“Fire your market researcher” is this morning’s cry from the media as the market research industry’s surveys on UK voter intentions proved to be as accurate as the beliefs of the Flat Earth Society.
For months, the market research industry, armed with surveys, Big Data and powerful modelling has been unanimous in its prediction that the UK would return a hung parliament with no overall majority. The industry also predicted that the Conservatives would lose seats in parliament, and Labour would win seats. And the market research industry was utterly wrong on all counts. Just as we were so utterly wrong with the recent Israeli elections. The lesson? Never trust market researchers with predictions, especially if they agree with each other.
But before market researchers go down in history as the first against the wall when the election came, it’s worth noting that the market research industry, whilst being excruciatingly bad at predicting what people will do or buy (in this case parliamentarians), were far more accurate at reporting what people did do or buy – the exit poll of how people had behaved (voted) was nearly spot on.
So is the insight for market researchers that we should get out of the prediction business? No, not at all. It merely implies that we should base our predictions on real-word behaviour through real-world piloting and market testing. The proof of the pudding is in the buying. Period. Instead of asking what people think, or their opinions about this or that option, we should ground our research – qualitative and quantitative – wherever we can on what people actually did, or at least say they did.
This is not to say we shouldn’t ask for consumer opinions, far from from it. Interviews, discussion groups and surveys can still be extraordinarily useful in understanding the ‘why’ questions so necessary for improving and optimising; most notably by asking people why they did what they did.
The simple insight is the reminder that behaviour eats opinions and intentions for breakfast.