Databricks Data + AI Summit 2024 | George Fraser | Fivetran

theCUBE.net

Clips
News
More from Databricks Data + AI Summit 2024

George Fraser

CEO

Fivetran

How Fivetran is revolutionizing data movement, one connector at a time

Data pipelines are the lifeblood of enterprise operations, driving everything from business intelligence and analytics to artificial intelligence. By undertaking enterprise data movement operations, Fivetran Inc. aims to simplify data replication, AI workflows and cloud migrations.By simplifying the complexities of enterprise data movement, Fivetran empowers businesses to harness the full potential of their data with ease and efficiency, according to George Fraser (pictured), chief executive officer of Fivetran.Fivetran’s CEO George Fraser talks data replication

play_circle_outline Short-term value realization for customers starting their data journey with Fivetran

play_circle_outline Fivetran's niche market and future growth potential

play_circle_outline Expanding Data Lake Offerings: A Look into Fivetran's Vendor-Neutral Growth and 10x Scaling Hope

Info

George Fraser | Fivetran

George Fraser

CEO Fivetran

search

>> Good afternoon, nerd fam, and welcome back to Databricks Data and AI Summit. We're here in San Francisco and I am delighted because we happen to live here, which never happens for this tech crew traveling around the world. My name's Savannah Peterson here with theCUBE, but more importantly, very excited to be joined by George from Fivetran. George, thanks so much for being here.

George Fraser

>> Very nice to be back.

>> Thanks for bringing your svelte velvet too. You know how many blue Brooks Brothers suits I see sitting up here and how refreshing it is to see something that doesn't look like that?

George Fraser

>> This did not come from Brooks Brothers.

>> That definitely did not come from Brooks Brothers and sorry to everyone who wears the Brooks Brothers, but you can do a little better than that blue blazer. I just got to say it while I have the opportunity. George, I'm going to give you the opportunity essentially the CEO, this is your company, this is your baby. Give us the one-line pitch for Fivetran straight from the maestro's mouth.

George Fraser

>> Fivetran gets all of the data from all of your company's business systems, be they databases like Oracle, tools like Workday or Salesforce or ServiceNow or you name it. We get the data from all those systems and we replicate it and keep it up to date in your data platform of choice, for example, Databricks.

>> And I've heard rumor from other guests we've had from Fivetran on the show that it's super easy. Is that what your customers say? Is that what you've been a fan favorite?

George Fraser

>> Indeed. Fivetran is very easy. I think sometimes it's too easy for some people. Some people like to have knobs and fiddle, but yes.

>> In the dating world, maybe there's such thing as too easy. Is there such thing as too easy when it comes to tech?

George Fraser

>> Well, for some people who have been engineering their own data pipelines for a long time, this can be a little bit bittersweet to delegate responsibility to someone else-

>> Oh, it's giving up a little bit of control a little bit there.

George Fraser

>> You do give up a little bit of control, but the good news is we've been doing this for 10 years, we're very good at it. We can deal with very difficult data sources. We have 500 connectors. We can go to extreme throughput and extreme latency. We can adapt to change and we make it all very easy to use. And for most people, that is a good proposition.

>> I like it. So I'm no super expert but pretty plugged into the pulse here. Databricks made some announcements this morning that could be very much interpreted as a little bit competitive with you.

George Fraser

>> That is true. Our destination partners, and we love them, do like to make a hobby out of writing connectors and so Databricks had some announcements around some connectors that they have built. So the reason this happens is because the nature of the connectors business is it's actually pretty easy to build a connector that works some of the time, even most of the time. A good engineer can write a Postgres connector that works in the straightforward cases in a few days. The problem is, and we know this because we've been doing it for a long time, is it is extremely difficult to make a connector work 99.9% of the time. And for that reason we feel very secure in the future of our business. We have a breadth and performance capabilities and reliability that is unparalleled. We have 7,000 customers who use this for this and there's a good reason. And so I feel very comfortable about the future. But yes, it is true, all of our destination partners do like to build connectors now and then do a little competing with us.

