The argument is that differentiation does exist. But it happens within, not between, brands. For example, a two-door sports car is different to a five-door family car. However, most competing brands sell both. But brands differentiating from each other is difficult. Mostly because they copy each other. Successful ideas and innovations are quickly imitated.
But differentiation can exist. I saw that first hand too. In the UK and Ireland, O2 had exclusivity to sell the original iPhone for two years before other mobile networks could. The iPhone was very different to every other brand of mobile phone at the time. Clearly the difference was valued. Thus, the price premium. And, for those two years at least, O2 was different – it was the only network that sold the iPhone. This was temporary differentiation of course. Peter Thiel, the co-founder of PayPal, believes in differentiation. Although he points out in his book, Zero to One, that you need genuine difference. Ten times better. Not trivial differentiation.
Does any of this actually matter? Or is it just an academic argument? It does generate some good debate among the academics.8 A debate worth having. But are we just mincing words when we say differentiate versus distinctive?
From a practical perspective, I think it matters. This is about making better marketing decisions. If we have a powerful differentiator that customers will value and desire, great. Let’s shout about it. I’m working with a brand at the moment that I think has, at least for now, something differentiated and valued enough that people will pay more for it. They’re going to advertise their point of difference and see what happens.
But if we don’t have meaningful differentiation, it does seem feasible that we could still win customers without it. Brands that we are familiar with, or that are popular, have done so. Knowing this, we might save ourselves hundreds of hours, headaches, fights and agency fees trying to invent a weak form of differentiation that lacks credibility. All because we believe differentiation is essential. Professor Patrick Barwise told me that this “undue obsession with differentiation based on trivial USPs” is one of the biggest mistakes marketers make.
While the subject of differentiation is important, the bigger lesson I learned was to challenge marketing ‘truths’ and their underlying assumptions.
Fox lesson: Challenge long held marketing truths and beliefs.
6
Serial killer
Of all the tactical mistakes we make, there is one that crops up in many guises. It is the serial killer of good decisions. It is the mistake of assuming causation.
A common example is when we assume our advertising is effective because sales are up. We don’t consider other possible explanations that may have contributed to sales. Such as improved distribution. Or increased media spend. Or a price promotion. Or decreased competitor activity. Or the possibility that the entire market is up. Or the weather.
Granted, this specific example feels like schoolboy-error stuff. But it is quite common. Even now, I have to work hard not to trip over it. Another expert that I got to quiz was Professor Jenni Romaniuk, one of the best-known experts on branding in advertising. She is an Ehrenberg-Bass Institute Professor, a co-author of How Brands Grow (Part 2) and author of the more recent Building Distinctive Brand Assets.
I asked her views about a famous advertising campaign widely deemed to be successful. Within just a few minutes, Jenni discovered a piece of information I had missed. Before the new campaign, the brand had paused advertising for a long period due to a product recall. She explained that, “if this is the case then anything put on air would have lifted the brand”.
Her point was that in this situation, we could not say that increased sales were due to effective advertising. Instead, it was likely the impact of having advertising versus nothing, which is not that helpful for understanding what type of advertising works, or how well it works.
The opposite is also true. If your market share does not increase, folks might jump to the conclusion that your advertising isn’t working. But it is, of course, quite plausible that your market share could erode far quicker without your wonderfully effective ads.
Causation mistakes can turn up in more subtle ways. For example, what would you think if I told you that, in the UK, people prefer Costa Coffee to Starbucks? That’s probably not a fun piece of trivia if you’re the Starbucks marketing team. People in charge will want answers. The question on their minds would be whether or not Starbucks have an underlying brand health problem? And, if so, will this impact customer footfall or sales in the near future?
We might hypothesise that more people prefer Costa and we need to do something about it. Maybe because it is British? If so, perhaps we might decide to increase local brand advertising to drive up preference scores. Fear not – there is a less worrying explanation. Costa Coffee have about twice the number of coffee shops than Starbucks. The very fact that they have more stores means that more people drink Costa coffee. And when people are asked in surveys what coffee they like, most will name the coffee they buy and drink. Since more buy Costa, more say Costa. And so… more prefer it.
We tend to assume that preference drives sales. This feels right. We like something more, so we buy that product. In that sequential order. Preference first – then sales. But for Starbucks, according to them,9 it was the other way round. Sales – then preference. Their brand tracking data showed that the size of the business correlates with both usage and preference. If they open more stores, their preference scores will go up.
The unchallenged assumption we make is that advertising changes attitudes – which then leads to sales. It can. But it might not. The causation might be the other way around. The Starbucks team know this, of course. They understand that metrics such as preference, trust and brand consideration simply reflect that the brand has grown in a previous period, rather than the effect of communications.
Identifying cause and effect can be difficult. About a decade ago, when I was working for O2, we saw that our Google advertising was more effective than other advertising channels for driving our online store sales. We moved budget from our daily newspaper press ads to Google. But, when speaking to customers, we realised that the press ads were often what got them first interested. They would stumble across O2 ads for new phones while commuting to work on buses and trains. These days, we’d whip out our phones and search for more information or even buy immediately. But this was ten years ago. So, what they did was go online when in work or back home that evening. They’d search Google, and click. The press ads were creating the sales. Google was capturing them. We needed both.
You may have heard of something called the Rosser Reeves Fallacy.10 In research, we often find that people who notice our ads are more likely to buy from us. This may lead us to conclude that higher advertising recall will lead to more sales. However, the causation is probably the other way round. If you buy a brand, you’re two-to-three times more likely to remember its advertising than non-customers would.
Last year, when doing some work in the pet food category, I noticed that one brand had the fastest growing consideration and penetration measures. I assumed that pet owners must like this food more. And that this was the reason more people were considering it and buying it. But, after about 30 hours of in-depth interviews with pet owners, and some store visits, another possibility emerged. This brand had been very successful in increasing its distribution. It was possible that, because it was available in more stores, more people were finding and trying it. And this trialling was leading to higher consideration.
It so happens that I was introduced to the differences between causation and correlation quite young. I was about 12. Perhaps my dad thought I’d find this interesting. I didn’t. I do now.
Fox lesson: Whenever you see relationships, hold off on concluding what is driving what. It could be either. It might be neither. Look for a third variable that might be driving both.
7
“Only a desperate, insecure idiot would buy a Vespa these days.” (I own two)
I love Vespa scooters. Adore them. Have