Where BI Is Heading (Must Head) to Stay Relevant
Originally written by Analytics Demystified on July 7, 2009
I stumbled across a post by Don Campbell (CTO of BI and Performance Management at IBM — he was at Cognos when they got acquired) today that really got my gears turning. His 10 Red Hot BI Trends provide a lot of food for thought for a single post (for one thing, the post only lists eight trends…huh?). It’s worth clicking over to the post for a read, as I’m not going to repeat the content here.
BUT…I can’t help but add in my own
drool thoughts on some of his ideas:
- Green Computing — not much to add here; this is more about next generation mainframes that run on a less power than the processors of yesteryear
- Social Networking — it stands to reason that Web 2.0 has a place in BI, and Campbell starts to explain the wherefore and the why. One gap I’ve never seen a BI tool fill effectively is the ability to embed ad hoc comments and explanations within a report. That’s one of the reasons that Excel sticks around — because an Excel based report has to be “produced” in some fashion, there is an opportunity to review, analyze, and provide an assessment within the report. Enterprise BI tools have a much harder time enabling this — when it’s come up with BI tool vendors, it tends to get treated more as a data problem than a tool problem. In other words, “Sure, if you’ve got data about the reports stored somewhere, you can use our tool to display it.” What Campbell starts to touch on in his post is the potential for incorporating social bookmarking (“this view of this data is interesting and here is why”) and commenting/collaboration to truly start blending BI with knowledge management. The challenge is going to be that reports are becoming increasingly dynamic, and users are getting greater control over what they see and how. With roles-based data access, the data that users see on the same report varies from user to user. That’s going to make it challenging to manage “social” collaboration. Challenging…but something that I hope the enterprise BI vendors are trying to overcome.
- Data Visualization — I wouldn’t have a category on this blog dedicated to data visualization if I didn’t think this was important. I can’t help but wonder if Campbell is realizing that Cognos was as guilty as the other major BI players of confusing “demo-y neat” with “effective” when it comes to past BI tool feature development. From his post: “The best visualizations do not necessarily involve the most complex graphics or charts, but rather the best representation of the data.” Amen, brother!!! Effective data visualization is finally starting to get some traction — or, at least, a growing list of vocal advocates (side note: Jon Peltier has started up a Chart Busters category on his blog — worth checking out). What I would like to see: BI vendors taking more responsibility for helping their users present data effectively. Maybe a wizard in report builders that ask questions about the type of data being presented? Maybe a blinking red popup warning (preferably with loud sirens) whenever someone selects the 3D effect for a chart? The challenge with data visualization is that soooooo many analysts: 1) are not inherently wired for effective visualization, and 2) wildly underestimate how important it is.
- Mobile — I attended a session on mobile BI almost five years ago at a TDWI conference…and I still don’t see this as being a particularly hot topic. Even Campbell, with his mention of RFIDs, seems to think this is as much about new data sources as it is about reporting and analysis in a handheld environment.
- Predictive Analytics — this has been the Holy Grail of BI for years. I don’t have enough exposure to enough companies who have successfully operationalized predictive analytics to speak with too much authority here. But, I’d bet good money that every company that is successful in this area has long since mastered the fundamentals of performance measurement. In other words, predictive analytics is the future, but too many businesses are thinking they can run (predictive analytics) before they crawl (performance measurement / KPIs / effective scorecards).
- Composite Applications — this seems like a fancy way to say “user-controlled portals.” This really ties into the social networking (or at least Web 2.0), I think, in that a user’s ability to build a custom home page with “widgets” from different data sources that focus on what he/she truly views as important. Taking this a step farther — measuring the usage of those widgets — which ones are turned on, as well as which ones are drilled into — seems like a good way to assess whether what the corporate party line says is important is what line management is really using. There are some intriguing possibilities there as an extension of the “reports on the usage of reports” that gets bandied about any time a company starts coming to terms with report explosion in their BI (or web analytics) environment.
- Cloud Computing — I actually had to go and look up the definition of cloud computing a couple of weeks ago after asking a co-worker who used the term if cloud computing and SaaS were the same thing (answer: SaaS is a subset of cloud computing…but probably the most dominant form). This is a must-have for the future of BI — as our lives become increasingly computerized, the days of a locally installed BI client are numbered. I regularly float between three different computers and two Blackberries…and lose patience when what I need to do is tied to only one machine.
- Multitouch — think of the zoom in / zoom out capabilities of an iPhone. This, like mobile computing, doesn’t seem so much “hot” to me as somewhat futuristic. The best example of multitouch data exploration that I can think of is John King’s widely-mocked electoral maps on CNN (never did I miss Tim Russert and his handheld whiteboard more than when watching King on election night!). I get the theoretical possibilities…but we’ve got a long ways to go before there is truly a practical application of multitouch.
As I started with, there are a lot of exciting possibilities to consider here. I hope all of these topics are considered “hot” by BI vendors and BI practicitioners — making headway on just a few of them would get us off the plateau we’ve been on for the past few years.