#DataCulture mini-series: part 4

Part 4

What does bad look like?

And why to leave ‘ugly’ to the data experts

Here’s the last part of our mini-series and am sure, the one you’ve been waiting on to get confirmation that this is not you. But what if it is? (Let’s face it, for many of us this state will be the truth.) Then jump back up to part 2 and give me a call.

The good news is that ‘bad’ is easily visible

As facilitators of workshops and coaches, we hear the ‘bad stuff’ from workers within organisations of all types, sharing their frustrations about the progress rate of their organisation or lack of, in data exploitation. Rather than catalogue an exhaustive list of negative statements we hear, we’ve selected below a list of those specifically and commonly aimed at the Business Intelligence folks (a breed for whom we feel particular empathy). We have come to observe these comments are often to do with the stereotypes of the different functions and their perceived ability to work with each other.

• Slow
• Disorganised
• No visibility
• Underwhelming responses
• Reports, not insight
• Not Actionable
• Didn’t explain to me
• Not tagged properly
• Don’t track this, that…….
• Don’t understand / ‘get it’

As a BI professional, for the patterns of data to deliver rewards in your working day, you will over-index on numeracy, precision of command, accuracy of statement, etc. You could possibly find busy people in other functions difficult to pin down to exact requirements or specific business problems and logics that could be proven or disproven, as such their briefing leads you to provide compromised insights that run the risk of the response ‘so what’?

How can this example be helped?

The criticised can be helped by a facilitator encoding what the business folks need to know in greater detail, with greater precision and in response to specific prepared questions. The business folks can be helped by the ‘so what’ question being asked in first round feedback from the facilitator, not the internal client. This leads to refined, more impactful insights and recommended tests from the BI professional.

That’s learning; by showing.

What’s ugly?

Forgive the play on words, but actually, ‘good’ doesn’t really exist – it’s the process of hacking about with data, with a valuable business question to answer that means you discover the patterns, the value, from an ugly, undefined process that means you try, try and try again until the value is extracted. It’s an ugly process, best left to the data scientist, because when they have assembled a Proof of Concept model that can make a prediction using the available data, they will doubtless have done it the quickest, not the most orderly or fundamentally ordered manner in which they could cut to the insight. Good is probably ugly, but it’s valuable and once you know that you can derive organisational value from such a model, you can give it back to the engineer for re-building in a more stable and sustainable infrastructure.

They almost always ‘get there ugly’…………but good teams get there fast.