>> Well, I think you have a great attitude about it because that 99.99999% reliability and consistency is exactly what customers are going to need. And speaking of customers, you work with some of the biggest customers in the world. You got a couple examples for me, right, Morgan Stanley perhaps?

George Fraser

>> Sure, yeah. We are very proud of our customer base. We have customers across all sizes, all industries. We started years ago more in the technology industry. Many of our early customers were in the Bay Area, but at this point we cover all industries and all sizes. So Morgan Stanley is a great example of a customer and a great example of why our business is so difficult to do at scale. So Morgan Stanley uses us to replicate databases. Probably their most important type of database is Sybase, of which they have many, but Morgan Stanley has a lot of subsidiaries, so they have many different types of database. These databases are... Some of them are on-prem, some of them are in the cloud.

>> I was going to say they've got to be all over the world.

George Fraser

>> Yeah, they're all over the place and these databases generate huge volumes of data and they have huge requirements around latency. And Fivetran's ability to reliably replicate data at scale and when the databases are under pressure, which is when things really get hard, is why they choose us to do this, as well as our breadth of support that we have connectivity to all of their systems of record.

>> And you've got a few other household names. Do you mind dropping them?

George Fraser

>> Sure, yeah. Another great customer example is LVMH, a company that I think is well-known, speaking of velvet blazers. Although I don't know if they make one. I really-

>> I love that you were going for fashion brand there though. I was like, "Bring it. Let's do it."

George Fraser

>> Yeah, yeah. I need to get something from them to wear to things like this.

>> Hopefully they're listening right now and just fit you and me, while we're at it just for the record.

George Fraser

>> Exactly, exactly. So LVMH, they have 75 subsidiaries and they've chosen Fivetran as their data movement tool of choice across the entire global company. And so you can imagine the diversity of data sources involved in creating a single version of the truth for a company like LVMH. There are databases, there are marketing data sources, there are inventory data sources. It really runs the gamut of everything you can imagine and it's going to take multiple years to implement all of these, but we're very excited to do it and they're very excited to work with us. In order to get that customer 360, where you can connect all the dots, all the data points, in order to do that, you need to have that breadth of coverage and FiveTran has unparalleled breadth.

>> I love it. And I think what you're talking about as an end consumer of their product offerings and their retail, you will be able to have that hyper-customized experience that we've always kind of been dreaming of when it comes to our fashion or our fit or whatever our luxury goods might be.

George Fraser

>> Absolutely. It's funny, yesterday I happened to buy a new tennis shirt at Vori and I was trying to use the checkout process to explain to my boyfriend the concept of master data management and how this works. Like, "Well, this company will probably want to figure out that I'm the same George Fraser as the one who bought something online, and how does that all work behind the scenes?"

>> Okay. I adore that. You're at the POS system. We're giving just a quick little hyper-customization, AI chit-chat to your boyfriend. Well, that's the reality though. And maybe we'll get rewarded or prompted with something really cool there or who knows, or at least they'll know our shopping patterns too to a really minute degree, which may or may not be a good thing for our wallets throughout the course of that experience. I'm curious when it comes to, so you talked a lot about... I mean that is a really broad scale implementation. I was going to ask how long it takes. You mentioned a couple of years to fully roll out, which makes perfect sense given the number of systems that you're integrating here, what are some of the short-term value realization that these customers will have on the front end of starting this journey to help momentum stay through all that?

George Fraser

>> Well, one of the great things about Fivetran is because it is so automated and because our approach to data replication is really mirroring, we create a mirror image of all of your systems of record inside your data platform as though the data just lives there. And that means that you can very quickly set up a lot of different systems, get a lot of data into your central repository even without having to know exactly what you're bringing in. And you can start with your most important question, whatever that is for you as a business. So for a technology startup, it might be about funnel optimization. For a consumer packaged goods company like Procter and Gamble, it's about inventory management and making sure that the right goods are in the right warehouses. So there's a nice iterative element of implementing Fivetran. The total implementation may never end because the business is always changing, but you can really start with your most important question and then go from there. Rather than do some massive enterprise data management single schema project, which by the time it finishes, the whole business will be completely different and it will be unusable.

