Why we should question our faith in data

There is a pervasive dogma in our industry that we must be data-led. It’s something you’ll never get fired for saying; it feels transgressive to question it. But the reality is more nuanced.

Of course, data can add value. If you can measure how big Hello Fresh and Gousto are, and which consumers are buying from them, that is massively helpful.

But often data, unwisely used, can cause big problems. The retailer who convinced themselves from analysis of their customer satisfaction data that price was not of pre-eminent importance to their customers. The food manufacturer who switched from value to volume share as their key measure of success, after seeing a strong run in very heavy packs. There are lies, damned lies and statistics.

So what can you do to ensure data plays a positive and appropriate role in your business, as part of your toolkit?

First, all decision-makers need to understand the measures they are using. It is not enough to leave it to your insight people. You must at least know the fundamentals. Scanning measures do not properly include discounters. Kantar Worldpanel excludes food bought and eaten out of home. Customer satisfaction tells you about existing customers, not potential customers who shun you. Penetration can be measured over any time period (an especially persistent and meaningless mantra is the notion that brand growth is all about penetration).

Second, use a variety of different measures, including seeking out new ones. Resist the urge for a simple life and “one number we can all focus on”. Life and business is not that easy. I have sometimes heard people criticising Kantar’s numbers because they get the growth of a small category in a small retailer wrong by a few per cent, and then writing Kantar off as a result. What a waste. Kantar Worldpanel is a treasure trove of insight – the one tool I’d take to my data desert island.

So accept multiple measures, and understand their limitations and strengths. Take notice when different measures clash – it probably tells you something. Look for new measures as markets develop – if you’re in vitamins, how are you tracking Symprove, which is mainly subscription-based?

Thirdly, and most importantly, keep faith in your observation, experience and instinct. If what you see, believe and sense clashes with what the data says, have the courage to challenge the data. See if the data holds up and be ready to find your intuition is wrong. But equally, be ready to find the data is wrong or misleading.

The great companies and the best people in our industry are not data-led. They are data-influenced. They resist the temptation to be seduced by the idea that the answers are all there in the data, if we could only see.