How many recipes do you actually need?
You have saved hundreds. You cook the same handful. Guess the real number before you scroll, then build the twelve that would carry an ordinary month.
The gap between what you saved and what you cook is the whole article.
And the smaller number turns out to be much smaller than most people would admit to.
Somewhere on your phone is a folder of recipes you fully intended to make. Meanwhile Tuesday happens and you cook one of about six things, again. It is easy to read that as failure, and it is not. A survey of 2,000 Americans put the number of recipes people know by heart at 15, and a separate, larger survey found 86% of adults repeat the same meals at least some of the time. This does not mean people are lazy. Most households run on a small group of familiar meals. Below you can guess the number first, see the real one, and then build a starting set of twelve: five for weeknights, three for when there is no time, two for weekends and two for when people come round. Twelve is a template to adjust, not a magic number, and there is a way to work out your own size further down.
Guess first
Nothing leaves this pageHow many recipes do you think the average person can cook from memory, with no recipe in front of them?
Fifteen by heart, hundreds saved
The figure comes from a survey of 2,000 Americans run by OnePoll for HelloFresh in January 2023: fifteen recipes known well enough to cook without looking anything up, and, among Americans who cook regularly, about ten home-cooked meals a week. Meals, across the whole week, not ten dinners. A week only contains seven of those, and the difference matters for the arithmetic later on. The 86% comes from somewhere else, a larger HelloFresh report with Wakefield Research covering 5,000 adults, which found that most people repeat the same meals at least some of the time.
Sources: OnePoll for HelloFresh, 2,000 Americans surveyed 5 to 6 January 2023, reported by StudyFinds, for the 15 recipes and the 10 home-cooked meals a week. The 86% is from the HelloFresh State of Home Cooking report with Wakefield Research, 5,000 US adults, 2025. Both are commissioned surveys of self-reported answers, which is worth holding in mind: people estimate their own repertoire generously, and a company that sells recipe boxes paid for the questions.
Now put that next to the folder. Nobody has counted how many recipes the average person has saved, and we are not going to pretend a figure exists. What we can say is that the two numbers are not in the same league, and one food blog put the mismatch nicely in a headline about having four hundred saved recipes and cooking the same six.
The observation, from a meal-planning blog rather than a study: eatwellplanner.com.
Home-cooked meals a week among Americans who cook regularly. That is breakfasts, lunches and dinners together, so call it five to seven dinners for most households. Against fifteen recipes known by heart, something has to come round again inside a fortnight, and that is with a repertoire on the generous side.
Why repeating yourself is not a failure
The saved folder is full of recipes that took somebody an afternoon to photograph. The Tuesday version of you has forty minutes, a tired child and whatever is in the fridge. Those are two different jobs, and only one of them is happening tonight.
There is also a practical reason the repertoire stays small, and it has nothing to do with willpower. A recipe you know by heart costs almost nothing to run: no reading, no checking, no shopping list surprises, and you can start it while thinking about something else. A new recipe wants attention at exactly the hour you have least of it. Repeating familiar meals saves time and attention when you are tired.
Repeating meals is fine. The problem is that many people never chose their regular meals on purpose. Ask someone why they make the same pasta every Wednesday and the honest answer is usually that it happened, not that they picked it.
Your working set of twelve
Runs in your browser · nothing is sent anywhereOnly things you can already cook. Not the ones you mean to learn, not the ones you saved. Leave a slot empty if nothing comes to mind, because an honest gap is the most useful output here.
WHAT YOU HAVE
The shape of the twelve, and why those four groups
The split is ours, not a finding, and it is a starting template rather than a prescription. It comes from one observation: a repertoire fails not because you are short of recipes, but because you are short of the right kind of recipe at the moment you need it.
The backbone. Forty minutes or less, ingredients you tend to have, no special trip. Five is enough that a week does not repeat, and few enough that you know all five cold.
Fifteen minutes, from cupboard staples, edible when you are past caring. These are the ones that stop the takeaway decision, and most people have two rather than three.
The ones worth an afternoon. Not for Tuesday, and the mistake is filing them with the weeknight ones and then feeling bad about never making them.
Scales up, can be half made in advance, does not need you standing at the hob while people arrive. Two is genuinely plenty. Nobody is auditing your rotation.
Here is the honest arithmetic, because it is easy to get wrong. If you cook six dinners a week, a month is about 24 dinners. Twelve dishes across 24 dinners means each one turns up roughly twice, which most households find comfortable. Cook every night and it is 28 to 31 dinners, so twelve dishes appear two or three times each, and the ones you like will drift towards three.
Work out your own number instead of taking ours. Multiply the dinners you actually cook in a week by four, then divide by how many times a month you are happy to eat the same thing. Six dinners a week and a tolerance of two repeats gives twelve. Cooking every night with the same tolerance gives fourteen. A tolerance of three repeats brings it down to eight or ten, which is a perfectly good working set. Twelve is where a fairly common set of answers lands, not a rule.
A few things quietly change that maths, and they all push the number down, which is good news. Cooking a double batch and eating it twice makes one dish cover two nights. Leftovers do the same. A household of two eats through a repertoire more slowly than a household of five, because there is more of everything left over.
One filled set, and two weeks of it
Abstract until you see one. Here is a working twelve for a household of four, cooking six nights a week.
WEEKNIGHT (5) Pasta with tomato sauce · Chicken and rice · Tacos Stir fry · Sausages and mash FAST (3) Omelette and salad · Beans on toast · Soup from the freezer WEEKEND (2) Roast chicken · Lasagne GUESTS (2) Curry · Baked potatoes with fillings
Week one. Monday pasta, Tuesday the roast chicken leftovers, Wednesday tacos, Thursday omelette, Friday sausages, Sunday roast chicken. Six dinners, five dishes cooked, one night carried by leftovers.