>> And everyone's exhausted and isn't sure the tool is actually going to solve any real problem.

George Fraser

>> Yes, everyone's already left for other jobs.

>> Yeah, right. I could just see those meetings. We've all been in them as consultants or outside forces. It's brutal. And maybe this is obvious, maybe it's not. Do you feel like large enterprise know what their most important question is or is that a little bit of a competing challenge across different business units?

George Fraser

>> Well, if it's a large enterprise, they're not going to have one most important question, but you can start with the first batch of key questions. And I mean sometimes they do. I think of Geico for example, as a customer and their most important question is the effectiveness of their ad spend hour by hour, even minute by minute during things like the Super Bowl. So there is an example of a big company which very clearly has one most important question.

>> I kind of like when it's that simplified and you can really see if you're toggling it. Obviously a million of complex things going on there, but it's cool to think about. All right, I'm going to shift gears and this is a little bit of a stretch of a question, but you got to roll with me because I think your company background is really interesting, especially living here in the Silicon Valley. A lot of very similar founder stories sometimes I've been here for too long, I'm clearly a little jaded. Lots of talks about lakes today and announcements about lakes. It turns out that you and your co-founder, Taylor's family met over a hundred years ago building lake cabins. All right. Stay with me, audience. I know you think I'm nuts right now, but I promise that we're going for something. I think that's rather serendipitous given the conversation, but I got to ask because I've done business with friends and I'll keep my opinion out of this. Would you recommend starting a company with a friend like that?

George Fraser

>> It's interesting. Way back when we started the company, some people warned me about starting businesses with friends and they told me it was a bad idea. It has worked out really well for us.

>> I mean, clearly you're sitting here right now, so it's a good sign.

George Fraser

>> Yeah, I think we've worked incredibly well together. My co-founder and I are very different and I think that helps.

>> Yeah, I agree.

George Fraser

>> And it has really been a great part of my life working with him to build this company together. So for me and I think for him as well, it's been fantastic. Your mileage may vary.

>> Well, no, I think that's great and it's actually really heartwarming to hear. There's a lot of AI founders out there, most of them concentrated right here in the Silicon Valley in our backyard. What would be your advice to those who are early on the founder journey? You've been in it for over a decade, so you've seen some of the ups and downs of this hype cycle. What's your advice to the young founders today?

George Fraser

>> Well, if you are like we were and you're founders without a pedigree, you have to claw your way up from zero and you will have no credibility with anyone. And you have got to find a problem that somebody needs solved so badly that they will buy a solution from this tiny company that is started by two people, not from the industry, with no credentials. And you need to be very real about that. If you're that kind of founder about, you've got to find something that for some segment of the market is so discontinuous that it will get you from zero to one. It's a little different if you're like Databricks and you're coming out of a famous lab at a university or if you are like Snowflake and you're started by people with long pedigree in industry. They play a little bit of a different game. But you asked about a lot of these startup founders-

>> That's not what I'm asking about, yeah....

George Fraser

>> today are people like I was with no pedigree. And you have a very specific challenge you have to solve when you're in that situation.

>> It's know your niche and know it to the tiniest little note of that niche.

George Fraser

>> The key though is that niche needs to be the beginning of something that grows very big, otherwise you'll have a very small company. And for Fivetran, that initial-

>> I was just going to ask....

George Fraser

>> that initial niche was people who prioritize time to value above everything else. Because Fivetran is so automated, you really can set up a full-fledged data warehouse in an hour with Fivetran if your data sets are not too big. And for some very small companies, technology startups in the Bay Area, that was a real thing that they wanted to do. They wanted it up and running today and they would have maybe one person who was doing everything from managing the platform to delivering reports to the executive team. And so our extraordinary degree of automation was what was so discontinuous about us. And now as the years have gone on, we sell to all kinds of companies. We have other properties like our performance and our reliability that we've developed over time. The automation is still very valuable though. It is valuable in a slightly different way at big companies. For small companies, the reason automation is so valuable is because it's so fast to set up. Big companies, fast to set up is nice but is not actually that important for them. It's what automation does for big companies is it makes Fivetran very easy to maintain. Maintenance is the nightmare at big companies.

