Please tell us about your role and the team/technology at Fivetran.
Now, we’re at just over 300 employees. We’ve grown a lot in the last three years and expanded our customer base significantly, which now stands at a little over 1,000 customers.
What was your initial vision when you founded Fivetran? What parts of that vision have you already fulfilled? What’s left to accomplish?
Our vision really took shape when we found product-market fit and when we secured our first customer. That is – how can companies get ridiculously easy access to data?
For businesses, data lives in so many different places and that poses a very large challenge. We’ve addressed about 5% of the problem. That’s actually a lot and we’ve made some great progress, but there’s still plenty to do. That said, the bigger we get, the easier it gets.
As far as what’s left to accomplish, we need to have connectivity to every data source in the world, other than the ones we have now, and we need to make those connectors perfect. They should have super-short latency –a latency of seconds– and we should be able to send the data to any destination.
Also, anything that looks like a database, we should be able to support as a destination. There’s still a lot of ground to cover and I know our team is up for the challenge.
What makes Fivetran different in its competitive landscape?
Our main competitors are people building their own connectors. Data integration as a category has been around forever. Most data integration is accomplished by a couple of engineers writing some Python code or using a tool like Informatica. It’s ultimately up to them to specify all the details of how that pipeline works. They have to maintain that and watch over it.
Fivetran is a true fully-managed service that’s automated. We solve the problem for you instead of providing you with the tools to solve it yourself.
Fivetran’s technology bridges the gap between Applications and Cloud data warehouses. What other technology is required to leverage Fivetran’s connectors?
First, you need a plan for what to do with the data. We send the data to a database that you choose, and then you execute a plan to do something with it. That could be as simple as an Analyst who’s going to write a SQL query or that could be a bunch of application code written by Engineers who are using the data to operate your business.
Or it could be used for BI tools, building dashboards, data warehouses or databases, so you can really do anything with it. The Fivetran technology is essentially horizontal, meaning it can be used to do all kinds of things, very much like AWS.
Most of our readers know what ETL technology does, but Fivetran talks about an ELT approach instead. Can you please elaborate?
The way that you do data integration is you read data from some sources such as business systems like your database or Salesforce or your Marketing tools. And then you change the data and load the edited data into the data warehouse. So that’s the “T”, transform, in ETL.
ELT simply means that you load the data without changing it into the target database, and then you transform it in place. The reason why this is better is really simple–it’s because you still have an unmodified copy of the data sitting there. If you want to go back for any reason, for example, if you make a mistake, you can easily go back and fix it.
Which industries and/or regions have been the most receptive to Fivetran’s solutions? Any particular industry, you would like to see a great adoption?
Fivetran is like a utility that provides easy access to data – you can use it to do whatever you need for your business, regardless of industry.
What else do you want our readers to know or understand about Fivetran’s technology that we haven’t yet addressed?
We have an OEM software program in its early stages called Powered by Fivetran, where you can embed our connectors. And, if you’re building a product that uses data and needs to consume data from different sources, you can save yourself a lot of pain and suffering by embedding our connectors rather than building your own. It will also be much cheaper than building it yourself.
With classic ETL tools, you only ever use them to work with your own data and everything is bespoke. You have to go in and turn the knobs for every table. But with Fivetran, by virtue of being so automated, you can actually use it as a component of a larger system. It’s like a robot. You can just plug it in and it just works. That’s really different from a classic ETL tool.
If you were at an AI company building a product that needs to consume data from Zendesk for example (let’s say you’re doing AI for support), you would never want to use a non-automated tool because you would have to hire people to set it up for every single one of your customers.
With Fivetran, it’s actually automated and you don’t have to do that. Instead, you can just write some code that initiates these connections and we will give you a uniform, stable API to all of these data sources, which is going to make it a lot easier to build your product.
Thank you, George! That was fun and hope to see you back on AiThority.com soon.
George Fraser is the CEO of Fivetran, the leading provider of automated data integration. Previously, he was a Scientist at Emerald Therapeutics. He received his Ph.D. in neurobiology from the University of Pittsburgh in 2011.
Fivetran is the smartest, fastest, and most reliable way to continuously load data into your Cloud warehouse. Prebuilt, zero-maintenance Fivetran connectors deploy in minutes and completely replicate your applications, databases, events, and files — powering comprehensive business insights. Fivetran builds the best schemas and ERDs around, and we protect the continuity of your data projects with automatic schema migrations and micro-batch updates. Fivetran culture flows from a relentless determination to solve a critical data engineering problem — and then perfect the solution.