I was recently involved in a message board conversation on Focus.com where the question was asked “How do you build confidence in the information being provided by a Business Intelligence Solution?” There were some pretty interesting responses there that suggested greater human intelligence and other possible solutions. In my response, I thought it boiled down to good in-good out.
Here’s my full response:
Building Confidence in BI
I see the problem a little bit differently. I see lack of confidence coming when you’ve got a meeting with three different departments who are all talking about the same measurement(e.g. sales volume), but have very different numbers. They spend half of the meeting arguing about what the right number IS instead of what the number MEANS and what to do about it.
So, building confidence (I think) boils down to two things: Good Stuff In/Good Stuff Out
Good Stuff In
First: you need to have taken a strategic approach to building out your enterprise data warehouse making sure that everyone in the organization (or at least everyone who’s going to be collaborating with the relevant data) is working from the same version of the truth. So that’s good stuff (vs. garbage) in
Good Stuff Out
Second: you need to invest in a BI front end tool that can allow users to see this information on their own terms. So for example: Sales sees the same data, but their view is of territory, and performance against target. Operations sees the same data but their view is of fill rates and delivery time to customers. Their interest may overlap when, say, Sales misses its revenue target due to Operations not achieving its target fill rates. Then, with a well managed EDW, they’re confident about the results they’re seeing (may not like it, but they at least have confidence in the information). So that’s good stuff (vs. garbage) out.
- Of course this is an oversimplification and a lot of other factors have to go into building confidence such as:
1. investment in master data management skills so that the data is (and remains) accurate
2. strategic alignment between business and BI teams so that the information remains relevant — i.e. give me access to data that I really need to care about
3. training business users so that you can have greater self service(the further removed you are from the data, the less trust you have)
4. feedback and improvement cycles so that you can continue to provide BI applications that stay relevant to business needs.
There’s a whole different level of confidence with more complex BI (e.g. predictive analytics) which has more to do with the predictive models and assumptions used to make projections. But I think that’s a different question.
By Adrian Alleyne, Director BI Market Research
© DecisionPath Consulting, 2011