Alchemy. Rory Sutherland (2019)

What it argues

That human behaviour is not logical, so solutions derived purely from logic will systematically miss the ones that work.

His central example is the Eurostar journey time. Engineers proposed spending an enormous sum to make the trip faster. Sutherland’s point is that a fraction of that, spent on making the journey more pleasant, would have produced more satisfaction per pound. The logical solution was to reduce the number. The psychological solution was to change how the number felt.

From that: the opposite of a good idea can also be a good idea, psychological solutions are usually cheaper than technical ones, and the reason they are not chosen is that nobody gets promoted for proposing something that sounds silly.

What it gets right

The institutional critique is sharp and correct. Organisations select for defensible decisions, not effective ones, because a decision-maker is judged on whether the reasoning was reasonable rather than on whether the outcome was good. That produces a systematic bias toward the boring option, and Sutherland names the mechanism precisely.

He is also right that cost-effectiveness and logic are not the same thing, and that the cheapest lever in most businesses is psychological rather than operational.

And he is extremely good company on the page, which matters more than it should for whether a book changes anything.

What is wrong with it

It is unfalsifiable as presented, and that is a real problem in a reference section that otherwise insists on evidence.

The argument that illogical things sometimes work cannot be tested, because every success confirms it and every failure is explained as the wrong illogical thing. There is no version of the world the thesis would rule out.

The examples are also selected after the fact. We hear about the irrational ideas that worked. The ones that did not are not in the book, and there are necessarily many more of them.

Read alongside Sharp, the two are close to incompatible. Sharp says the answers are in the data and the industry’s problem is ignoring it. Sutherland says the answers are not in the data and the industry’s problem is worshipping it. They cannot both be the whole picture, and a reader should notice the conflict rather than enjoy both.

Who should read it, and who should not

Read it if your organisation kills every interesting idea at the business case stage.

Do not read it first. Without a grounding in what is evidenced, it functions as permission to do whatever you already wanted and call it alchemy.

Do not quote it to a finance director. It will not land, and it should not.

The soundbite

Organisations choose the defensible option over the effective one, because nobody is ever fired for being reasonable.

If you run the company rather than the marketing

The useful part of this book for you is not the marketing advice. It is the observation about how your company makes decisions.

If every proposal that reaches you arrives with a business case, and every business case rests on a projection, then you are selecting for proposals that are easy to justify rather than proposals that are likely to work. Those are different populations, and the gap between them is where the cheap wins sit.

The practical move is small. Keep a deliberate allowance for a few things a year that cannot be justified in advance, with a real budget and a short deadline, and judge them on what happened rather than on whether the reasoning was sound. Most firms do this informally for things the chief executive personally likes, which is the worst version of the same idea.

The caution: this book is the most quotable on the list and the least falsifiable. It is a corrective to over-measurement, not a licence to stop measuring, and in a company that already runs on instinct it will do harm.