How to set a grocery budget that survives contact with a supermarket
Food is the category most households want to control and the one where guessed budgets fail fastest. How to derive a number from your own spending, why the first figure is usually too low, and which tactics actually move it.
Food is the category households ask about most, for good reasons: it is large, it is genuinely variable, it happens several times a week, and unlike rent it is a place where a decision this afternoon changes the number. It is also the category where budgets fail fastest, and almost always for the same reason — the number was invented rather than derived.
Start by finding out what you actually spend
Ask anyone what their household spends on food per month and you will get a figure. Compare it to a recorded month and it is very often thirty to fifty percent low. This is not dishonesty, it is how memory works: the big weekly shop is memorable and the other things are not. The petrol station sandwich, the top-up loaf, the school snacks, the two bottles of wine, the delivery on a night nobody could cook — each is small and forgettable and there are twenty of them.
So the first month has no budget in it at all. Record every food purchase for a full month, including the awkward ones. Do it at the moment of paying rather than in a weekly catch-up session, because weekly reconstruction misses precisely the small frequent purchases that make up the gap you are trying to measure.
One month is the minimum and two is better, because months differ more than people expect: a birthday, a visit, a school holiday with children at home all day. A budget derived from a quiet February will fail in August for reasons that have nothing to do with discipline.
Separate groceries from eating out
These two behave nothing alike and merging them hides the only thing you might want to change. Groceries are relatively inelastic — a household of four needs feeding, and the range between careful and careless shopping is real but bounded. Eating out and delivery are highly elastic and are usually where the surprising money is.
Merged into "food", a household sees a large number, concludes they must buy cheaper vegetables, and misses that a fifth of it was delivery apps on tired weeknights. Split, the picture is immediately actionable, and it points at a scheduling problem rather than a shopping problem.
This is why Fambook lays out second-level categories flat rather than behind a menu: groceries, lunch, dinner, takeaway, coffee, drinks, snacks and fruit are each one tap, and the order adapts to what you actually record. The distinction is only worth having if making it costs you nothing at the till.
Setting the number
- Take the real recorded figure from your month or two of tracking.
- Check it against the calendar. If the recorded month contained something unusual, adjust for it — but adjust honestly. Every month contains something.
- Set the limit slightly below the real figure, not at your aspiration. Five to ten percent below is a target you might reach. Thirty percent below is a target you will miss in week two, and missing in week two is how budgets get abandoned.
- Set it on groceries and eating out separately, with different amounts and, if it helps, different attitudes. Many households find it works better to be strict about delivery and relaxed about the supermarket.
- Leave it alone for three months before adjusting again. One month is noise.
The reason to aim only slightly below is that the first successful month is doing most of the work. A limit you come in under is evidence that the system functions; a limit you blow immediately teaches you that the numbers are decorative, and once that is learned it is hard to unlearn.
A monthly limit in Fambook is per category and shows the remaining amount while you are entering an expense, which is the only moment the number can change anything. The bands are plain — on track, close, over — with no score and no advice.
Tactics that actually move the number
Advice in this area is abundant and mostly untested. What follows is sorted by how confident anyone can reasonably be about it.
Reliable, because it is arithmetic
- Compare unit prices rather than pack prices. The larger pack is often but not always cheaper per unit, and shelf labels in many countries are required to show the unit price precisely because the comparison is otherwise hard.
- Reduce waste. Food thrown away is money already spent with nothing received for it, and household waste estimates in wealthy countries are consistently large. A week of noticing what gets binned is usually more informative than a week of price comparison.
- Fewer, larger shopping trips. Every trip has a fixed cost in impulse purchases. This is one of the more robust findings in retail behaviour, and it also saves time.
Probably useful, but the evidence is thinner than the confidence with which it is repeated
- Shopping from a list. Widely recommended, and most of the supporting numbers circulating online trace back to marketing material rather than to studies. It plausibly helps, and the size of the effect is not well established.
- Never shopping hungry. There is laboratory work suggesting hunger increases acquisition, including of non-food items, but the studies are small and the effect on a monthly grocery bill has not been convincingly quantified.
- Meal planning. Works well for some households and collapses in others, and which one you are is not predictable in advance. Worth trying for a month; not worth feeling bad about abandoning.
Contested, and interesting anyway
Paying in cash is often claimed to reduce spending. There is a well-known experimental result from Prelec and Simester, published in 2001, in which participants bidding for tickets were willing to pay substantially more when told they would pay by credit card than by cash. It is a striking finding and it has limits worth stating: it was an auction for event tickets rather than a supermarket shop, the sample was small, and later attempts to replicate the general "payment method changes willingness to pay" effect have produced mixed results. Treat it as a reason to try a cash experiment for a month if you are curious, not as an established fact about your grocery bill.
What to do about a bad month
You will have one. A birthday, a visit, a bout of illness where nobody cooked anything for eight days. The instinct is to compensate the following month by setting a lower limit, and it is the wrong instinct: an artificially tight month follows an over month, is itself missed, and the whole thing turns into a cycle of restriction and rebound that ends with the app deleted.
Better: leave the limit where it is, note in one line what happened, and carry on. The number you care about is the average over a quarter, not any single month. Recorded reasons are what makes a ledger worth rereading — three over months with "visitors", "visitors" and "nothing in particular" attached mean completely different things, and only the third one is worth acting on.
When the budget is not the problem
If your recorded food number is already close to what a careful household in your area spends and it still does not fit, the problem is not the grocery budget. It is that food is absorbing a share of your income that reflects the size of your income, which is Engel’s original nineteenth-century observation and remains true. In that case, the leverage is in the large fixed costs and in income, and continuing to squeeze the food number will produce a lot of effort, a small saving and a fair amount of misery.
Knowing that is worth something in itself. One of the more useful outcomes of tracking a category properly is discovering that it is already tight, and being able to stop worrying about it.
Frequently asked questions
How much should a family of four spend on groceries per month?
There is no portable answer — it depends on where you live, what is available, dietary needs and the ages of the children. Published national averages exist in some countries and vary enormously by region within them. Your own recorded month is a far better reference point than any average, because it is the only one that reflects your actual prices and household.
Why is my grocery budget always wrong?
Usually because it was set from memory. Recalled food spending is commonly thirty to fifty percent below recorded spending, since the weekly shop is memorable and the twenty small purchases around it are not. Record a real month before setting any number.
Should groceries and eating out be one category or two?
Two. They respond to completely different things: groceries are relatively inelastic, while eating out and delivery are where surprising money usually is. Merged, you will conclude you need cheaper vegetables when the actual finding is four delivery orders on tired weeknights.
What do I do when I go over?
Nothing dramatic. Leave the limit where it is, write one line about why, and look at the quarterly average rather than the month. Compensating with an artificially tight following month produces a restriction-and-rebound cycle that ends in abandonment.
Does paying with cash really make you spend less?
It is plausible and not settled. The most cited evidence is a 2001 experiment by Prelec and Simester in which people bid more for tickets when paying by card than by cash — an auction, a small sample, and later work on the general effect has been mixed. Worth a one-month experiment if you are curious; not worth treating as fact.
How often should I change the limit?
Every three months at most. A single month is noise: a birthday, a visitor or a school holiday moves the number more than any change in your shopping habits will.
Try it for one month
Fambook gives a household one shared ledger: anyone can add an entry in seconds, every entry says who spent it, and the month adds up in one place instead of two. Records with no signal and syncs afterwards. Recording, categories, budgets, statistics, CSV import and export, sync and sharing for two people are free — the subscription only buys you less typing.