Week two. Monday stir fry, Tuesday lasagne leftovers, Wednesday chicken and rice, Thursday beans on toast, Friday pasta again because it is Friday, Sunday lasagne. Across the fortnight only the pasta appeared twice, and nobody noticed.
That is the whole point of the arithmetic above. With leftovers doing real work, twelve dishes cover a month at about two appearances each.
The shape should change with the household. The four groups are the part to adapt, and swapping one out is expected rather than cheating.
| Who is cooking | A shape that works better | Why |
|---|---|---|
| One person | 4 weeknight · 4 fast · 2 weekend · 2 batch | Portions rarely divide neatly, so a batch group that freezes matters more than a guests group |
| Family with young children | 6 weeknight · 3 fast · 2 weekend · 1 guests | Weeknights carry almost everything, and the guests slot is mostly theoretical for a few years |
| Vegetarian household | 5 weeknight · 3 fast · 2 weekend · 2 guests, built around three bases | Building on beans, grains and eggs gives more variety from the same shopping |
| Packed lunches to make | Swap the guests group for 2 that travel cold | Guests happen a few times a year. Lunch happens every day |
| Small freezer, no batch cooking | Swap a weekend dish for a fifth fast one | Without leftovers, the week needs more short options rather than more ambitious ones |
Look at it once a season. Not to redo it, just to notice what stopped happening. Summer sets and February sets are genuinely different, and one or two swaps in March and September is the whole maintenance job.
The empty slots are the useful bit. If you could not fill in the fast group, that is where your week keeps breaking, and one new recipe there is worth more than another twenty saved into the folder.
What to do with the other four hundred
Leave them alone, mostly. A saved recipe folder is not a to-do list you are failing, it is a reference library and a record of things that once looked good. Both of those are fine things to own.
The only change worth making is one that costs seconds rather than an afternoon of sorting.
- Promote by cooking, not by tidying. Add a recipe to the main list only after you have cooked it twice successfully. That is the only way in, which is what keeps the set honest.
- Save with one word attached. Quick, weekend, guests, freezes well. That word is what makes the folder searchable later, and it is nearly all of the value of writing anything down.
- When something graduates, something leaves. Not deleted, just out of the working twelve. A set that only grows stops being a set.
- Stop counting the folder. Four hundred saved is not a debt. Nobody is going to look at it, including you, and that is completely fine.
We have written about the storage side of this before, in why the recipes you pin never get cooked and the recipe screenshots living in your camera roll. This page is about the number rather than the filing.
Fair questions
"Twelve sounds boring"
It is a floor, not a ceiling. Twelve is what carries the ordinary weeks, so that trying something new is a choice you make on a Saturday rather than a gamble you take at six on a Wednesday. Most people's cooking gets more adventurous once the boring part is solved, not less.
"Does a variation count as a separate dish?"
If it changes what you buy or how long it takes, yes. Pasta with tomato sauce and pasta with pesto are one dish wearing two hats, because the shopping and the effort are the same. Pasta with a slow ragu is a different dish, because it needs a different afternoon. The test is not the recipe card, it is whether the two feel different to cook on a Tuesday.
"I cook for people with very different tastes"
Then your five weeknight slots are doing more work than anybody else's, and the honest move is to pick ones that split easily: a base plus toppings, a sauce plus two pastas. That constraint is worth building the set around rather than treating as a failure of the set.
"Fifteen from memory seems high to me"
It seemed high to us as well. Remember what it is: a self-reported figure from a commissioned survey, and people are generous about their own abilities. That is exactly why the tool above asks for things you have actually cooked, not things you believe you could.
Where we come into it
Your set is the useful thing here, and a paper copy is easy to lose. dEssence is somewhere to keep it, along with the links and the screenshots and the note about which one freezes well, and then ask Tracy in ordinary words, along the lines of what are my fifteen-minute ones, and get the answer back rather than scrolling a folder of four hundred.
Being clear about what it is not: it does not write recipes, it does not plan your week for you, and it does not know what is in your fridge. If you want a week arranged out of the dishes you already cook, we built that as a separate thing: the seven-day what's for dinner generator.
You do not need more recipes. You need twelve you actually chose.
The saved folder was never the working set, and treating it as one is what makes an ordinary Tuesday feel like a failure.
Where the numbers come from, and what is ours
The survey figures. 15 recipes known by heart and about 10 home-cooked meals a week, meals rather than dinners, come from a OnePoll survey of 2,000 Americans commissioned by HelloFresh, carried out on 5 and 6 January 2023, reported by StudyFinds. The 86% is from a different and larger study, the HelloFresh State of Home Cooking report with Wakefield Research, 5,000 US adults in 2025. All of it is self-reported and commercially commissioned, so read it as a good indication rather than a precise measurement.
How many recipes people save is not a number we have. We looked and did not find a credible one, so this page does not claim it. The four hundred figure appears only as a blog headline and is quoted here as an image, not as data.
The twelve and the four groups are ours. Five, three, two and two is an editorial starting template, not a finding from anywhere. It assumes roughly six cooked dinners a week and a tolerance of about two repeats a month, and the article shows how to work out your own number instead. Note also that the survey figure is ten home-cooked meals a week, not ten dinners, so any arithmetic built on ten dinners would be wrong.
No crowd average. We do not collect your guess, so we cannot show you what other readers said. Everything on this page runs in your browser, nothing is uploaded, and closing the tab throws it away. The card is drawn here and saved straight to your device.