Fivetran has closed its acquisition of HVR. CEOs George Fraser and Anthony Brooks-Williams explain how it happened and what it means for customers.
Well, it’s official: Fivetran and HVR are now one company. After the deal closed, HVR CEO Anthony Brooks-Williams and Fivetran CEO George Fraser sat down to discuss how it unfolded, what it means for customers, and what they see ahead for the combined company. (They also elucidated the role of ice cream in the merger.)
We were focused on an enterprise customer base, replicating data to the cloud from core on-prem databases that large enterprises run — Oracle, SQL Server, Db2. But we knew that more and more data would be generated by SaaS-based applications, and we were pushing into that space. From our perspective, Fivetran technology was the clear leader there.
And we were basically in the inverse situation. We were in the position of, How do we get good at everything HVR is good at? Replicating mission-critical database management systems for enterprises is an incredibly hard problem, and HVR is incredibly good at it. It would have taken us years to get to where HVR was. So the two companies have very complementary capabilities.
The problem is that there are so many SaaS data sources, and they’re all different. They all have different quirks, and you don’t really see those quirks until you have customers at scale using these systems, and then the quirks reveal themselves. We’ve spent many years studying the intricacies of hundreds of SaaS tools that modern businesses use, and we’ve figured out a way to hide that complexity and just present analysts with a simple, ready-to-query schema for each source.
You have to find an unobtrusive way to acquire the data from these database systems — you don’t want to put additional load on them. The best way to do that is to read the transaction logs, to actually parse the log files. Alternatively, you could use a method called triggers, but then you’re putting a load on the database, so it’s slowing down while people are trying to use it.
And at Fivetran we’ve experienced this. It’s common for these kinds of databases to run very close to the edge of their capabilities, so any added intrusiveness will show up. Your bank’s website will run a tiny bit slower because of a trigger. The tolerances are very low, which is why it’s so important to be efficient.
And today the volumes of data are so great — 90% of the world’s data has been created in the last two years. So the sheer magnitude of data is a challenge, and sub-second, near real-time replication is key. Our technique is the most optimal way to do it, but it takes years to develop. You can’t just spin up a team of developers and throw it at them. It takes a certain amount of skill and a certain mindset, and it will always take a certain amount of time. There’s so many different permutations and deployments that customers do — to be able to cover all of those, it takes years of effort.
And from the outside, we recognized this. We had been studying all the capabilities HVR had that we didn’t, and realizing, Oh my gosh, this takes a certain amount of time and there’s no getting around it.
It gives them one place for all their data movement needs and challenges. We have customers saying to us, That's great that you're taking our business-critical SAP data and putting it into Snowflake, but we want to put the SuccessFactors or Salesforce data into this environment as well. A lot of them are stuck in a legacy data integration world, which is like trying to fit a square peg into a round hole. They need to be able to make decisions based on the freshest, quickest data. Now they will have one place dedicated to moving all their data. That’s important because data targets or destinations may come and go, but the data is always going to move.
That’s the critical point I want to double down on, from the perspective of the customers: There's incredible value, especially for large customers, in having one company that acts as a sort of data Switzerland that will move your data from anywhere to anywhere. A company that's been around for a long time, and that’s going to be around for a long time — and no matter which tools and databases I adopt in the future, I'm going to be able to move my data reliably.
That’s the company that we concluded we could build together. If we had stayed separate, there was a likelihood that we would both end up as specialists in the domains we grew up in, and I think that would be a worse outcome for everyone.
Our mission statement is: Data is the lifeblood of businesses, and our mission is to keep it circulating. And Fivetran’s is: Making access to data as simple and reliable as electricity. If you think about it, it's all about getting as close as possible to a more autonomous environment and setup for customers. That's where I think they truly do complement each other.
I think the similarity of the mission statements reveals that both companies are primarily focused on the problem of data movement, which is unusual. Most companies in this general neighborhood of software focus on data transformation and getting insights from data. Very few companies have ever really said, No, what we are all about is data movement. Because it’s such a hard problem — you're down in the plumbing of all these systems. It gets short shrift from investors and founders and engineers, but it's an incredibly important problem, and we have both been working on different aspects of it for years.
When you’re able to consolidate many more data sources, new analytical use cases open up. You can combine data that you never could do before. You can capture it from a majority of places where it’s generated and make decisions based on it. In a sense, you've now got a mix of the new world and the old world. The core of what runs businesses today sits in the legacy world, but more and more is coming from SaaS, so we will be able to give customers a much better view of their business.
I see a blurring of the line between analytical use cases and traditional database use cases. The conventional way of thinking about analytical use cases is reporting. You’re doing your nightly reports — or your monthly reports, which a lot of people are still doing. There’s a lot of data, from everywhere, so it’s a consolidated view of the world, but it involves long latency and it's very static.
Whereas a traditional database like Oracle is very fast, but it's not going to have all of your data. There are intermediate use cases where you want to get a consolidated view, but you also want it to be low-latency. The only way to enable that is to have a data pipeline that combines the attributes of Fivetran and HVR — one that supports your databases, supports your tools, supports everything, at low latency and high volume.
So a lot of new opportunities will open up in that gray area between analytical use cases and traditional database use cases.
As I got to know more people than just Anthony at HVR, I noticed how similar a lot of the characters were at HVR and Fivetran. It was like that episode of Seinfeld where they go and meet the bizarro universe version of themselves. There were always people at HVR who were kind of like the counterparts of people at Fivetran.
And I suppose from Anthony’s perspective, vice versa. So it's a very similar group of people. HVR seems to have a very similar spirit as Fivetran. I don't know why that is, but somehow that turned out to be the case.
Yes, George said people who work in moving data all have a similar personality and similar traits. But for me, leaving aside the difficulty of doing such a big deal in such a short period of time, this has been a good process, and a quick one. My team have certainly got deal fatigue because we were out there raising funds for the last four or five months before engaging with this process. But we ended at a great place, and a home where HVR can continue to have a huge impact on customers.
If you look at the timing, it was a very quick deal, and that’s because it made so much sense. All parties are happy. It’s a win-win, with massive upside potential, not only for us at HVR but more importantly for our customers.
Yes, George reached out and we agreed to meet in Orinda; he was only about a 10-minute drive away. I said, Do you want to grab a beer or an ice cream? We were standing outside an ice cream shop called Loard’s, and I mentioned that my wife owned it, and George said, Well let’s get an ice cream then! So we went and sat in the back booth and spoke about a merger.
I had salted caramel.
I can’t remember.
I think you had salted caramel and cookies and cream or something like that.
Oh yeah. Salted caramel and cookies and cream. It was a weird combination.
Nothing. George just got to the point and said, Well, are we joining our companies or not?
That’s true. I’m not big on small talk.
And I thought, maybe he knows me, because I lose attention after about two minutes if someone doesn't get to the point. My team always jokes: Shiny object! Shiny object! He’s looking for where the next thing is.
But George was like, Hey, great to meet you, and he was super-complimentary about what we had achieved and the products we’d developed and the business and how hard all that is — and he seemed really genuine. So that's where it started.