A Simple Plan
QlikTech CEO Mans Hultman sets out to prove there is a simpler alternative to the traditional BI technology stack.
Forty miles across the Oresund Strait from Copenhagen, the city of Lund, Sweden, rises like a picture postcard of Europe in the Middle Ages, complete with an ancient cathedral and a university center that ranks among the oldest and largest in Scandinavia. It was amid the cobbled streets and gardens of Lund that entrepreneur Mans Hultman first came across a small private company with a business intelligence (BI) tool called QlikView. Hultman had worked the IT business in the 1980s but more recently had come from a wholesale/logistics venture, the successful sale of which left Hultman and his partners with a holding company and investment capital. "I found this little company in a science park and was thrilled about the product," Hultman says. "I had tried business intelligence tools in the 1990s and honestly couldn't understand how they worked. When I tried QlikView, not only did I understand it, I was able to build my first application in about six hours with no training. I figured it was something everyone should have."
Hultman joined the company as a board member and his partners followed with investments. In 2000, Hultman became CEO of what would become QlikTech and rolled out a global sales plan. The company set up a U.S. headquarters near Philadelphia, and though it is still a small upstart, QlikTech now calls itself the world's fastest-growing BI software company with record 2005 sales that grew 80 percent year over year. DM Review Editorial Director Jim Ericson recently spoke with Måns Hultman about QlikTech's success and plans for the future.
DMR: You seem to have stumbled across a novel BI tool that had been around for a while but wasn't well known. How did this happen?
MH: The product had always been excellent, and there was nothing preventing us from selling it when I joined the company - but the company I joined was in rather bad shape. It was very product oriented and [yet] the management didn't understand how to sell products. We realized that we needed to change management, and so I took the CEO position. Also, we needed a global strategy. You can't be successful in Scandinavia and get away with it because eventually, regardless of how good you are, the global guys will come and take you out just because they are bigger. My thinking was global, but of course that's bold when you come to a company with revenues of $1.5 million and no sales organization or brand awareness.
DMR: Your product seems to be straightforward and even simple. What is the pitch you take to customers?
MH: We often use a quote from DM Review when we benchmark against the rest of the market. The 2004 DM Review readership study showed that the average deployment is about 17 months, the average budget for BI/DW is $12.89 million and only 35 percent of projects were rated successful. That deployment compares with less than five days in our case. The cost is not really comparable, as we have yet to invoice a customer $1 million. Our success rate is over 99 percent, and we reach between 50 and 85 percent of potential users in our customers' organizations.
DMR:I'm going to guess large prospects are going to be skeptical of the "cheap and easy" approach. What's underneath your product that makes it suitable for an enterprise?
MH: Most [competitors] we watch fit into a stack of tools in one way or another. All of those solutions are based on a stack that consists of data warehousing, ETL tools, cubes, report generators, distribution tools, security and so on. QlikView is actually an extremely well-normalized database, to the fifth degree or so. We can sit on a data warehouse if the customer wants that, but most customers who didn't have a data warehouse before they bought QlikView didn't go on to develop one afterwards. They keep reloading QlikView from whatever operational data they have or from the Web, XML or Excel. What we do is load the underlying data and remove all redundancy. The trick is that we do queries and aggregations in memory. That's the secret sauce. It is not as much in the data set and its structure as it is in the way we make queries.
DMR: With so much emphasis today on data quality, I don't see how you can sidestep that issue so simply.
MH: Well it's funny to me in a way. I hear lots of discussion from competitors and from the industry about data cleansing. We never hear that from our customers. In fact, our customers are happy that we show them data that is not cleaned because that creates a business process refinement event. The first thing our customers will discover in a QlikView application is that something is wrong.
DMR: You're saying that data cleansing is a business process and not an automation problem?
MH: Yes, and it jumps up on the agenda for management when they understand that they have a problem. At the end of the day, you have to make your people understand that they must enter data in the right way. If you just explain it to them or write down a procedure, people will read it but won't understand. When they see that they actually have the same customer entered twice under different names, they understand the need to change. What happens in practice is that people don't really clean historical data, but they shape it up. From the point they shape it up they have better data quality and know whatever preceded that point is bad quality.
DMR: What kind of analytic functionality comes with QlikView?
MH: We don't do things like data mining, but we have statistical functions that cover at least as much as Excel covers. We have automation capabilities and an API; you can trigger events from outside. You have the ability to write your own functions in QlikView if you want alerts and a certain amount of automation. But when we refer to analytics, we mean that we work in areas of analysis rather than reporting. It's a little hard to position because everyone is using the same words. For example, we had a bit of a wake-up call at the Gartner show in London. [Going in] we claimed to be the simplest tool on earth, and our vision statement says we want to simplify analysis for everyone. Then, at the Gartner show we saw everyone else saying that they are simplifying analysis for everyone. It just happens to be a little truer in our case. I think it's ridiculous that people talk about "BI for the masses" when their products are so complex that not even well-educated CEOs can use them. How do they expect blue-collar workers to use these tools if the best-educated workers can't use them without extensive training? To me it comes down to simplicity. Wasn't it ridiculous when people started talking about "BI for the masses" and all of a sudden report generators were fashionable again? I found that strange.
DMR:In that report generation doesn't empower analysis for more people?
MH: Yes, in my mind people started looking for more interactive tools because they were lacking interactivity with report generators. When we develop technology we have to respect that we are developing tools for human beings of all sorts.
DMR: About 45 percent of your sales comes through partners. Are you affiliated with the usual big names in North America?
MH: No (laughs). Our success has been with small partners. We have been courted for five or six years now by more or less all the big names. But the fact is we haven't been a very good fit for the big consulting companies or system integrators so far because they are looking for products that drive services, and we don't drive services.
DMR: So, how do you get noticed out on the sales front?
MH: In competitive situations we ask prospects: when our competitor says you need an ETL tool, is that because you want an ETL tool or is it because they tell you to buy one? It's typically not that they always wanted an ETL tool that they buy one; it's that they needed a solution to their problem. The same goes for data warehouses. When we say you can buy that stuff but you don't need it, of course, some people don't trust us. But an increasing number of people believe we will survive and can look around at other companies that believe it and are happy with us.
DMR: Where does QlikTech fit into the landscape of big and small competitors and all the consolidation going on?
MH: I see not only the consolidation, but also the fact that there is some disappointment with the traditional business intelligence stack. That's my opinion, and of course, I would benefit from that disappointment. I would say that over the last two years, a number of companies that bought into the big stack have started to see projects that have been a failure for the business and for management. CIOs have started looking for alternatives, many times quietly and secretly, and there haven't been many alternatives. They have kept to their investments in hopes they will eventually make it work, and some of them make it work, too. I'm not saying it's a total failure, but a couple of years ago when we came into large companies it was typically through the back door at the subsidiary level. Sometimes it was with the knowledge of the CIO and other times it was without. It took us a long time to move forward internally in those companies. Nowadays we see a totally different type of interest. It almost immediately goes up to the corporate level, which is sort of annoying for us in a different way because we want to keep the sales cycle short. As for consolidation, we are happy to be independent. If there is anything I could wish for, I would wave a magic wand and make my sales force 10 times bigger, because the market opportunity is enormous.