Post Office scandal shows the perils of bowing down to data

The Post Office scandal is shocking. But can our industry learn from it?

It is often said that in business and life, we must be data-led. But the Post Office provides a cautionary tale about being blindly led by data. In our industry we sometimes let data blind us, though hopefully not with such devastating consequences.

Consider, for example, that most major fmcgs and retailers used to use either scanning or IGD pooled sales data to judge who was winning or losing in the market. Neither measure included Aldi or Lidl, which were growing fast. That would be a bit like saying Liverpool won the Premier League this year, because you’re discounting the teams that came first and second.

The flaw in the data drove a delayed response from the traditional big four. It should have been obvious that the measure was flawed, but for many, it wasn’t obvious.

That shouldn’t undermine the importance of data. You just need to use it well, rather than following it blindly. So how do you ensure that is the case?

First, however boring it may sound, decision-makers must understand the data. How is it sourced? What is included? For example, in fresh food, Kantar panel measures kg volume, but NIQ measures unit volume (packs). Pack sizes can change hugely with supply, and at different times in different retailers. So you see big differences between volume measures according to the source. Fail to understand the data, and you may come to the wrong conclusion about your performance, and possibly make flawed decisions.

Second, challenge the data if it smells wrong. If you’re a retailer and your store managers are highlighting an issue or opportunity, but the data fails to confirm it, don’t fall at its feet. Is it to do with the data collection method? The sample? Data should shape and challenge your thinking, but you should also respect your observation and intuition. I saw a mass-market retailer convincing itself via a survey that price was not that important – common sense should have put paid to that.

Third, be careful with targeting measures. An example: mystery shopping. You work out what matters most for shoppers, so you create a list of what your team should do in response, and measure those factors via the mystery shopper. Your team responds in kind. But they begin to think the goal is to get a good mystery shopper score, not to serve customers well. Worst case, they ignore anything not on the list. It doesn’t have to be like this – but I have seen it happen. Try and focus on the end outcome (sales, or customer satisfaction) to manage this risk.

Data isn’t bad. The bigger the organisation, the more important it becomes. But we need more than just data. Sometimes we are too quick to bow down before the data – and at the Post Office, the result was catastrophic.