Two stars, and I'll defend them.

My book club read this in January, which tells you something about January. Deborah picked it, and Deborah is a physiotherapist who has kept our group alive for six years, so let the record show that I like her enormously before I take her choice apart in public.

Start with the claim. We are looking at a self-help book that has been received as a psychology book, and that category error is worth an extraordinary number of copies. Nothing here is research. It is research-shaped. You get a citation, then an anecdote about a cyclist or a dentist or a man who stopped smoking on a Tuesday, then a numbered rule. The three are stacked so that the anecdote reads as evidence for the rule and the citation reads as evidence for the anecdote. Neither relationship is ever asserted outright, so neither can be checked. Read four consecutive pages slowly, with a pen in your hand, and you can watch the shell move.

Then there's the compounding. One percent better every day; thirty-seven times better in a year. Better at what? Measured how, on what scale, in what units? Compound growth is a property of quantities that compound, and almost nothing a person actually wants to improve behaves that way. Sleep doesn't. Patience doesn't. Whatever has gone wrong between you and your brother doesn't. It's a metaphor with a decimal point glued to it, and it has colonised every away-day slide in the country. The colonising isn't his fault. Putting the number on the cover, where it functions as a promise rather than an illustration, is.

Now the objection I take seriously, which Neil made at about eleven o'clock with his coat already on. You are reviewing a tool as though it were a book, he said, and tools get judged on whether they work. Fair enough. Do they? Two of them do. Habit stacking works — attach the new behaviour to an existing one and stop relying on memory to trigger it. Environmental friction works. I moved my phone charger into the hallway on this book's advice, slept better for six weeks, and I won't pretend otherwise just to win a point at Deborah's kitchen table.

But a tool has to be honest about the conditions it requires, and this one isn't. There's a chapter here that would look a single parent on two shifts in the eye and tell them their difficulty is friction. The word systems gets used throughout to mean things you personally control, a definition that quietly disappears everything else: money, illness, other people, the shift you can't refuse, the flat you can't leave. Show me the page where a reader learns that some habits are unavailable rather than unchosen. Cialdini is honest about the limits of his material. Duhigg is honest about his. This one sells certainty, because certainty is what the market pays for.

Let me put it another way. Hand me these ideas as a fifteen-page pamphlet with no anecdotes and no arithmetic and I'd call it useful and post it to my nephew. Padded to book length, dressed in a borrowed lab coat and shelved beside behavioural science, it turns into something worse: a machine for persuading people that failure is an information problem they could have solved with a wall chart. I've watched two people I love read this and conclude that their exhaustion was a design flaw in themselves. That isn't a literary objection. I don't much care.

The concession — and it's a real one — goes like this. The prose is clean. The structure is generous to a tired reader, which most non-fiction writers can't manage and won't attempt. The four laws are easy to hold in your head a fortnight later, and that's a craft skill, not an accident. I still use the two-minute rule most weeks and it still works. So two stars sit awkwardly beside a habit that has improved my mornings, I've got no tidy way to reconcile those facts, and I'm giving you both of them and letting you sort it out.

Deborah gave it a nine. We now have a rule that whoever chooses the book speaks last, and everybody knows that rule is about me.