has never happened in the history of software. So, you just cannot have a strategy being Microsoft or Amazon that says your only choice will be what we offer you. And I think that both Amazon and Microsoft, based on my dealings with them, do have conviction. Does it sometimes lead to friction? Yes, but the friction is not that. I mean, we like to say, you know, steel sharpens steel. We make them better. And they have said so again and again. You make us move faster and we're going to have better products because you exist. Now, it's obviously incumbent upon us to run hard and fast and furious towards the future. But at the end of the day and I hate using that expression — everybody's going to be better off because everybody is going to end up with more innovation more choices than they otherwise would have. So, you know. We're looking for the win for the customer. If we're not the better solution, then we don't deserve to be there. And so, our whole game is to be that best option. That's what we bank on.

  • Ravi Kumar S.
    17:49
    Ravi Kumar S.

    Thank you, Frank, that is so well said. In fact, they're going to find this situation always in the market where you are going to compete and collaborate with many, many enterprises across the world. You know, how much do you believe data privacy regulation will come on the way as you as you put exchanges and you start to share consumer data across the spectrum of a value chain? Do you see governments now starting to react about it?

  • Frank Slootman
    18:15
    Frank Slootman

    Yeah, data governance is one of the single biggest topics in our conversations with customers. I mean, three years ago it was sort of an ancillary conversation. Now you see in large institutions that the data governance is really the gating between the data and the business. In other words, the organization before the business sees any data in any form, they got to be able to ensure that the data is secure and compliant from various privacy and other types of regulations. So, I went from sort of being adjacent to being in front of the supply chain, if you will. Data will not be getting to the business unless, you know, the institutions can guarantee and people can completely own the security and the privacy of that data. It's a huge topic of conversation and it's one of the reasons that Snowflake has had an incredible run because people; a platform like Snowflake, we know what's coming in, we know what people are doing, it's completely auditable right? And it's so important that you have a platform where we know where things are coming from, we know who is using the data, we know where they're putting it, we know what they're doing to it, those kinds of capabilities are going to be non-negotiable going forward. It's no longer the Wild West of the data lake you know where we're just dumping tons and tons of files on there, and people are sort of in the Wild West style having their way with the data and having a nice time. Those days will, that can't happen, and you talk to a large institution out there for whom their brands are sacrosanct. There is no way they will compromise or negotiate on their brand. And so, in other words, governance is number one and performance and scale and all these other great things that we're talking about here you know are coming after that. And that's sort of a change and, you know, we have invested a great deal on our ability to implement data clean room, the ability to share data without exposing data between the sharing partners. And that has enabled and we can anonymize the data and we can maintain the compliance; we never will relinquish custody of the data either. All these techniques that we now have as part of our architecture really enable different parties to share data without getting compromised from either a security or a compliance perspective. So, it's actually very promising that you need these platforms to enable these kind of data operations.

  • Ravi Kumar S.
    20:15
    Ravi Kumar S.

    Thank you, Frank. Thank you so much for your conversation today. Every time I've actually spoken to you, I’ve always come out enriched with more learning, and you always inspire us in the corporate world. Thank you again for talking to our teams at Infosys.

  • Frank Slootman
    21:05
    Frank Slootman

    You bet. Appreciate the opportunity. Thanks, Ravi.

  • Ravi Kumar S.
    21:08
    Ravi Kumar S.

    Thank you so much.