Honest question: what exactly is meant by “differentiation,” and why does it matter? I just finished “Differentiate or Die” by Trout and Rivkin, and I am left feeling like the concept is mostly meaningless.

For those not familiar, “Differentiate or Die” is an famous marketing from 2000. The central thesis of the book is that the products and companies that are perceived as different by their customers will beat those that aren’t. While there are bits of this book that are interesting, it is mostly stories that are more clever than instructive.

There are two critical issues that run through the whole text:

Differentiation can mean anything: it seems that almost any product feature, benefit, or attribute could qualify as “differentiation”. This makes it nearly impossible to choose a point of differentiation for your own company or brand. The book also seems to be unclear if it is talking about marketing, products, business, or brand.

The book is built on anec-data: there are lots of short stories of individual companies or products that succeed or fail. While they sound impressive on their face, they don’t provide enough data or context to prove anything.

These two factors together give this book a wicked case of hindsight bias. Company A beat Company B. Why? Because Company A was differentiated on Attribute A of course! This analysis is trivially easy to do on almost any case study.

Is “Differentiate or Die” Unfalsifiable?

This relates to a concept from the philosophy of science: falsifiability. Falsifiable claims are those where certain results could be imagined that prove it wrong. If it is impossible to disprove a claim the claim is not meaningful or useful.

For example: If gravity didn’t work on high mountains or fast speeds, you could say our theory of gravity was flawed. Gravity does still work under these conditions, which is a sign it is a good theory.

On the other hand, imagine I said there was a lizard living in my bathroom that controls gravity. The lizard is invisible, and it is so tiny you can’t hold it, influence it, or otherwise measure it. It is virtually impossible to disprove my theory of lizard gravity because it is untestable!

“Differentiate or Die” is designed to be unfalsifiable.

In reality, it would be possible to test whether product differentiation mattered. If you think Company A is winning because it is differentiated on Attribute A, then you should be able to run a survey of people who buy from the category to see if that is true. Beyond that, you should be able to point out brands that predict how a differentiated product will perform in a crowded category.

In fact, there are some predictions in “Differentiate or Die,” and they are consistently bad. The authors suggest Nike shouldn’t get into sports equipment. They failed to predict that Chick-Fil-A would overtake KFC and Popeyes. They imply that products like smart phones would be failures. Etc…

Maybe I’m being unfair given that it has been over 20 years since the book came out, but I can’t remember hearing a single prediction where I thought “wow, they really nailed that.”

Differentiation might be useful in some cases, but the theory (as laid out in this book) is seriously undercooked. Is it better to have a differentiated product/company rather than not? Sure, but “avoid becoming a commodity” and “sell better products” are extremely banal observations.

It is no wonder that Byron Sharp and the Ehrenberg Bass Institute made such a splash with “How Brands Grow” – they actually had data!