Qlik started out in Sweden in 1993 when data science and analytics weren’t as popular as today. So, how did it decide to pursue this idea?
It started out as a small consultant company and then they realised that what they were using for consulting was actually a breakthrough technology in business intelligence.
It was unique in that they pulled data into memory, which gave higher performance than any other tool at the time. And that enabled businesses to ask questions and get answers instantly, which was such a radical difference from other tools at the time. I worked at Microsoft at the time it started and we were building tools that would give you answers in minutes, so it was a dramatic difference.
How does your business intelligence (BI) software, QlickView, deliver?
You click on a data model and you can explore it and it returns data quickly. With a traditional system, you’d ask a question, but you had to ask the right question because it took so long. One of more interesting customers is the Swedish police. They had a serial shooter on the loose shooting people at bus stops randomly. So how do you find someone like that? Traditionally you had to think exactly what scenario you were asking for, run the system, wait maybe a day or two and then finally get an answer. But with Qlik, they were able to narrow it down to this guy and catch them.
How does QlikView distinguish itself from all the other players in the BI market?
The key thing is the simplicity for the business user to be able to do it themselves. The process we use is called natural analytics, which means we try to make it work the way the human mind works. We have this conviction that software is too often difficult to use with the human mind, it tends to work against it. For example, keyboards are really complicated things the first time you use them, but think about just speaking into your phone, which is far more natural. The human mind is really good at identifying patterns, so how can you get that into software so that business people will find the patterns and make sense of it.
What challenges have there been with the software and how have you overcome them?
Big data is a challenge because part of our secret to making this work is getting the data into memory that has to be done with very high performance. When data is in memory, the performance is much faster than when it’s sitting on a desk drive somewhere. If we could load that data into that chip and store it, we can investigate it and query it, which is what we’re achieving. There’s too much information out there so then we’ve got provide a way of navigating that information.
Also, if you think of the potential market for BI, several businesses could make a better decision if they had better information. So, in theory, we should be able to reach all business decision makers but in practice the whole BI market only reaches about 25%, and that has been static for the past 10 years. So our challenge is how to get to the 75%.
How do you find yourself competing against bigger BI tool providers such as Oracle, Microsoft and IBM?
Those players are stack vendor, meaning they want to sell you the entire software suite that will run your business not just the BI part. Microsoft, for example, wants to sell you all these tools. We don’t do anything like that. We do this little piece and they do this bigger piece. And in this little BI space, they really are in trouble. Are they doing well?
Why are they in trouble?
The reason is because they cannot make it simple enough. It’s ironic but it’s not in their interest to make it simple. If you think of Oracle, it makes its money from selling consultancy and complicated products that need a lot of servicing, which are simply not natural projects. Even their most recent ones are still too complex. The reason for that is they try to do too much and they’re too complex but that’s how they make their money…by selling big complex products to companies who are willing to send people on training courses to use them. They can be complex and they can get off with it in a way.
But, at some point, customers will get fed up with them in the same way that BlackBerry was too complex when it lost to the iPhone.
Would you ever partner with them if they approached you?
If they did want to partner, I guess we would. But you know what, they really won’t ask because that’s not their business model. Their business model is to sell on complexity.
They used to call this problem the ‘Washing machine repairman problem’ at Microsoft. When you’re washing machine breaks down, a repairman comes around to fix it with this huge tool box. The chances are he uses one spanner, one screwdriver and a little electric metre and that is all you need to fix a wash machine. But you would not pay him the call out charge if he turned up with only that. The truth is he uses a tiny little part of that to fix your washing machine. In the Microsoft world, they’re selling that big tool box so they make it complicated and complete because otherwise nobody wants to pay that amount.