Q&A with Jill Dyché: Racing 'Round the Reporting Track
In which Jill dons a lovely hat, sips a mint julep, and watches the race.
I was minding my own business at the TDWI Las Vegas Executive Summit reception recently (read: I was waiting in line at the bar), when -- clearly knowing exactly where to look -- a former client found me. He'd recently changed jobs and wanted my opinion. I'm paraphrasing here, repeating a question that never seems to go away.
I'm a manager in my company's BI Competency Center. It sounds important, but all we really do is build reports. Business users email us data in spreadsheets, then we create reports and email them back.
The good news is that our BICC has a stellar reputation. We rank high on internal surveys and our budget has stayed steady.
The bad news is we're not doing anything new. My last company was all about self-service BI, with business users using their own tools to run their own reports. But here? I'm starting to think business people are over-relying on us, and we'll never be able to break free of all this manual report writing. My boss isn't listening -- he thinks the surveys say it all.
Can I breathe new life into my BICC?
--Manuel in Manitoba
I feel your pain. In this age of the Citizen Data Scientist, who the heck is still writing reports for someone else?
You are, and the truth is, you're not alone! Many companies continue in the mode of what I call the "report factory," churning out reports, and then new versions of those reports, and yet more revisions. Heard of data glut? Well, welcome to report glut!
This phenomenon happens when well-meaning BI managers want to be "approachable" and have "seamless rules of engagement." Need a report? We'll get it to you in 24 hours or less. Even if it kills us.
The report factory model is inefficient. It's not scalable. It's not interesting long-term for the analysts building the reports, who could and should be doing so much more. It can pigeonhole an entire team as a bunch of dashboard jockeys. (Giddyup!)
You're thinking your team should transform that report factory into a self-service model, allowing users to create their own scorecards and reports, thereby saving all that rich report-writing talent for more complex analytics, right?
Not so fast, Secretariat. Like any high-performing organization, your BICC should meet its customers where they're at. (I know, you really shouldn't make grammatical mistakes when you're finger-wagging someone, but that was a sentence in which I was trying to drive a point home.) So ask yourself the following question:
Are we able to segment our users based on their analysis skills?
Are there users who take your reports and drill down on the answer sets, building other reports that reveal actionable findings, maybe eventually even feeding predictive models? Do certain people bemoan the lack of a true BI portfolio of tools that they can choose from according to their analytics needs? Have you heard the words "visualization" or "data mining" being uttered within the four walls of a business unit?
If so, consider creating a discrete segment of users. Call them "power users" -- it's a fluid term that brings prestige. Assign a mentor to guide them in exploiting the functionality of your visualization tool. Teach them data curation. Introduce a new tool and give them a sandbox, forthwith!
When I was a consultant, I guided clients in segmenting user communities, a lot like marketing organizations do with their customers. We'd examine user behaviors, evaluate usage frequency, monitor help requests, listen to gripes, and ask a lot of what-if questions. When an end user insisted on being called a business analyst, I knew we were on the right track, so to speak.
What? Your users are happy with the status quo? They're content with receiving reports, and going back to the trough for revisions? Sadly, those users are already out to pasture. If you can't instill change quickly, you will be, too!