I’ve been geeking out with SSIS for the last week or so and posted a blog over at DynamicsCare.com with some hard won knowledge.
I had the great honor of presenting at the SQL Pass meeting in Winston-Salem, NC on February 28. The subject was the “Design of SSAS Solutions for ERP Data”. A sure conversation killer on a date, but pretty cool stuff for the SQL gearheads I hang out with.
The preso contents are in my SkyDrive public folder (darn SkyDrive URL’s). Let me know what you think.
Data is like the new oil. Everyone wants it, the Chinese crave it, and the press talks about nothing but it. But, who’s using it?
According to this article from CIO.com, not the guys you really want using it. In essence, it claims that the managers who pay for the IT infrastructure to create the business intelligence data don’t trust it and don’t use it. In the interest of full disclosure, Tableau is responsible for the research and they have a clear agenda to push, but their findings do match my experience so I’ll assume it to be true (because, as you, my faithful readers, know – I am a genius).
In every BI project I’ve done (and it’s a lot), I’ve found this same thing: managers don’t trust the data they pay for and make decisions off the cuff that the data doesn’t support. Why?
Here’s my thoughts:
- They can’t get at the data and the details themselves. When they can’t validate the data themselves (by seeing details or drilldowns) its easier for them to dismiss an undesirable result by making unfounded guesses about the underlying data. Examples: “Oh, that probably doesn’t include May’s results”.
- They find errors every time the look. When a business users hits a report and immediately spots an error (or, god forbid, makes a public claim about a business condition to only find later it was wrong), they will throw the entire system into doubt and stop using it.
- They’re stupid. I hear this from IT all the time. You know how many stupid people I’ve met at my clients in the last 10 years? 1. He got fired. I don’t think the users are stupid – I think they are overworked and expect us to provide good result without them functioning as QA.
How to deal with it? Outside of all the good, usual consulting practices, I think the answer is pretty simple. Let them get personally attached to the system by:
- Getting them involved early in the project through agile rollout and frequent communication.
- Give drill downs to underlying data everywhere it makes sense.
- Train them on how to get at the data themselves AND give the simple, intuitive interfaces to get their information.
- Lastly, do the hard work of validating the data when you write the load packages and create the interfaces.
On January 23rd, Microsoft announced a change to its BI roadmap through discontinuing portions of Performance Point Server and downgrading other functionality into MOSS Enterprise. Read my post on this subject at the IBIS ERP Blog.