Remembering which products you actually liked

Four different things get called “what we like”

They overlap often enough that we treat them as one thing. They are not one thing, and the difference is the whole problem.

Only one of the four is an opinion, and it is the only one nothing records automatically. The other three leave evidence behind. That asymmetry is why households end up with a detailed account of what they bought and almost nothing about what they thought of it.

Buying it twelve times is not a verdict

A household buys the same cereal twelve times in a year. That is a real fact about the household. It does not establish that anyone prefers it.

• It was what that store had.
• It was on offer often enough to become the default.
• One person does the shopping and has stopped deciding.
• It was next to the thing you actually went for.
• You never tried the alternatives, so there is nothing to prefer it to.
• Half the house likes it and the other half eats it.

Any of those produces twelve purchases. So does genuinely liking it. From the outside the twelve look identical, which is why no amount of counting gets you to the answer — the information was never in the count.

It runs the other way too. The best thing you ever bought might appear once, because the shop stopped stocking it and you never wrote down what it was called.

A household does not have one opinion

“We liked it” is usually a summary of a disagreement. One person liked it, one person ate it without comment, and one person has not tried it. Three different facts, flattened into a word.

A single rating hides exactly the part you needed. Knowing the house gave something four out of five tells you nothing about whether to buy it for the person who did not like it. Knowing that one particular person did not like it tells you what to do.

It is also why “favorite” is the wrong shape. A favorite is one winner per category. Real households run two — the one most people prefer, and the one the fussy eater will accept — and both are worth remembering.

None of which is visible to whoever does the shopping this week — making a list someone else can shop from →

Remember the decision, not every thought

The temptation here is to start cataloging. Don’t. A system that asks you to review your shopping is a system you abandon in three weeks, and the abandoned half is worse than nothing, because you no longer know what the gaps mean.

The test is narrow: will this change what I do the next time I am standing in front of the shelf? If not, it is not worth the keystrokes.

Six short sentences, and every one of them decides something. None of them is a review. You are not rating the product for strangers — you are leaving a note for the version of yourself who will be standing there in four months with no memory of this.

Most products need nothing at all. The ones worth a sentence are the ones where you chose between options and the choice was not obvious.

Opinions move, so the record has to

A child who refused something at four will eat it at six. A brand reformulates and the thing you liked is no longer the thing in the packet. A product you rejected on price becomes reasonable when the alternative goes up.

None of that makes the old opinion wrong. It made sense when it was formed, and what changed is the world around it. The mistake is treating a verdict as permanent — “nobody here likes that” hardens into a rule nobody revisits, and the house stops buying something everyone would now eat.

Which argues for a record that keeps its history rather than one field that gets overwritten. Changing your mind is not a correction. It is the interesting part, and it has a date.

A receipt remembers the purchase, not the verdict

Receipts are good evidence and they are worth keeping. A receipt establishes that on March 12 you bought something, at that store, for that price, under whatever description the till decided to print. None of that depends on anyone remembering it.

What it does not contain is anything you thought:

• Was it any good?
• Was it better than the one before?
• Did somebody refuse to eat it?
• Was it worth what it cost?
• Would you buy it again?

So the two halves come from different places and neither substitutes for the other. The purchase is read off the document. The verdict has to be said by a person, once, at a moment when they know the answer — which is usually not in the aisle.

What an accumulated purchase record holds, and what you can look up in it, is set out on its own page →

If the question is specifically what one product has cost you over time, that is here →

And if you are still deciding how to keep receipts at all, the ordinary methods compared →