The Grocery Number: What Every Household Gets Wrong

Ask a household what it spends on groceries and you will get a number given with confidence. Then count it.

Four households, four underestimates, median 26 percent. The two-teacher household in our Household Audit guessed $950 and spent $1,240. The household in our first Ledger guessed $1,050 and spent $1,412.

The only household that came close was the one-income household, where the adult who is at home keeps a running total in her head and was accurate to within $15 on day nineteen. That is not a coincidence, and I will come back to it.

The gap is in the small trips

Here is the thing that makes this interesting rather than just embarrassing: the households were not wrong about the big shop.

Every one of them recalled the planned trips accurately, within a few dollars. The planned shops are memorable — they take an hour, they involve a list, somebody complains about the car park.

What is missing from the household’s own accounting is the nine to fifteen unplanned trips: milk, bread, the thing for the lunchbox, the thing somebody promised to bring. Average value $26 to $39. Duration eight minutes. Cognitive footprint approximately zero.

Between a third and a half of the monthly total, in every household, in a category the household does not experience as shopping.

Why the small trip costs $31

Nobody spends $31 on milk. What happens is the trip has a stated purpose — the milk — and a set of unstated ones that get resolved on the way past: the good crackers, something for tomorrow, a bag of the mandarins that were there and looked fine.

Every one of those is a reasonable decision. The problem is that a household making a hundred reasonable decisions a month, four at a time, in eight-minute windows, in no way experiences that as a hundred decisions. It experiences it as buying milk, eleven times.

This is the same structure as the mental load and the same structure as the subscription gap: the individual instance is too small to notice and the aggregate is never presented. It is the most common shape of household money that gets away from people — a season of youth sport is the same thing at a larger denomination — and it is why nobody’s budget survives contact with a month.

What actually changed it

Two households reduced their grocery spending during our reporting, both without anyone reporting feeling deprived. Neither did it by buying different things.

One moved from thirteen trips a month to six. Their basket contents barely changed; their monthly total fell by about $180. The other set a rule — one trip between the planned shops, and if something runs out after that, the household eats something else — and fell by roughly $140.

The marginal cost of a trip is not the item you went in for. It is the four other items, and the only reliable way to not buy them is to not be there.

Planning, and the thing that breaks it

Everybody knows meal planning is the answer. Most of our households had tried it and stopped.

They stopped because a plan written from scratch produces a shopping list that duplicates what is already in the cupboard and ignores what is about to go off, so the household buys rice it owns and throws away chicken it forgot. That is also precisely why generated meal plans failed in eleven of the fourteen households in our reporting on AI at home: a plan that does not know what is in your fridge is a plan that shops against your own inventory.

The households where planning worked spent four minutes on stock first. Look in the fridge, look in the freezer, write the five things that must be used, plan around those. The Sunday reset household does this as a matter of course, in the twelve-minute block after the calendar, which is the correct order — you cannot plan food until you know which nights nobody is home.

The woman who was accurate to $15

The one-income household is worth returning to, because the obvious reading of it is wrong.

She is not more disciplined than the other households. She is under a constraint they are not: at $58,000 for four people, the grocery line is the only flexible thing in the budget, and it functions as the household’s shock absorber. When the car needed $214, the grocery number moved. She tracks it precisely because it is the only lever she has.

That is not an argument that every household should track groceries to the dollar. It is an argument for the much smaller thing, which is that no household can manage a number it has never once counted — and that counting it for thirty days, exactly once, ends the argument about what it is.