The finding

John Gottman’s lab spent decades watching couples interact and coding what they did. One of the more durable results concerns what he called bids — small attempts to get the other person’s attention. A remark about the weather. Showing you something on your phone. Asking if there’s milk.

Each one is trivially small. What he found was that the rate at which partners turned toward those bids — answered, engaged, acknowledged — separated couples who were still together years later from those who weren’t, by a wide margin.

Not the arguments. Not the holidays. The answering.

Worth being precise about what this is: observational research on couples in a lab, with the usual limits. But the direction has held up across a lot of replications and it matches what people report about their own relationships.

Why a favour is a bid with a cost

“Can you pick up the parcel on Wednesday?” is a bid with something extra attached: it costs the other person something, and it can be refused.

Which is exactly why it carries information. Answering a remark about the weather is cheap. Rearranging half an hour is not. When it happens, it means something; when it’s ignored, that means something too.

And unlike a grand gesture, it happens forty times a month.

Why they go wrong

Small asks fail in ordinary, fixable ways:

They arrive at the worst moment. A text at 18:10 asking for something at 19:00 lands as pressure, regardless of what it says. The same request on Sunday would have been nothing.

They’re never actually answered. “Maybe, I’ll see” is not an answer, and now the asker is carrying it all week. The unanswered ask is worse than a no — a no is information.

One person does all the asking. Then asking starts to feel like being a burden, and they stop — which reads to the other person as everything being fine.

They vanish. Agreed on Tuesday, forgotten by Thursday, and the person who asked has to choose between chasing and dropping it. Both are corrosive.

What helps

Ask early. Not because of politeness — because an ask with lead time is a request, and an ask with an hour’s notice is a demand. Same words, different thing.

Answer with one of three things. Yes, no, or a specific counter (“not Wednesday, but I could Thursday”). Never “we’ll see”. A clean no is a good outcome; it lets the other person solve the problem another way while there’s still time.

Say when it’s done. Small and disproportionately effective. It closes the loop, and it means the asker stops holding it.

Notice who asks. If it’s always the same person, the other one probably has needs they’re not voicing — often because they’ve decided their needs are less important, which is its own problem.

Not everything is a date

This is why Hours for Two has favours as their own thing, separate from plans. A request to be collected from the station isn’t a date; it shouldn’t compete with one, and it shouldn’t be a text that arrives at the wrong moment.

Each favour has a state — asked, promised, done, or declined — so the answer is explicit and nothing lives unresolved. The stats screen counts them, which sounds trivial until you notice a month where one person did eleven and the other did one.

The part that’s easy to get wrong

None of this means grand gestures are pointless. A weekend away is lovely.

It means the weekend away is not what’s holding things together, and it won’t compensate for four months of unanswered small asks. The arithmetic is simply against it: one gesture a year against a thousand small moments.

Be good at the small ones. The big ones are a bonus.

TL;DR

Gottman’s bids research: turning toward small attempts at attention predicts who stays together
A favour is a bid with a cost attached, which is what makes it informative
They fail by arriving late, going unanswered, being one-sided, or vanishing
Ask early; answer yes, no, or a specific alternative — never “we’ll see”
Grand gestures are nice and they don’t substitute for the small ones