Blog

Company Updates

Exclusive Insights: An Interview with Michel Tricot at the Snowflake Summit 2023

Gain exclusive insights from Michel Tricot at Snowflake Summit 2023, unlocking invaluable insider perspectives.

COMPANY UPDATES

June 27, 2023

15 min read

Michel Tricot

Summarize with AI:

Lisa: Good evening everyone. Welcome back to theCube Live Coverage Day one Snowflake Summit 2023 from Caesars Forum in Las Vegas. It's Happy Hour. It's Happy Hour in Las Vegas. It's been happy hour for a couple hours, but Dave and I are here. We're having some amazing conversations today. Up next, a conversation with the co-founder and CEO of Airbyte.

Lisa: We're gonna be talking about open source. Data integration. Dave, a lot of supported connectors from any one vendor has typically plateaued out at around 150 until now, and we're gonna be talking about

that with Airbyte.

Dave: Oh, I didn't know that stat. Yes.

Lisa: Okay. Yes, please welcome Michel Tricot, co-founder and CEO of Airbyte.

Lisa: Michel, welcome to theCube.

Michel Tricot: Thank you for having me. Tell us a little bit about Airbyte the catalyst to launch the company, what it is that you do for your customers. Yeah, so I've been working in the data space for the past 15 years and. Every time I've been working on data, [00:01:00] I've noticed that teams always do the same thing, which is, and they always have the same problem, which is, how do I have, how do I access data across all the system that I have, whether it's a Salesforce silo, a HubSpot silo, a Google ad silo, a database, et cetera, and every team is redoing the same thing over and over again, which is building a connector.

Michel Tricot: Pulling data out of the system and centralizing that data somewhere. And with modern warehouses like Snowflake, that has become even more of a need for companies. And early, early 2020, with my co-founder, we started Airbyte and our mission was really how do we address the long tail of data connectors? So not just the top 20 that everybody has, it's also all these other places, all these other connectors and.

Michel Tricot: We decided to go for something open source so that people can, one, build their own connectors, maintain connectors But also and grow that long tail very fast.

Lisa: As I mentioned in the [00:02:00] intro and as some of the research I was doing on Airbyte, the number of supported connectors is typically plateaued out at 150 because the ROI isn't simply there to justify more.

Lisa: Why is Airbyte different? How did you get past that plateau?

Michel Tricot: Yeah, so we do it because we are not the only one maintaining connector. So we do a part of it, but then we. We work with our community, we work with our customers to actually share the load and the burden of the this maintenance because the moment you start using a connector, you have an incentive to make sure that this connector always works.

Michel Tricot: And that's the way of spreading the load. And instead of having every single thing doing it, you just have one team that does it. And suddenly the rest of the teams I have access to the same quality of connectors and the same type of connectors.

Dave: So connector creep. Yeah. Kills you right now, Michel.

Dave: This has been tried before. I think if I think back to the big data days, I think of a company like had [00:03:00] adapt. They were trying to eliminate connectors, didn't they didn't quite get it right Now. Maybe that was cuz it was hadoop that killed them. I don't know. But So what's the fast forward to 2020 when you started the company?

Dave: What's changed?

Michel Tricot: Yeah. So when you go back to the days of the Hadoop, the key way of accessing data was files. Yeah. And. People were just trying to dump as many files as possible, and you did not really have that many places where company had data. It was in general one or two main product databases, one ERP, a few APIs, but that's all what you've seen over the past 10 years is.

Michel Tricot: People are using and consuming a lot of SaaS products and just a recent study showed that we went from 16 on our ratio 130. So these are all 130 little silos of data. And yeah, like the demand for that data is not coming from engineers anymore. When it was Hadoop, it was engineers, data engineers looking to get data [00:04:00] somewhere.

Michel Tricot: Now it comes from a marketing ops team, a sales ops teams, a finance ops teams, and they need that data to do their job. And they are asking for breaking down these silos. And

Dave: can you paint a picture of the value chain? Because maybe be starting with the business person, marketing team going through however they get the data.