>> No headaches once we're on.

George Fraser

>> Yeah, once you're on, it's very set it and forget it. And people at big companies love that about us, but definitely, we started in a niche, which is weird people who want to have a data warehouse set up tonight. That was kind of Fivetran's original, original customer base and then we grew out from there.

>> I love that. Where are you going to grow out now that you're at 7,000 customers, where are you going to blossom to from here then?

George Fraser

>> Data replication is a huge business. It's kind of a hidden business. It's like plumbing. People don't tend to talk about it. It's not the thing people tend to bring up.

>> The sexy dinner conversation?

George Fraser

>> It's not. It's not.

>> Although you're doing it at the POS moment buying clothes,

George Fraser

>> There you go. Well, I try to connect it to the real world, but it is a giant. It is ubiquitous. Every company above a certain size and age will have a central data platform where they replicate all their systems of record and they will either need to use Fivetran or they will need to build it themselves is what actually most people do today. So what the next few years look like is just continuing to persuade people, many of whom were here at this conference who are accustomed to writing code to build data pipelines and doing it all themselves that like, "Hey, for the replication piece, it just really doesn't make any sense anymore." We are so good at it. We are reasonably priced. I think people will look back and it will seem as ridiculous in retrospect to write your own connectors as to write your own email server, which is a thing people used to do. I think-

>> I was alive then still for the record, yes. Yeah.

George Fraser

>> That's the journey we are on is convincing everyone to maybe put down your keyboard and let someone else solve the problem of replicating your Postgres database into your Databricks.

>> I love that. Put down the keyboard. I'm ready to be done with my keyboard. You all ready to be done with your keyboard? I'm looking at the production guys. Yeah, I think they're definitely ready to be done with their keyboards over there. All right. I got two more questions for you and this one's kind of cool. So when I think of charismatic tech CEO, I don't always think of PhD neuroscientist. This is your background. How did you make the leap from studying that to knowing you wanted to be an entrepreneur? You talk about being a non-pedigree founder. You're clearly an educated founder, but there must have been some sort of pivot that catalyzed you to want to come do this.

George Fraser

>> It is true. Before I was a tech CEO, I was a scientist and it was a different chapter of my life. I loved the work I did when I was a scientist. I'm very proud of it. I think I contributed in a small way to the enterprise of science, which is the most important thing I think that humanity does. I decided not to stay in academia and I joined a biotech company after graduate school, started by two college friends actually, and I was there for a couple of years. And I had the idea of entrepreneurship going way back to being a kid. I had sort of set it aside and taken a different path for eight years, but then I ultimately decided to go back to that. I joked that I wasn't brave enough to start a biotech company. So I started a software company and there was a process of evolution. Data integration was not our original idea. So there was a process at the beginning there, but we were very determined and we were very patient and we were very smart. I sometimes joke that the secret of Fivetran's success is that we are the smartest people who are willing to work on this problem.

>> I love the humility and the honesty there. So your advice to those founders or to someone in school would be to go for it, yeah?

George Fraser

>> I mean, I think there's absolutely nothing wrong in changing careers. I think sometimes people think that their whole life journey has to make sense. Everything has to be leading into the next thing and it absolutely does not. And you're not too old to change direction. And the people who have been doing it for 10 years, experience is less valuable than people think. You can catch up.

>> Oh, I love that now. All right, final question for you, just so that we can use the highlight before we have you on the show next time. We'll call it a year from now. It doesn't have to be this show. We'll see. I suspect we'll see you before then. But what do you hope to be able to say a year from now that you can't quite say today?

George Fraser

>> I hope that our data lake offering is 10 times bigger than it is today. It is one of our fastest growing platforms, and there's a lot of talk about data lakes right now, including at this conference. One of the really cool things about data lakes is that you can build a truly vendor neutral data lake and still use the platforms you love like Databricks, but you can control your data to an even higher degree. And Fivetran can help you do that. And it's growing really well and I hope a year from now, it'll be 10 times bigger than it is today.

