How is your approach to BI/analytics different to what’s currently on the market?
I believe the traditional systems companies and governments have used to help people access and understand data have become heavy, complicated, inflexible, slow moving and expensive. There is a disruptive new technology that is putting all this back in the hands of people.
And how does it do that?
The first component is approaching the problem with computer graphics. Our founder, Professor Pat Hanrahan, who is now CTO, was also one of the founders of Pixar. In the mid-90s he decided to start applying computer graphics technology to information, which no one had really done before. Our technology is based on the idea that he originally posed: What if anyone of just about any skill level could query a spreadsheet or database using pictures?
We all agree that pictures are almost always your end destination of analytical work – you want to turn it into a picture so you can understand it. So why is that saved to the end? Why not start with it?
That’s visualisation. The second part is that we enable companies to pull slow-moving databases into memory with no IT investment. The price of memory has dropped incredibly and we now live at a point where you can fit billion row datasets into memory that you can buy on Amazon.com for $20.
It’s a sea change for analytics. Relational structures can’t keep up with the speed of questioning and volume of data. When those systems were architected it was so expensive they didn’t architect them to process in memory.
So we make it easy for a customer to just suck data into this high-performance data memory engine we have and couple that with interactive visualisation to achieve what I believe is the ultimate goal of any analytics product: to help people answer questions. If it doesn’t do that it’s useless. What we do is take raw, boring data and make it visual and useful so people can answer questions.
But is there really much of an issue with the old guard?
The fact is people today who have data and understand it and have questions are reliant on a priesthood of people called business intelligence developers who know how to operate these convoluted systems.
We live in a period of insanity and we’ll look back on it and laugh at it like we do with typing pools. We find that humorous now, so why don’t we find it humorous that whenever I have a question about data I have to submit a change request to another department? It’s laughable. We need to let anyone who understands data ask questions of it.
We have a very disruptive technology and business model and we’re making some people very upset. There is this priesthood of people who operate these old school systems but the way of human improvement is to allow more people to participate.
Who are you upsetting?
I’m not singling any one company out; I’m more interested in the class of technology. They have names like Cognos, Business Objects, MicroStrategy, Hyperion and they’re all owned by various companies now. These are heavy, complicated, slow-moving, development-intensive and extremely expensive systems. And they are dying, they are going to die. Right now they are the incumbent and we’re the salesforce.com. We respect them; they were a very valuable generation of technology, but we’re a symbol of the consumerisation of enterprise technology.
Can you provide an example of that in your product?
We use drag and drop, we are very visual and put lots of principals of psychology into the product. For example if you are looking at geographic questions, the software won’t ask you if you want a map, it’ll just make one. We also make really careful use of colour – for example we don’t insert red unnecessarily because it’s a very emotional colour and draws attention to itself, making that data seem the most important.
You’ve spoken a lot about being disruptive and doing things differently – what about price?
If you were looking at a classic BI project from one of the traditional vendors the licence and software cost alone is $100,000. That doesn’t include services on top. Tableau, by comparison, can start at $1,000. That will scale up as you add more users, of course, but all the way up and down the curve you’ll find that Tableau is much more affordable.
