Starting a Company in the Crowded Big Data Space
By Ben Werther, Founder & CEO, Platfora
Big data is a topic on the tip of everyone’s tongues right now, but the notion of starting from a clean slate and building a comprehensive visual analytic solution for big data was controversial when I started Platfora a little less than four years ago. Business intelligence vendors dismissed us, enterprises didn’t understand why they needed us and many VCs thought we’d never be able to pull it off.
I came to big data after a circuitous route through Silicon Valley, with stops at Siebel Systems, Microsoft and data analytics firm Greenplum. At the time it seemed the data warehouse market was all played out. But, two developments came together and changed the course of my life.
First came the explosion of new types of data — web, mobile, social, machine data, IoT, sensors and more; it quickly became apparent that the escalating volume and variety of data — and all the exciting new ways to use it — would quickly overwhelm the massively inefficient data warehouses prevalent in the corporate world. The second shift was the emergence of Hadoop, the (then largely unknown) open source software for storing and processing massive amounts of data on clusters of commodity computers.
The light bulb went on one day when a Dell executive wondered when Hadoop would be ready to take over the functions of a legacy data warehouse system. I realized there was no point making Hadoop fit the old paradigm because legacy systems were too slow and inflexible to carry us into the future. In the future, Hadoop would become the perfect dumping place for all that data; the challenge was to develop intelligent software that could quickly extract and make sense of all that information.
We concluded that none of the existing tools would work and we needed a clean-slate approach. With this new perspective came an idea (with an obligatory sketch on a napkin) of a new kind of product that, if we were able to build it, would transform data analytics.
Armed with this big idea, a road map and some early code, we set out in search of funding. Many VCs tried to temper our ambitions, asking us to instead focus on just a piece of the solution and how we would fit into the existing product ecosystem. Several VCs also pointed out that we were taking aim at incumbents in three distinct markets: business intelligence, scale-out data warehousing databases, and the market for data preparation and ETL (i.e. software used to extract, transform and load data).
We had few illusions about the technical challenge; ours was a high-wire act that could have easily failed if we didn’t get the engineering just right. But we held firm, confident there was no point trying to find a niche in the existing market that was doomed to fail. In hindsight, we didn’t appreciate how difficult it would be to convince an entire industry that our native end-to-end solution on Hadoop was the right path forward. But it was just as well, because there really was no lesser way to tackle this challenge.
Fortunately, Andreessen-Horowitz gave us the investor backing we needed to start down the road to achieving our vision, and a number of other world-class investors subsequently added their support. We’ve assembled a remarkable team and built a product that is transforming the way data is used at some of the biggest and most innovative companies in the world. There have been many challenges along the way, including hires that didn’t fit our culture and engineering problems we managed to overcome. Perhaps nothing prompted as much internal consternation as the debate over our go-to-market strategy.
Our software works on a scope and scale that makes us a natural fit for the enterprise market. We knew our success depended on signing a handful of key early adopters, big castles on the hill that could trumpet our products and validate our approach.
But how should we swarm those castles? Hadoop was still largely untested, so we needed to make a name for ourselves in a very crowded marketplace. The easy option was to throw a freemium version of our software over the wall, muck around in the trenches with business analysts and IT people and wait for Platfora’s software to catch on. This cheap and relatively easy go-to-market approach has turned many software startups into overnight success stories with huge valuations.
We decided against that approach, opting instead to knock on front doors and head straight for the executive tower. Our software is powerful enough to transform the way companies set strategy and interact with customers, and we needed to make sure C-suite executives understood our value proposition. This was the harder path, but there was no other way forward, not when we’re asking our Fortune 500 clients to make strategic bets based on our products and our commitment to be their partner for decades to come.
Despite our progress, we have still not reached the tipping point for big data. At times, the marketplace seems noisier than ever. But analysts are catching on, the media is taking note and customers are buying in to our vision. These are all real and quantifiable reasons to believe the market is slowly coming around to our way of seeing the future — as wildly controversial as it may have seemed a scant few years ago.
