Taking BI beyond the C-level executive

Vendors like Microsoft are also looking to place BI tools in familiar environments like Microsoft Office, to make BI more accessible to more people within an organization.

In May, Tibco announced its intentions to acquire BI company Spotfire for US$195 million, in a move bring BI to the "front-line" users.

ZDNet Asia sat down with Spotfire founder and CEO Christopher Ahlberg, who was in Singapore recently, to discuss this new direction and why the "commoditization" of BI is good for Spotfire's business.

How relevant is BI to the regular employee?
Ahlberg: BI is a market that's taken a while to come along. Back when it started, it was considered something quite esoteric and niche, but now we're seeing customers putting this out to tens of thousands of users. It doesn't mean it's everywhere yet, though, but what has changed is that people are now thinking about it in a broad way. It used to be the specialized researcher or high-level executive using our tools, but now it's the general knowledge worker. That's a very exciting development.

With so many vendors wanting to bundle BI into their typical offerings, where do you fit into the picture?
Traditionally, BI has been about reporting. But now that's a limited application; the industry is looking for more, and companies want a competitive edge, so they need BI to be predictive.

Vendors like Microsoft want to put BI in a familiar environment like Excel. That works for some people, but often we hear that employees want to do more with that data. With Microsoft commoditizing the reporting side of things, people are coming to vendors like us for predictive BI. They're saying: 'Help me to understand what's going to happen in the future.' I think predictive analysis is more of 'what-if'-ing--something to allow people to test ideas and explore scenarios based on the data they have.

How would you build BI into an IT infrastructure?
Right at the start, for sure. It's never advisable to run BI as a mega operation, but rather to inject it in steps. The worst way to design your infrastructure is to put BI in at the last, because nine months later, when the report is churned out after you're done building and gathering information, the report will no longer be relevant. Putting BI at the end makes the odds of it being right so small. In the nine-month period, things would have changed. So assuming that you built a data warehouse [for credit information] and nine months later people look at the reports. What are the odds that it's going to still reflect the current situation?

Often, in mega implementations, the report rolls out two years after the infrastructure was planned, and gets rejected by the CIO. IT says: 'But that's what you told me two years ago.' And the CIO then says: 'That's not how I run my business anymore.' It's one of the biggest diseases of IT.

With all the information the office worker has to deal with today, and now you're saying they may have to deal with BI as well, it sounds like people will face bigger information overload.
Yes, definitely (laughs). But that's why information visualization is gaining such popularity. My PhD research on the 'human side' to computing was what led to the development of Spotfire. Ten years ago, people were used to dealing with information in a text-based way. If you look at visual interfaces, everything is going graphic, with computer games as a good example. Hardly anyone plays a text-based game, anymore.

But information is still largely dealt with in text. Even with search engines, I'd have to go through hundreds and thousands of results to find out more about something, and it's very hard to deal with this because all I'm getting is text returned to me. If you think about the human visual system, if I gave you text versus good graphical displays, I'd argue that you can deal with a thousand times more information with the latter. It's like the classic expression: a picture is worth a thousand words.