>> Awesome. And then we've got some custom fits from your fabulous customers perhaps.

George Fraser

>> Exactly.

>> Yeah. George, thank you so much for this. This was absolutely fantastic. I loved it. We covered everything from the future to fashion and we can't wait to have you back on theCUBE.

George Fraser

>> Very nice to be with you.

>> Yeah. And thank all of you for tuning in wherever you might be on this beautiful rock. We're in San Francisco, California at Databricks Data and AI Summit. My name's Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.

George Fraser | Fivetran

search

>> Good afternoon, nerd fam, and welcome back to Databricks Data and AI Summit. We're here in San Francisco and I am delighted because we happen to live here, which never happens for this tech crew traveling around the world. My name's Savannah Peterson here with theCUBE, but more importantly, very excited to be joined by George from Fivetran. George, thanks so much for being here.

George Fraser

>> Very nice to be back.

>> Thanks for bringing your svelte velvet too. You know how many blue Brooks Brothers suits I see sitting up here and how refreshing it is to see something that doesn't look like that?

George Fraser

>> This did not come from Brooks Brothers.

>> That definitely did not come from Brooks Brothers and sorry to everyone who wears the Brooks Brothers, but you can do a little better than that blue blazer. I just got to say it while I have the opportunity. George, I'm going to give you the opportunity essentially the CEO, this is your company, this is your baby. Give us the one-line pitch for Fivetran straight from the maestro's mouth.

George Fraser

>> Fivetran gets all of the data from all of your company's business systems, be they databases like Oracle, tools like Workday or Salesforce or ServiceNow or you name it. We get the data from all those systems and we replicate it and keep it up to date in your data platform of choice, for example, Databricks.

>> And I've heard rumor from other guests we've had from Fivetran on the show that it's super easy. Is that what your customers say? Is that what you've been a fan favorite?

George Fraser

>> Indeed. Fivetran is very easy. I think sometimes it's too easy for some people. Some people like to have knobs and fiddle, but yes.

>> In the dating world, maybe there's such thing as too easy. Is there such thing as too easy when it comes to tech?

George Fraser

>> Well, for some people who have been engineering their own data pipelines for a long time, this can be a little bit bittersweet to delegate responsibility to someone else-

>> Oh, it's giving up a little bit of control a little bit there.

George Fraser

>> You do give up a little bit of control, but the good news is we've been doing this for 10 years, we're very good at it. We can deal with very difficult data sources. We have 500 connectors. We can go to extreme throughput and extreme latency. We can adapt to change and we make it all very easy to use. And for most people, that is a good proposition.

>> I like it. So I'm no super expert but pretty plugged into the pulse here. Databricks made some announcements this morning that could be very much interpreted as a little bit competitive with you.

George Fraser

>> That is true. Our destination partners, and we love them, do like to make a hobby out of writing connectors and so Databricks had some announcements around some connectors that they have built. So the reason this happens is because the nature of the connectors business is it's actually pretty easy to build a connector that works some of the time, even most of the time. A good engineer can write a Postgres connector that works in the straightforward cases in a few days. The problem is, and we know this because we've been doing it for a long time, is it is extremely difficult to make a connector work 99.9% of the time. And for that reason we feel very secure in the future of our business. We have a breadth and performance capabilities and reliability that is unparalleled. We have 7,000 customers who use this for this and there's a good reason. And so I feel very comfortable about the future. But yes, it is true, all of our destination partners do like to build connectors now and then do a little competing with us.

>> Well, I think you have a great attitude about it because that 99.99999% reliability and consistency is exactly what customers are going to need. And speaking of customers, you work with some of the biggest customers in the world. You got a couple examples for me, right, Morgan Stanley perhaps?

George Fraser

>> Sure, yeah. We are very proud of our customer base. We have customers across all sizes, all industries. We started years ago more in the technology industry. Many of our early customers were in the Bay Area, but at this point we cover all industries and all sizes. So Morgan Stanley is a great example of a customer and a great example of why our business is so difficult to do at scale. So Morgan Stanley uses us to replicate databases. Probably their most important type of database is Sybase, of which they have many, but Morgan Stanley has a lot of subsidiaries, so they have many different types of database. These databases are... Some of them are on-prem, some of them are in the cloud.

