Flop-shoppers – how big data may help companies discontinue products quicker

Are you a Harbinger of Failure? If so, then the marketeers would love to meet you.

Or that’s my take on some research soon to be published in the Journal of Marketing Research by management gurus, Eric Anderson, Song Lin, Duncan Simester and Catherine Tucker. (A big, big hat-tip to Tim Harford, of the Financial Times and Radio 4’s More or Less, whose excellent blog brought this to my notice. I recommend signing up to it.)

As I understand it – and that might be a complete lack of understanding! – the researchers have identified what they call ‘Harbingers of Failure’ – consumers who consistently buy products that go on to discontinuation. Or, as Harford neatly puts it, they are the flop pickers.

As Harford explains it, the researchers looked at a data from a chain of over 100 convenience stores – more than 100,000 customers, 10 million transactions and closing on 10,000 new products. Of those products around 4,000 (40 per cent) were no longer being bought after three years. That’s definitely some Big Data.

What’s behind the flop pickers’ superpower? We’ll have to wait for the research to be published in order to dig into the detail.

But we – by which I probably mean those with some deep and sophisticated knowledge of statistical economics (economical statistics?!?) – can start thinking about what to look for.

Straight up the old adage about correlation and causation comes to mind. It seems unlikely that the flop pickers are making the products fail. (If they are then their services would be in great demand by rival companies, each trying to make their competitors’ products fail by getting the flop pickers to buy them.)

So, is it something about the marketing of those new-but-doomed products that attracts a particular type of consumer? New products tend to come with both razzmatazz and a load of deals. Maybe the flop picker is less discerning – happy to keep buying the discounted failure-to-be long after others have gone back to an old favourite or moved on to the Next Big Thing?

Or is it something about the study? It sounds big – 100,000 customers and 10 million transactions. But that’s a mean average of 100 transactions per customer. There were closing in on 10,000 new products – and the researchers were looking at repeat purchases. So each customer can’t have individually been buying many of the new products. And it turns out that the Harbingers of Failure represent about one in eight of these customers. Half or more of the products they bought failed – though we’ll need to know what proportion of the non-Harbingers’ purchased also flopped.

That’s not to say that the researchers’ insights can’t be statistically significant – stats is a powerful discipline when it comes to discerning signal from noise. Even some basic understanding can take you a long way.

What is particularly interesting is that it’s not just that these flop pickers pick flops. Buying flopping products doesn’t just identify them. It seems that so can their habit of picking up the niche products that others don’t.

The researchers, in true researcher style, conclude that more research is needed. But if they are on to something then it could mean companies can pull products that will never turn a profit far sooner than they currently do.

Next time you’re in the supermarket have this in mind… if you seem to be tailed by a shifty-looking person with a clipboard sneaking a peak in your basket or trolley, alert security! (The real data is coming from records based on one of those loyalty cards that the stores are so keen for you to have.)

  • If you want to get on top of basic stats – using and communicating them – then why not try one of my workshops.

Leave a comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s