Tesco’s downfall: should a business be “data driven”?

Harvard Business review has an article on the recent troubles at mega-retailer Tesco. It seems Tesco’s much-vaunted customer analytics capabilities have done little to help it hold off its competitors in its core UK market.

Michele Kiss at Web Analytics Demystified has a nice commentary and reflects on the role of analytics in business decision-making.

The way that hard data (like customer analytics or digital analytics) links to decision-making has always fascinated me, and often frustrated me. Some management decisions go wrong because available data is ignored, or given insufficient weight. Others go wrong because too much attention is paid to hard data and not enough to other factors. The term “data driven business” is a great example of the second type of thinking, and as Michele says, it’s not the best mantra for most situations.

<IMHO>

Strategy should rarely, if ever, be data driven. (If a decision is so straightforward that it can be made purely “by the data”, that’s not strategy, that’s operations.)

Strategy should however be informed by available data: hard data should be part, but not all, of the decision-making process. Good strategic management involves appropriate efforts to obtain and understand relevant hard data, and informed judgements about how much weight to give to hard data and how much to other factors.

One of the key roles of a mature data analytics capability is to provide appropriate data to inform strategy. [Aside: another key role, a distinct one, is to develop new strategic ideas via analysis of data.]

</IMHO>

Nomenclature often leads us astray in business and I think this is an example where the language we use when talking about data and data analytics doesn’t help us get the right outcomes. I agree with Michele that we need better terminology than “data driven” and I like her term “data informed”. Early in my career we used to talk about “decision support” as a label for software systems designed to help with decision-making rather than for operational purposes – I like that term too, and I’m sad it seems to have gone out of fashion.

I also like the term “evidence based decisions”. The medical professions have taken this seriously with a whole discipline around evidence-based medicine. And politicians sometimes talk about evidence-based policymaking. There is also a less-developed discipline of evidence-based management, but that seems a bit heavyweight for the sort of decision-making that most data scientists and data analytics people are likely to be involved with.

What terminology do you favour when talking about data and decision-making? Let me know in the comments.

Posted in: Uncategorized