Dave: It's less and less it's dashboards. It's more answers. Yeah. And then where you guys fit in that value chain.

Michel Tricot: Yeah. So let's say you're a marketing ops person to understand how your eyed is performing, to understand how. Your email campaign is performing. To understand all of this, you basically need to connect to your ad platforms.

Michel Tricot: You need to connect to your email provider, you need to connect to your marketing CRM, and this is where the needs come from. And then when you go look just at the data value chain, what's gonna happen is that [00:05:00] marketing ops person is gonna connect to Airbyte and say, okay, push the data that is on my HubSpot instance and make it available in Snowflake so that now I can connect, put my dashboard on it, put my automation on top of my snowflake, and start extracting this information so that my next campaign, my next ad campaign, my next email campaign is higher quality.

Michel Tricot: And I've learned from everything I've done in the past to do make the future better and,

Dave: okay. And so you essentially enable that. And the value. How do you describe the value proposition from an economic standpoint?

Michel Tricot: From, yeah. I I think at that point it's really why do we have a rev ops or why do we have a sales ops team? Ah

Michel Tricot: like these functions are here to [00:06:00] build more. Operational efficiency within a sales team. For example, if you can bring the right data to the right salesperson while they're talking to a customer, then boom, they might close this customer faster, or they might just close it plain and simple. Or they might be able to upsell that customer if they realize that, oh, this customer is consuming 50% of all our customer success and our support time.

Michel Tricot: Maybe we need to have sell them something on support so that's becomes one of the value prop. If you're looking at it from the lens of a sales ops team,

‍

Dave: You're accelerating the existing value proposition of a very important function of the company that's already been established and justified.

Dave: Exactly. So it's a no-brainer. What's your relationship with Snowflake? What's that partnership

Michel Tricot: like? Yeah, so the first thing when we released Airbyte, the first time, snowflake was the top places where people wanted to put data in. And we very quickly develop that [00:07:00] ability to push data into Snowflake.

Michel Tricot: And one of the things you get with Open Source is control. Control over where your data is going. And the way our connectors are working, they can run directly into Snowflake. So we have our own platform, which is a cloud product, or we have our enterprise solution, but we also now have the ability to run Airbyte connector directly into Snowflake native apps.

Michel Tricot: And that's something we announced today, where yeah, if you don't want data to be accessed by anyone else, then you, and maybe Snowflake, this is something you can do and open source give you that control because yeah, you can push that connector over there.

Lisa: Talk a little bit more about the native app that you just announced this morning.

Lisa: You mentioned some from a security perspective, but talk to us about some of the value drivers there and some of the outcomes that customers could expect to get from that.

Michel Tricot: Yeah, I think the overall vision here is ease of adoption [00:08:00] in a sense that once you're using Snowflake, you have a stamp around security, around your data.

Michel Tricot: So if you can run. Most of your data processing within the Snowflake environment, and that's the case for us. Like we run data ingestion, then everything happen within the security of Snowflake. So I think even for us as vendors, it removes friction. Meaning that if you go to if you're trying to sell your product and you get stuck at the security level, You don't have that problem anymore because now you're already vetted by the Snowflake Secure environment.

Michel Tricot: So that's a big advantage both for us as vendors, but also for our customers because it speeds up adoptions of new tools.

Dave: So you guys have raised a Crunchbase is accurate of a fair amount of money. Yeah. You've done an acquisition, so give us the [00:09:00] highlight of the company where it's at. I think you're at series B.

Michel Tricot: Yeah. So yeah we raised about 180 million. And as I said before, we've been really focused on making Airbyte as ubiquitous as possible. Meaning the moment you need to bring data from one system to another system, Airbyte should be top of mind. And I think this is really what we've done over the past three years, which is being top of mind, a top of mind solution for everything, data ingestion, etl, ELT, related Now, Our focus has really been on this top four product, open source.

