Brand Loyalty and Boyle’s Law

Andrew Ehrenberg passed away this year. I had the pleasure of meeting him on several occasions and have him explain to me first hand his findings on brand loyalty. I never really understood it then and have still not entirely got my mind round it. Andrew applied the laws of physics to brands and in a study of over 100 fmcg brands in Europe, Japan and America proved something that none of us marketers wanted to believe, still don’t – that brand penetration and average purchase move in a constant relationship. Put in plain English, big brands are big because more people buy them and on average they buy more of them. Small brands are small because fewer people buy them and buy less of them.

“It is like Boyle’s Law, Mark, as the volume of a gas declines the pressure goes up in a constant relationship” he explained to me (I did not have the heart to point out I gave up Physics at the age of 13 years). But what causes the change, I asked? “Ah, that is for you clever marketers to find out!” he replied.

The implications of this struck me as staggering. For years I’d believed that most of marketing was designed either to build penetration or increase loyalty and that we could somehow manipulate these as separate objectives. His findings were clear – they work in tandem. If you make a brand stronger then both penetration and loyalty (in terms of how often/how much people buy) go up.

Over the subsequent years I was exposed to many marketing models that failed to take this into account. McKinsey have a funnel that dissects a brand into sub groups from aware to consider, to buy occasionally, to buy most often, which they use to define the specific marketing objectives for a brand – its opportunity to grow. Wrong according to Andrew Ehrenberg. Strengthen the brand and every one of these measures will rise.

At one level this makes sense. It is hard to think of any marketing activity which addresses only one aspect of the funnel. Even if you flight an ad that encourages people to try a brand for the first time it will also encourage some to use it more often. If you run a promotion that rewards heavy purchase it will get the attention of non-buyers and persuade them to try. But surely there must be different levels of effect that would cause a skew in Ehrenberg’s findings? He was certain that there were not – he had the data to prove it – unless there was some structural market anomaly. Unleaded petrol was just coming on the market at the time I met Andrew (tells you how long ago it was) and he used this as an example which could skew the results because the distribution of the new type of fuel was controlled artificially and patchy. It was, in economists’ terms, ‘an imperfect market’. But in most fmcg markets were brands are freely available and fair competition keeps prices in line (i.e. a more perfect market) Ehrenberg’s Law held true over time. Those last couple of words – ‘over time’ – are important. I am quite sure that in the short term, a month or so, certain marketing activity will disproportionately affect either penetration or loyalty but over the course of a year they effect will work their way through to either a stronger or weaker brand in the way Ehrenberg described.

During my time at SAB we focused hard on the nature of brand adoption (and ‘unadoption’) and loyalty, and I got heavily immersed in studying it from many different angles. This included a lot of data-based analysis building on Ehrenberg’s work but also ethnographic work (studying consumers’ behaviour in real time). We even looked at the differences in people’s learning styles from the findings of educationalists to see if we could segment the different ways people assimilate new information about brands. We studied the best circumstances in which to sample a beer brand – in a busy pub when the party is in full swing or when you had quieter groups of friends in the early evening. We found that if people were ‘out of their comfort zone’ for example in a new venue with an unfamiliar group of friends, it improved their receptiveness (in a lasting way) to new ideas and new brands. We looked at various conversion models and even developed one of our own. But I never felt we got to the ‘answer’ and I certainly never felt we fully took account of, let alone disproved, Ehrenbergs challenging empirical analysis and findings.

So I’m still thinking about it. Here are my latest thoughts.

The journey to loyalty is not a journey

At any one moment in time we can divide a brand into various usage/attitudinal groups:-

•    Blissfully unaware
•    Passively aware but don’t know much about it
•    Saliently aware – both know the brand and know something about it that makes it interesting/relevant/attractive.
•    Actively considering
•    Using occasionally
•    Using most often
•    Using exclusively and downright evangelical about it

You can look at a brand in this way – people have to fall into one, but not more than one, of these groups. BUT it is not a journey from “aware to loyal” with several neat steps along the way. You can go from blissfully unaware to evangelical in one brief moment.

There’s a difference between behavioural and attitudinal loyalty

Financial people only care about bevioural loyalty because that shows up on the P&L. If someone is buying a brand on more than 80% of their purchase occasions we can say they are behaviourally loyal. But they may or may not be attitudinally loyal – they may be buying out of indifference and/or inertia or they may be buying very deliberately because of their strong feelings towards the brand. Guess which is better in the long run?

Cheating only helps in the short term

It follows from the above that we can manipulate behavioural loyalty with short term incentives – a discount, a promotion. I’m not knocking this – if you can do that cost effectively and there is no competitive reaction it can be profitable. If you force trial and therefore force people to see how much better the brand is it can also be of value. But that is the point – you need to affect the attitudinal loyalty, the strength of relationship with the brand to make it so strong it will still be bought as often by as many people long after the discount ends.

We want to shift the demand curve

As Andrew Ehrenberg said, as marketers we need to figure out the best ways to create the difference in “volume and pressure’ in other words to strengthen the brand, in fact to shift the whole demand curve so that at any given price more people will consider, buy and buy more often. There are no short cuts to loyalty – it is a proxy for brand health and it takes time and effort.

I did not know Andrew Ehrenberg well but he struck me as a highly intelligent, decent and thoughtful man. He was a scientist and he enriched our understanding about marketing through the appliance of science. He injected some facts based on experimentation into a ‘marketing discipline’ that is often anything but.

Leave a Reply