11 Meanings of Why-My-BI-Application-Is-Not-Useful

When a user of a BI application complains about the application not being useful - something that I hear way too often - what does that really mean? I can count at least 11 possible meanings, and potential reasons:

1. The data is not there, because

  • It's not in any operational sources, in which case the organization needs to implement a new app, a new process or get that data from an outside source
  • It is in an operational source, but not accessible via the BI application.

The data is there, but

2. It's not usable as is, because

  • There are no common definitions, common metadata
  • The data is of poor quality
  • The data model is wrong, or out of date

3. I can't find it, because I

  • Can't find the right report
  • Can't find the right metadata
  • Can't find the data
  • I don't have access rights to the data I am looking for

4. I don't know how to use my application, because I

  • Was not trained
  • Was trained, but the application is not intuitive, user friendly enough

5. I can't/don't have time do it myself - because I just need to run my business, not do BI !!! - and

  • I don't have support staff
  • I am low on IT priority list

6. It takes too long to

  • Create a report/query
  • Run/execute a report/query

7. I need to report/analyze on something that SQL can't do, such as

  • Faceted search
  • SQL on data with uneven, unbalanced, ragged, recursive hierarchies

8. It's a wrong BI application

  • I have strategic decisions to make, but the app is designed for operational decisions, or
  • I have operational decisions to make, but the app is designed for strategic decisions

9. The app is not integrated with other applications, processes or desktop so

  • I loose context
  • Have to switch apps, cut & paste
  • Don't know how to act on the info that I find

And my personal favorites

10. I don’t know what I am looking for, but my application is asking to

  • Run a specific report
  • Pick specific facts and dimensions

so I don't know where to start.

11. The app stops short of helping me, directing me to make the actual decision, even if I know how to use the app, have access to the right data, and can find what I am looking for. It's a loooooong way between finding the right info and actually make a decision based on the info.

What did I miss?


re: 11 Meanings of Why-My-BI-Application-Is-Not-Useful

Because no-one thought about the decision I have to make, the actions I have to take, when they designed any of this. And anyway I don't have the analytic skills/interest or the time to wade through the data anyway - I just want my system to help me do the right thing.

Most people in most companies don't want to do BI, they want to run their business and they need their systems and processes to actually help, not simply present data to them.


re: 11 Meanings of Why-My-BI-Application-Is-Not-Useful

Great list - I'd add a few:

It's not fun - my BI tool looks like it was designed for Windows 95 - it's no fun to use old UI.

It's not social - BI is an activity conducted by an analyst in a corner, there's no collaboration (see Forrester's great research on collaboration)

It's not agile - projects take too long to complete, and they are usually off target. There are no incremental improvements.

re: 11 Meanings of Why-My-BI-Application-Is-Not-Useful

Adobe produced a funny ad in 1993 about some of this when they launched Acrobat.


At Exalead, we're using semantic indexing technology to overcome some of the issues cited in this blog. Check out www.yakaz.com. It's a classified ad aggregation site where you enter what you're looking for and categorized inventory is returned, over-laid on Google maps. Exalead parses price and other attributes out of the unstructured text. It scrapes and indexes over 10,000 (!) ad web sites. An auto manufacturer is considering this kind of thing to track used-car prices...

re: 11 Meanings of Why-My-BI-Application-Is-Not-Useful

Great list, Boris. How about…

12) The reports present only historic data, but I’m interested in being more predictive. Are these trends meaningful? How do the measures and metrics relate to one another? Are these trends likely to continue? How confident should I be?

This is where the tightly integrated statistics and data mining capabilities that you and Jim Kobielus talk about become so critical.

re: 11 Meanings of Why-My-BI-Application-Is-Not-Useful

To develop custom applications suited for individual exploration activities, I recommend the book "Interaction Design for complex problem solving", Barbara Mirel. A whole chapter deals with exploration and sensemaking strategies and delivers valuable insights to design applications based on BI data.