Advice about impulse buying is nearly all the same shape: put obstacles between yourself and the purchase. Wait twenty four hours. Delete the shopping apps. Unsubscribe from the emails. Remove your saved card so checkout takes effort.

It is decent advice and it works about as well as you would expect, which is to say it works while you are motivated and stops working shortly afterwards. This article is about a different mechanism, one that adds something rather than taking something away.

Why willpower advice runs out

Every technique in that list is a cost you pay at the moment of temptation. Waiting is a cost. Re-typing a card number is a cost. The theory is that if resisting is cheap enough and buying is expensive enough, you will resist.

The problem is what happens on the other side. You resist, and then nothing happens. No confirmation, no receipt, no change to any number anywhere. You paid a real cost and received, in observable terms, nothing at all.

Do that forty times and the honest conclusion your brain draws is that resisting produces no result. It is not wrong. Under every normal setup, it genuinely does not.

The asymmetry nobody fixes

Buying something is heavily instrumented. There is a confirmation screen, an email, a notification, a delivery date, a package, and a line in your bank feed you will see again later. The purchase is recorded in half a dozen places without you doing anything.

Not buying something is recorded nowhere. It produces no artefact of any kind. Within an hour you have forgotten it happened, and by the end of the month there is no evidence you made the decision at all.

So the two options are not competing on equal terms. One of them comes with a parade and the other is invisible. It is remarkable that anyone resists as often as they do.

Give the decision a receipt

The fix follows directly. Record the decision not to buy, with the amount attached.

Two numbers: what you were ready to spend, and what you actually spent. Usually the second is zero. The difference is written down and added to a total.

What that does is give the non-purchase an artefact. It is no longer nothing; it is sixty eight dollars, in a list, next to the forty other decisions you made this month. The invisible option becomes as visible as the visible one, which is all it ever needed.

Financial advisers have recommended a manual version of this for years, usually phrased as "move five dollars into savings every time you skip an impulse buy". The instinct is right, and the reason it rarely sticks is bookkeeping: five dollars is not the real amount, and nobody remembers to make the transfer. Recording the actual figure solves both.

What changes after a couple of weeks

Three things, reliably.

The total surprises you. Almost everyone underestimates what they talk themselves out of, because they have never once added it up. Seeing four figures over a couple of months reframes you from "I am bad with money" to "I make good decisions constantly and had no idea".

The next decision gets easier. You stop choosing between a jacket and abstract virtue, and start choosing between a jacket and a number that is about to get bigger. Concrete beats abstract, which is the whole reason the jacket was winning.

You notice the moment. Logging requires catching yourself deciding, which is a small act of attention right at the point where impulse buying happens. This is most of the effect, and it is the same reason manual expense tracking beats automatic for changing behavior.

How to actually do it

Log the real number. What you had genuinely accepted paying, not the most expensive version you can justify. Inflated totals stop being believable, and once you stop believing the number you stop logging.
Count partial decisions. Ninety planned and forty spent is a fifty dollar decision. These are more common than outright refusals and leaving them out discards most of the data.
Do not aim for completeness. Two or three a day is plenty. There is no perfect score to protect, so a missed one costs nothing.
Never record a negative. Buy the thing and there is simply no entry. The moment your tracker can punish you is the moment you start avoiding it, and an avoided tracker records nothing at all.
Move the money weekly if you want it to be more than a record. The total is the exact figure to transfer.

Where the twenty four hour rule fits

It still works, and it pairs well. Put the thing in the cart, close the tab, and if you still want it tomorrow, buy it. What recording adds is the other half: on the days you do not still want it, you now get something for it instead of nothing.

Keep the Diff exists to be that record. Two numbers, a running total, and no way to log a negative. Free on iPhone, iPad, Mac, Apple Watch, Windows, Android and Wear OS, with no account and no bank connection. You can try it in your browser without installing anything.

How to stop impulse buying by recording what you didn't buy