How brands really grow: People change their minds

To begin with, let's get some areas of agreement out of the way.

Loyalty exists. There's a common misperception that double jeopardy (DJ) destroys the idea of loyalty. In fact however, what characterises behaviour in a DJ world is what Ehrenberg calls "polygamous loyalty" (see "Understanding brand performance measures...", Journal of Business Research, 2004). Loyalty can be low (as in a person who buys a brand loyally but only 20% of the time); or it can be high (as in that same person buying another brand 60% of the time). It's a matter of degree. I use "share of wallet" to measure the extent to which a person is behaviourally loyal. And so in my world, behavioural loyalty is a percent that runs from 0-100.

Users who use the brand seldom are typically responsible for a lot of its sales. DJ analysis has made an important contribution to marketing by drawing attention to the value of what I'll call "casual" users. They may be people who seldom buy the category; or they may be low-level loyals. Either way, there are a lot of them and they're not like the highly loyal. They respond to light-touch mental reminders (like eye-catching adverts just before they shop). And they respond to marginally stronger physical nudges (like a centimetre or two of additional shelf space). They're not obsessing; and it'll be hard for you to get them to obsess.

So if we agree on this, what's the problem?

At the heart of 'double jeopardy' lies an important discovery: its aggregate patterns result from applying what's known as the Dirichlet distribution to individual buyer behaviour. And that's the problem; because one of the central tenets of the Dirichlet is that no person ever changes their level of loyalty to any service or brand.

You read that right: no person ever!

It's worth repeating - what holds the DJ edifice together is the Dirichlet distribution; and the Dirichlet asserts that individual preferences are stationary. That may be one reason why Ehrenberg insisted on the idea that "advertising can't be persuasive". In a world ruled by DJ, people can't be persuaded because their minds are made up. And so nothing persuades - neither friends nor relatives, neither experts nor celebrities, not a string of bad experiences; and especially not advertising.

Now - there's a lot that I agree with when it comes to 'double jeopardy'. I agree that marketing is about physical and mental availability. And I agree that sheer presence, whether physical or mental, is an important tool in the toolkit. But you can't argue, on the one hand, that brands only grow by acquiring more users; and yet assert, on the other, that people never change their preferences - for what, after all, are all those additional users if not people whose preferences have changed?

DJ encompasses a remarkable range of behaviour; and a lot of what looks like people switching back and forth between brands is just probabilistic variation around stable underlying preferences. But DJ isn't a world in which businesses or brands can grow. If I wanted to be somewhat confrontational, I would argue that Byron and his colleagues should have called their book "Brands can't grow - except in a non-Dirichlet world".

Let's end by going back to the question about how brands really grow.

From the first blog: it's not only about acquisition. People's loyalty is shifting both up and down all the time. Your business or brand will only grow if the rate at which some increase their loyalty, is greater than the rate at which others decrease it.

Now I would add: you can't grow share without changing what people want. So the big question is: how do you change what people want?

In the next few posts I'll talk about that.

Summary: if individual behaviour were perfectly described by the double jeopardy, then businesses and brands would not be able to grow. Fortunately, it isn't. Brands can (and obviously do) grow. But your business or brand will only grow if you change what people want. The question is: how?