Errplane founders Paul Dix and Todd Persen had an idea for a company last year around anomaly detection in data center monitoring, but they soon realized that field was crowded and it would take a long time to build out the infrastructure for the company. At the same, time they heard from customers they were more interested in the underlying infrastructure than the service they were offering, and they did something brave. They decided to pivot and build an open source product that would meet the needs of the entire market, rather than try to compete directly.
And the approach has been successful almost from day one. Today, they announced they are receiving $8.1M from Mayfield and Trinity Ventures to continue build the product and a commercial entity around the open source component.
It seems that Dix and Persen hit a nerve with their idea, which was an open source time series database. It differs from a typical database in that it collects and processes data that changes over time. “You generate a ton of data in a time series database, and it’s hard to scale out in regular database,” Dix explained.
The founders went through Y Combinator and were building Errplane as a SaaS company focused on anomaly detection in the data center. Last year, Dix was at a conference in Berlin and he saw many competitors trying to do the same thing as his company, and he realized that it was going to be a very tough way to go. That’s when he had an idea to build an open source database on top of which all of these competing companies could build their products and services. It was a huge gap in the market and Errplane saw it as an opportunity.
At that point they decided to give it a shot and carved out five weeks to take a hiatus from their service and try and build what would become InfluxDB. After they built the first version of it, the plan was for Dix to give a couple of talks introducing the project in the New York City area where he lived. But even before that happened word about the project appeared in the O’Reilly Radar blog and that got picked up by Hacker News where it stayed on the front page for a day. The pieces linked back to the project home page and it started to pick up steam, even before they had talked about it publicly.
By January Dix was spending a good amount of time giving talks on InfluxDB and it became clear that the original idea for the company was no longer viable or sustainable and they informed customers they would be pulling the plug on that part of the business at the end of March.
As Dix pointed out, his company’s project didn’t take a normal path. Usually developers create a project and then build a company around it, but it was reversed in this case. They had the company and they built the product. The only similar situation he could think of was Docker, which was developed at dotCloud before spinning out as Docker, Inc.
And as with Docker, the community is certainly helping. Dix says they built the core libraries, but the community has chipped in and helped with additional programming libraries.
Today they are concentrating strictly on InfluxDB and plan to build out commercial services later on. For now, the idea is to build the community and as Dix puts it, “optimize for developer happiness” and see what happens. So far it seems to have worked pretty well.