>> I was going to say they've got to be all over the world.

George Fraser

>> Yeah, they're all over the place and these databases generate huge volumes of data and they have huge requirements around latency. And Fivetran's ability to reliably replicate data at scale and when the databases are under pressure, which is when things really get hard, is why they choose us to do this, as well as our breadth of support that we have connectivity to all of their systems of record.

>> And you've got a few other household names. Do you mind dropping them?

George Fraser

>> Sure, yeah. Another great customer example is LVMH, a company that I think is well-known, speaking of velvet blazers. Although I don't know if they make one. I really-

>> I love that you were going for fashion brand there though. I was like, "Bring it. Let's do it."

George Fraser

>> Yeah, yeah. I need to get something from them to wear to things like this.

>> Hopefully they're listening right now and just fit you and me, while we're at it just for the record.

George Fraser

>> Exactly, exactly. So LVMH, they have 75 subsidiaries and they've chosen Fivetran as their data movement tool of choice across the entire global company. And so you can imagine the diversity of data sources involved in creating a single version of the truth for a company like LVMH. There are databases, there are marketing data sources, there are inventory data sources. It really runs the gamut of everything you can imagine and it's going to take multiple years to implement all of these, but we're very excited to do it and they're very excited to work with us. In order to get that customer 360, where you can connect all the dots, all the data points, in order to do that, you need to have that breadth of coverage and FiveTran has unparalleled breadth.

>> I love it. And I think what you're talking about as an end consumer of their product offerings and their retail, you will be able to have that hyper-customized experience that we've always kind of been dreaming of when it comes to our fashion or our fit or whatever our luxury goods might be.

George Fraser

>> Absolutely. It's funny, yesterday I happened to buy a new tennis shirt at Vori and I was trying to use the checkout process to explain to my boyfriend the concept of master data management and how this works. Like, "Well, this company will probably want to figure out that I'm the same George Fraser as the one who bought something online, and how does that all work behind the scenes?"

>> Okay. I adore that. You're at the POS system. We're giving just a quick little hyper-customization, AI chit-chat to your boyfriend. Well, that's the reality though. And maybe we'll get rewarded or prompted with something really cool there or who knows, or at least they'll know our shopping patterns too to a really minute degree, which may or may not be a good thing for our wallets throughout the course of that experience. I'm curious when it comes to, so you talked a lot about... I mean that is a really broad scale implementation. I was going to ask how long it takes. You mentioned a couple of years to fully roll out, which makes perfect sense given the number of systems that you're integrating here, what are some of the short-term value realization that these customers will have on the front end of starting this journey to help momentum stay through all that?

George Fraser

>> Well, one of the great things about Fivetran is because it is so automated and because our approach to data replication is really mirroring, we create a mirror image of all of your systems of record inside your data platform as though the data just lives there. And that means that you can very quickly set up a lot of different systems, get a lot of data into your central repository even without having to know exactly what you're bringing in. And you can start with your most important question, whatever that is for you as a business. So for a technology startup, it might be about funnel optimization. For a consumer packaged goods company like Procter and Gamble, it's about inventory management and making sure that the right goods are in the right warehouses. So there's a nice iterative element of implementing Fivetran. The total implementation may never end because the business is always changing, but you can really start with your most important question and then go from there. Rather than do some massive enterprise data management single schema project, which by the time it finishes, the whole business will be completely different and it will be unusable.

>> And everyone's exhausted and isn't sure the tool is actually going to solve any real problem.

George Fraser

>> Yes, everyone's already left for other jobs.

>> Yeah, right. I could just see those meetings. We've all been in them as consultants or outside forces. It's brutal. And maybe this is obvious, maybe it's not. Do you feel like large enterprise know what their most important question is or is that a little bit of a competing challenge across different business units?