Michel Tricot: So very strong adoption, bottom up adoption. And the first production pipeline goes through open source and then we focus on the rest of, as you mature with Airbyte is how do you what are the additional requirements that you have around like SSO, Rback, and that became [00:10:00] our enterprise solution and where we provide support at that point.

Michel Tricot: And after that cloud, you don't want, you might not want to host airbags yourself, you'll find using a SaaS service good for you. We have a cloud product for you. And the last one was what we call powered by Airbyte, which is powering tools to bring more data. Into applications, and this has really been the, yeah.

Michel Tricot: The accomplishment over the past three years is addressing these four different angles for making Airbyte ubiquitous .

Lisa: So a lot of growth and acceleration in a very short time period during a pandemic I might add. How have customers and the Snowflake partnership and their customer base influenced this accelerated growth for that year seemed to be on?

Michel Tricot: Yeah I think what you see there is that every time there is a crisis, people are just going to dive down a little bit on growth and focus on fundamentals. And what has happened really during, and we've seen it [00:11:00] firsthand during the pandemic, is people stop investing in some growth areas. And they started to refocus on this long project that they know are necessary for the future.

Michel Tricot: But that might not be the most yeah, but growth might have taken all the resources at the time. And what we've seen during the pandemic is people really looked at their infrastructure and what do they need to bring to make, to prepare their companies and their organization for the future? So they used that calm, more calm time of covid.

Michel Tricot: I don't know if you can say it's calm, but it was a little bit quieter and they just made this investment. And for us, that was a big accelerator because suddenly people are adopting solution like snowflake. And the first question people ask themselves when they acquire Snowflake is, how do I bring data into Snowflake?

Michel Tricot: And Airbyte was was there to to help.

Lisa: What my last question for you is, [00:12:00] did you happen to catch the keynote, the fireside chat last night between Frank and Jensen?

Michel Tricot: I did. I saw the summary.

Lisa: One of the things that, that Jensen said, one of the many things Jensen said that were colorful and interesting was that, companies are, have to go from being data companies, software companies to AI factories.

Lisa: AI factory. How can the Airbyte snowflake? Partnership help customers become or generate AI factories so that they can be competitive, deliver new products and services and new apps to their customers? Yeah,

Michel Tricot: It's funny because just before I was making a, I was doing a talk and one of my statement was, today, whether you are in retail, whether you are in insurance, whether you're in banking, every single person, every one of single one of these companies is a data company.

Michel Tricot: And I think. It just talks to the foundation and the requirements that is necessary for companies. Like [00:13:00] every company must be good with data. And the reason why it's important is because now it enables this additional places where data has value. And AI is a big one when like you can have the best AI algorithm in the world if you don't have data to power this algorithm.

Michel Tricot: It's not gonna be helpful, especially when you start working with large language model where you want to inject your own proprietary knowledge into in into the product. If you don't have that data, your LLM is not gonna serve you well, and that's where this type of partnership becomes important because, People that are using Snowflake are also really thinking hard about how do I make AI part of my strategy?

Michel Tricot: And we just want to continue to funnel more of that data to power not only like analytics, not only like operational use cases, but really AI as a big one.

Lisa: [00:14:00] Ah, excellent. Michel, thank you so much for joining us Talking about Airbyte I see you now have 300 plus connectors available, so you really, we'll just wipe that plateau right off the table.

Lisa: We appreciate you coming on the cube, talking to us about Airbyte what you're doing with Snowflake. Thank you.

Michel Tricot: Thank you for having me.

Lisa: Thanks Michel. We're our guest and David Vellante. I'm Lisa Martin. You've been watching theCube Day one, mage of Snowflake Summit. Tomorrow. This is a wrap for day one tomorrow.

Lisa: Don't miss out. We have a great guest lineup all day tomorrow. Frank Sluman will be here in the morning doing a great one-on-one with Dave. We also wanna remind you that you can find all of our Snowflake assets and all of the cube content on the cube.net and all the great analysis on silicon angle.com.

Lisa: Thanks everyone so much for watching. Have a great day. We'll see you tomorrow.

‍