On keeping score.

You are lying in bed doing arithmetic. Not on purpose. You are running the list. The appointment you made, the early drop-off, the thing you canceled, the sink. And underneath the list is a conclusion that has already formed and is just waiting for you to agree with it: I am doing more of this than you are.

Almost everybody in a long relationship has had that evening. It is not a character flaw and it is not the beginning of the end. But it is worth understanding, because the thing that feels like the problem, the imbalance, is usually not the problem, and the thing that feels like the solution, proving it, reliably makes everything worse.

You are not keeping score because you are petty

Scorekeeping is what a fairness instinct does when it has no information.

Something in you is tracking whether effort in this relationship runs roughly both ways. That instinct is not the enemy: a person who has genuinely stopped caring whether the arrangement is fair has usually stopped caring about several larger things too. The instinct is fine. The data it is working from is terrible.

Because here is what you actually have access to: every single thing you did, in full detail, including how tired you were and what it cost you, and a thin, patchy, after-the-fact sample of what your partner did.

The asymmetry is a property of memory, not of your partner

This is the part that changes how the whole thing looks.

It is one of the sturdier findings in relationship psychology that when you ask each partner what proportion of the household work they personally do, and add the two answers together, the total comes out well above 100%. Both people are certain. Both people are honest. The numbers cannot both be right.

The reason is not that somebody is lying. It is that you remember your own contributions better than you remember anybody else’s, because you were present for all of yours and for a fraction of theirs. Your effort comes with the full experience attached: the reluctance, the interruption, the fact that you were already tired. Theirs arrives, at best, as a finished result you noticed in passing. Often it arrives as nothing at all, because the entire point of a lot of domestic labour is that when it is done well, nobody notices anything happened.

The psychologist Daniel Wegner described long-term couples as building a kind of shared memory system, each person becoming the one who holds certain things so the other does not have to. It is one of the real advantages of being a pair. It also means that a great deal of what your partner carries is, structurally, invisible to you. Not hidden, not unappreciated on purpose, just outside your view.

So when your instinct reports back that you are doing more, it is reporting accurately on the evidence available. The evidence is systematically incomplete, in a way that is not anybody’s fault, and in the same direction for both of you at once.

Why proving it does not work

The natural next move is to get better evidence. Write it down. Build the case. Bring the receipts.

And it does not work, for a reason worth being precise about: a tally between two people is only ever read by the person who is behind.

Whoever is ahead does not experience it as information. They experience it as an accusation with a spreadsheet attached, and they respond the way anybody responds to an accusation: by disputing the methodology. That is not defensiveness for its own sake. It is that the tally genuinely does not capture what they did: the thing they handled that never got written down, the week they carried alone, the fact that the twenty-minute task took them ninety.

Now you are not talking about your life. You are litigating a dataset. And every hour spent litigating a dataset is an hour in which the original question, do I feel like this is landing on somebody, goes unanswered.

Worse, the tally changes what the work is for. When something you do becomes a point you are owed, you have converted a gift into an invoice. People can feel that conversion happening, even when nobody says it out loud, and it is corrosive in a way the original imbalance was not.

But “just stop counting” is bad advice

The usual advice at this point is to let it go. Assume good faith. Stop counting.

For a lot of people that advice is quietly unfair, because the person being asked to stop counting is very often the person doing more, and “stop noticing” is not a solution to being unseen, it is a request to be unseen more quietly. Resentment does not dissolve when it is told to. It goes somewhere.

The instinct is not the problem. The scoreboard is a bad answer to a real question. So the useful move is not to suppress the question. It is to answer it with better information than your memory can give you.

What actually helps

1. Guess first, then look

Before either of you says anything: separately write down how many things you think your partner did for the two of you this week, and how many you think you did. Then compare both guesses to whatever record you have.

This is the single most useful exercise we know of, because the gap is almost never a surprise about your partner. It is a surprise about your own estimate, and an estimate you produced yourself thirty seconds ago is very hard to argue with.

2. Notice out loud, in specifics

“Thanks for everything you do” does almost nothing. “You dealt with the insurance thing so I never had to think about it” does a great deal, because it proves the invisible work was actually seen. Specificity is the whole active ingredient. If you are not sure what the invisible work in your house even is, the invisible labor list is fifty-odd rows of it on one printable page.

3. Let the person who received it say what it was worth

This is the rule that turns a tally into something else entirely. The person who did the thing does not get to decide what it was worth. The person it was done for does.

It sounds like a small procedural detail. It is the difference between a record of your own effort, which is a case you are building, and a record of what actually landed, which is something you built together and which neither of you can inflate alone.

4. Have the conversation on a schedule, not at midnight

Nothing about division of labor has ever been usefully resolved during the argument that surfaced it. Pick a time each week, look at the week together, and let the specific complaint wait for it. Most of them dissolve on the way; the ones that survive to Sunday are the real ones and deserve the attention.

If you want the questions and the order to run that conversation, we wrote them down: the weekly check-in. It works on paper and it does not need this or any other app.

5. Ask what the week actually was

A newborn, a move, a bad month at work, an illness, a loss. During those stretches the ordinary reading of a week is simply wrong, and holding either person to their normal standard is a way of being angry at someone for something that is happening to them.

The distinction that matters

Everything above rests on one line: a record and a scoreboard are not the same object.

A scoreboard exists to rank two people. It has a winner, it is read by whoever is losing, and its purpose is settled the moment somebody is ahead.

A record exists so that what happened is visible to both of you. It has no winner. Its purpose is to replace two unreliable memories with one thing you can both look at, so that the fairness instinct, which is not going anywhere, finally has something real to work from.

The distinction is not automatic and it is not preserved by good intentions. It has to be built in deliberately, and most things in this shape fail at it.

Where we come into it

We build one of these, so treat the next three paragraphs with the suspicion they deserve.

Every Little Thing is a record two people keep together, and it is designed around exactly the failure described above. Nobody grades anybody — rule three, taken one step further: on paper the person who received it decides, and here neither of you does. The app weighs each moment, one to three hearts with the why in a sentence, so no number is ever one person’s opinion of the other. Nothing in it ranks the two of you. Once a week the review puts the actual record between you instead of two competing memories. And you can turn the comparison off and keep the whole record.

If you arrive at it wanting a verdict on your partner, you will find it quietly unhelpful. That is not an accident; it is the main design constraint, and most of the arguments above are arguments we had with ourselves while building it.

And you do not need it. The five things in the previous section work on paper, in a shared note, or in a conversation on Sunday evening. If you take the guessing exercise and nothing else from this page, that is a good outcome. It is the part that changes the most for the least effort, and it costs nothing.