George Fraser

>> Well, if it's a large enterprise, they're not going to have one most important question, but you can start with the first batch of key questions. And I mean sometimes they do. I think of Geico for example, as a customer and their most important question is the effectiveness of their ad spend hour by hour, even minute by minute during things like the Super Bowl. So there is an example of a big company which very clearly has one most important question.

>> I kind of like when it's that simplified and you can really see if you're toggling it. Obviously a million of complex things going on there, but it's cool to think about. All right, I'm going to shift gears and this is a little bit of a stretch of a question, but you got to roll with me because I think your company background is really interesting, especially living here in the Silicon Valley. A lot of very similar founder stories sometimes I've been here for too long, I'm clearly a little jaded. Lots of talks about lakes today and announcements about lakes. It turns out that you and your co-founder, Taylor's family met over a hundred years ago building lake cabins. All right. Stay with me, audience. I know you think I'm nuts right now, but I promise that we're going for something. I think that's rather serendipitous given the conversation, but I got to ask because I've done business with friends and I'll keep my opinion out of this. Would you recommend starting a company with a friend like that?

George Fraser

>> It's interesting. Way back when we started the company, some people warned me about starting businesses with friends and they told me it was a bad idea. It has worked out really well for us.

>> I mean, clearly you're sitting here right now, so it's a good sign.

George Fraser

>> Yeah, I think we've worked incredibly well together. My co-founder and I are very different and I think that helps.

>> Yeah, I agree.

George Fraser

>> And it has really been a great part of my life working with him to build this company together. So for me and I think for him as well, it's been fantastic. Your mileage may vary.

>> Well, no, I think that's great and it's actually really heartwarming to hear. There's a lot of AI founders out there, most of them concentrated right here in the Silicon Valley in our backyard. What would be your advice to those who are early on the founder journey? You've been in it for over a decade, so you've seen some of the ups and downs of this hype cycle. What's your advice to the young founders today?

George Fraser

>> Well, if you are like we were and you're founders without a pedigree, you have to claw your way up from zero and you will have no credibility with anyone. And you have got to find a problem that somebody needs solved so badly that they will buy a solution from this tiny company that is started by two people, not from the industry, with no credentials. And you need to be very real about that. If you're that kind of founder about, you've got to find something that for some segment of the market is so discontinuous that it will get you from zero to one. It's a little different if you're like Databricks and you're coming out of a famous lab at a university or if you are like Snowflake and you're started by people with long pedigree in industry. They play a little bit of a different game. But you asked about a lot of these startup founders-

>> That's not what I'm asking about, yeah....

George Fraser

>> today are people like I was with no pedigree. And you have a very specific challenge you have to solve when you're in that situation.

>> It's know your niche and know it to the tiniest little note of that niche.

George Fraser

>> The key though is that niche needs to be the beginning of something that grows very big, otherwise you'll have a very small company. And for Fivetran, that initial-

>> I was just going to ask....

George Fraser

>> that initial niche was people who prioritize time to value above everything else. Because Fivetran is so automated, you really can set up a full-fledged data warehouse in an hour with Fivetran if your data sets are not too big. And for some very small companies, technology startups in the Bay Area, that was a real thing that they wanted to do. They wanted it up and running today and they would have maybe one person who was doing everything from managing the platform to delivering reports to the executive team. And so our extraordinary degree of automation was what was so discontinuous about us. And now as the years have gone on, we sell to all kinds of companies. We have other properties like our performance and our reliability that we've developed over time. The automation is still very valuable though. It is valuable in a slightly different way at big companies. For small companies, the reason automation is so valuable is because it's so fast to set up. Big companies, fast to set up is nice but is not actually that important for them. It's what automation does for big companies is it makes Fivetran very easy to maintain. Maintenance is the nightmare at big companies.

>> No headaches once we're on.

George Fraser

>> Yeah, once you're on, it's very set it and forget it. And people at big companies love that about us, but definitely, we started in a niche, which is weird people who want to have a data warehouse set up tonight. That was kind of Fivetran's original, original customer base and then we grew out from there.

>> I love that. Where are you going to grow out now that you're at 7,000 customers, where are you going to blossom to from here then?

