Sergey Nivens - Fotolia
Flatfile looks to be universal translator for data loading
Startup raises new funding to help advance its vision for making it easier for business users to extract, load and transform data from one source to another.
Onboarding data from one system to another is a basic element of any data architecture.
Among the vendors and technologies that enable data extraction and loading is Flatfile, based in Denver.
On March 10, the vendor said it had raised $35 million in a Series A round of funding led by Scale Venture Partners and including the participation of Workday Ventures. Flatfile will use the funding to advance its go-to-market efforts as well its technology platform.
Flatfile's platform enables data integration across different sources of data, so organizations can effectively transform the data into a format that is useful for different applications, particularly analytics and business intelligence.
In this Q&A, David Boskovic, co-founder and CEO of Flatfile, discusses the intersection of open source and enterprise software and where Flatfile is headed.
Why are you now raising a Series A?
David Boskovic: We raised as soon as we saw the opportunity to deploy the capital in a way that would help us scale, especially in the enterprise space.
We recently landed our first Fortune 500 customer and we saw an opportunity to build an enterprise sales team and go upmarket pretty quickly. As we go into more and more enterprise customers, that takes a lot more investment from the company.
For the next year, we will be focused on creating better and easier data onboarding, for any software product offering and anybody who's trying to get data in from their customers.
Long-term, our vision is to create what I like to call the Star Trek translator for business data [referring to the classic science fiction TV show's technology that enabled characters to seamlessly communicate with each other in different languages].
You should be able to go through any business transaction and exchange data without any effort whatsoever. That means I could just send you a file full of data, and it just works in your destination system.
David BoskovicCo-founder and CEO, Flatfile
What is the problem with traditional data loading and ELT [extract, load, transform] that Flatfile aims to solve?
Boskovic: At a high level we're doing pretty much what any other ELT tool does. The real challenge that we're looking to solve is high-variance input. Almost all ELT platforms are built around fully understanding the input data set to be able to build a mapping from point A to point B.
When you're dealing with transactional data, oftentimes every single file you're getting is a different format. So those rules for the data input have to get reestablished every time and for the market we are going after. People aren't able to use traditional ELT products as every input is different.
The innovation here is that we're simplifying the entire process, reducing the need to have a data scientist to help enable the data exchange process.
How does the Flatfile data loading technology work?
Boskovic: When we started Flatfile, there were two core things we enabled for data loading, mapping and the ability to correct formatting and data validation issues in real time.
What mapping is all about, is let's say for example I've got a spreadsheet and I need to get the data into a structure. If I'm going to do that manually, I'm going to prep the data inside of Excel and figure out how to provide an already mapped version to the user. With Flatfile, I can upload any spreadsheet and the system can validate the data against the target format. If there are issues like missing data, or invalid format, it'll show those and highlight them.
The advantage that we have is the ability to see an incredible volume of subject matter experts make a decision. At this point, we've seen billions of decisions go through the platform. Every time someone makes a decision about how something gets corrected, or how something should be mapped, Flatfile learns what users are trying to do inside of that content. So at this point, we can automatically suggest 95% of maps or corrections accurately.
Right now we provide a simplified ELT experience to non-technical users.
Does the term data middleware apply to what you do at Flatfile with data loading?
Boskovic: We are designed to essentially be in the middle; we were designed to be a component in another process, not to be a destination platform ourselves.
One of the key characteristics of middleware is that it's usually an integrated component of another product and that is exactly sort of how Flatfile gets plugged in. Customers build us into their product as either a white-label, or a core feature that makes us something that their customers trust as part of that product offering.
This interview has been edited for clarity and conciseness.