George Fraser

>> Data replication is a huge business. It's kind of a hidden business. It's like plumbing. People don't tend to talk about it. It's not the thing people tend to bring up.

>> The sexy dinner conversation?

George Fraser

>> It's not. It's not.

>> Although you're doing it at the POS moment buying clothes,

George Fraser

>> There you go. Well, I try to connect it to the real world, but it is a giant. It is ubiquitous. Every company above a certain size and age will have a central data platform where they replicate all their systems of record and they will either need to use Fivetran or they will need to build it themselves is what actually most people do today. So what the next few years look like is just continuing to persuade people, many of whom were here at this conference who are accustomed to writing code to build data pipelines and doing it all themselves that like, "Hey, for the replication piece, it just really doesn't make any sense anymore." We are so good at it. We are reasonably priced. I think people will look back and it will seem as ridiculous in retrospect to write your own connectors as to write your own email server, which is a thing people used to do. I think-

>> I was alive then still for the record, yes. Yeah.

George Fraser

>> That's the journey we are on is convincing everyone to maybe put down your keyboard and let someone else solve the problem of replicating your Postgres database into your Databricks.

>> I love that. Put down the keyboard. I'm ready to be done with my keyboard. You all ready to be done with your keyboard? I'm looking at the production guys. Yeah, I think they're definitely ready to be done with their keyboards over there. All right. I got two more questions for you and this one's kind of cool. So when I think of charismatic tech CEO, I don't always think of PhD neuroscientist. This is your background. How did you make the leap from studying that to knowing you wanted to be an entrepreneur? You talk about being a non-pedigree founder. You're clearly an educated founder, but there must have been some sort of pivot that catalyzed you to want to come do this.

George Fraser

>> It is true. Before I was a tech CEO, I was a scientist and it was a different chapter of my life. I loved the work I did when I was a scientist. I'm very proud of it. I think I contributed in a small way to the enterprise of science, which is the most important thing I think that humanity does. I decided not to stay in academia and I joined a biotech company after graduate school, started by two college friends actually, and I was there for a couple of years. And I had the idea of entrepreneurship going way back to being a kid. I had sort of set it aside and taken a different path for eight years, but then I ultimately decided to go back to that. I joked that I wasn't brave enough to start a biotech company. So I started a software company and there was a process of evolution. Data integration was not our original idea. So there was a process at the beginning there, but we were very determined and we were very patient and we were very smart. I sometimes joke that the secret of Fivetran's success is that we are the smartest people who are willing to work on this problem.

>> I love the humility and the honesty there. So your advice to those founders or to someone in school would be to go for it, yeah?

George Fraser

>> I mean, I think there's absolutely nothing wrong in changing careers. I think sometimes people think that their whole life journey has to make sense. Everything has to be leading into the next thing and it absolutely does not. And you're not too old to change direction. And the people who have been doing it for 10 years, experience is less valuable than people think. You can catch up.

>> Oh, I love that now. All right, final question for you, just so that we can use the highlight before we have you on the show next time. We'll call it a year from now. It doesn't have to be this show. We'll see. I suspect we'll see you before then. But what do you hope to be able to say a year from now that you can't quite say today?

George Fraser

>> I hope that our data lake offering is 10 times bigger than it is today. It is one of our fastest growing platforms, and there's a lot of talk about data lakes right now, including at this conference. One of the really cool things about data lakes is that you can build a truly vendor neutral data lake and still use the platforms you love like Databricks, but you can control your data to an even higher degree. And Fivetran can help you do that. And it's growing really well and I hope a year from now, it'll be 10 times bigger than it is today.

>> Awesome. And then we've got some custom fits from your fabulous customers perhaps.

George Fraser

>> Exactly.

>> Yeah. George, thank you so much for this. This was absolutely fantastic. I loved it. We covered everything from the future to fashion and we can't wait to have you back on theCUBE.

George Fraser

>> Very nice to be with you.

>> Yeah. And thank all of you for tuning in wherever you might be on this beautiful rock. We're in San Francisco, California at Databricks Data and AI Summit. My name's Savannah Peterson. You're watching theCUBE, the leading source for enterprise tech news.