What's the fastest way to decide what to cook tonight?
Three questions, sixty seconds, exactly three answers. Not thirty recipes from a database you have never opened, three dishes out of the ones you can already make.
It is half past five and the question is not what to cook. It is how to stop deciding.
Every article about this hands you more options. More options is the problem.
You are not short of dinner ideas. You are standing in a kitchen with a saved folder of four hundred recipes, a fridge with about nine usable things in it, and no energy for a decision. What ends this quickly is not inspiration, it is elimination: how long have you got, how much is left in you, and what needs using up. Three questions, asked in that order, cut the field down to a handful. Below you can type in the dishes you already cook, answer the three, and get exactly three options back. Three, not thirty, because thirty is where you came in.
The sixty-second decider
Runs in your browser · nothing is sent anywhereOne dish per line. Anything after the name, separated by |, is optional: minutes, how much effort it takes (easy, medium or hard), the main ingredients, and how many pans. Plain lines like Chilli 40 still work, they just give the tool less to go on.
HOW LONG HAVE YOU GOT
HOW MUCH IS LEFT IN YOU
ANYTHING THAT NEEDS USING UP
You are not short of ideas
Search for what to cook tonight and you get lists. Fifty easy dinners, thirty quick suppers, a hundred things to make with chicken. At half past five you probably have enough ideas. The hard part is choosing one.
Notice what actually happens in that moment. You open the phone, scroll a bit, feel slightly worse, and cook the same thing you cooked last Tuesday. That is not a failure of inspiration. It is what a tired brain does when handed thirty similar options with no way to rule any of them out.
So the fastest method is to rule out the dishes that do not fit tonight. Ruling out needs limits, and you already know all three of them before you pick up your phone.
The three questions, in this order
Fifteen, thirty, or an hour and it does not matter. Ask this first because it eliminates hardest and it needs no thought at all. You know the answer before you open the fridge.
Different from time, and the one people skip. There are evenings with forty minutes and nothing in the tank, and that combination rules out anything with a step you have to concentrate on.
The one thing in the fridge that will not survive another day. Not a full inventory, one item. It turns an open question into a much smaller one and stops you shopping for a fourth time this week.
The order matters more than it looks. Time first because it cuts the most for the least effort. Energy second because it is the honest constraint people lie to themselves about. Ingredients last, because starting from what is in the fridge is how you end up standing there with a limp pepper, thinking.
Why three options and not thirty
There is real evidence that too many similar options makes choosing harder, and it comes with a caveat that usually gets left out. The short version: a famous supermarket experiment found people bought jam far more often from six jars than from twenty-four, and a later pooling of around fifty similar experiments found the average effect was near zero. Too many options make choosing harder in some situations, but not in all of them.
Choosing dinner while tired ticks all three of the conditions where it tends to show up: the options are similar, you have no criteria ready, and you are deciding at the worst hour of the day. We set that out properly, caveat and all, on a separate page about list size.
The practical claim here is smaller and needs no study. Three options are easy to compare. A list of thirty usually sends you back to scrolling.
Some numbers we deliberately left out
Every article on this subject quotes the same handful of statistics. Roughly two hundred food decisions a day. A couple of hours a week spent deciding what to eat. A precise number of arguments about dinner per year.
We went looking for the sources. Each one led to another article quoting another article, and we did not reach an original study for any of them. So they are not on this page. They would have made the argument sound better and they would have been us passing along something we could not stand behind.
What we do have is a survey of 2,000 Americans, run by OnePoll for HelloFresh in January 2023, which found people know about 15 recipes by heart and that regular cooks average around 10 home-cooked meals a week. Meals across the week, not dinners, since a week only holds seven of those. A separate and larger HelloFresh report with Wakefield Research, covering 5,000 adults, found 86% repeat the same meals at least some of the time. Both are commissioned and self-reported, so not perfect, but at least traceable to a described piece of research.
Sources: OnePoll for HelloFresh, 2,000 Americans, 5 to 6 January 2023, via StudyFinds, and the HelloFresh State of Home Cooking report with Wakefield Research, 5,000 US adults, 2025. We look at what those figures mean for your own repertoire in how many recipes do you actually need.
If two of the three options look equally good, take the one with fewer pans. On a weeknight the washing up is part of the recipe, and it is the part you will still be doing at nine o'clock.
The sixty seconds, step by step
Without the tool, on paper, in a kitchen. This is the whole method.
- Ten seconds. Say the time out loud. Not the time you wish you had. Fifteen minutes, half an hour, or plenty.
- Ten seconds. Rate the tank: nothing, some, or actually fine. Be honest, because this is the question that decides whether the ambitious option is real.
- Fifteen seconds. Open the fridge and name one thing that needs using. One. Close the fridge.
- Fifteen seconds. From the dishes you know, name three that survive all three answers. Out loud, or on paper. Three, then stop.
- Ten seconds. Pick the one with the fewest pans, and start. The decision is over.
One evening, all the way through
Twelve dishes written down once, on a Wednesday, at twenty-five past five.
chilli | 40 | medium | mince | 2 pans omelette | 10 | easy | eggs | 1 pan fish pie | 60 | hard | fish | 3 pans pasta, garlic, chilli | 15 | easy | pasta | 2 pans stir fry | 20 | easy | whatever | 1 pan lasagne | 90 | hard | mince | 3 pans roast chicken | 55 | easy | chicken | 1 pan soup from freezer | 12 | easy | soup | 1 pan sausages and lentils | 35 | medium | sausages | 2 pans rice and fried eggs | 15 | easy | rice, eggs | 2 pans tacos | 30 | medium | mince | 2 pans baked potatoes | 70 | easy | potatoes | 1 pan
Tonight’s answers. Twenty-five minutes. Nothing left in the tank. The mince needs using.
| Question | What it rules out | Left standing |
|---|---|---|
| Twenty-five minutes | Chilli, roast chicken, fish pie, sausages, lasagne, baked potatoes, tacos | 5 of 12 |
| Nothing left in the tank | Nothing outright, but it pushes the easy ones up and anything hard down | 5, reordered |
| The mince needs using | Nothing, it promotes instead | No match under 25 minutes |
The three that come back. Omelette, ten minutes, easy, one pan. Rice and fried eggs, fifteen minutes, easy. Stir fry, twenty minutes, easy, one pan. The tool also says plainly that nothing under twenty-five minutes uses the mince, so the honest answer is that the mince waits for tomorrow or goes in the freezer tonight.
The pick. All three fit, so the tie-breaker is the washing up: omelette, one pan. Total time spent deciding, about forty seconds, and no phone was scrolled.
When nothing fits, have an answer ready. This is the rule most methods skip. Pick one emergency dinner now, while you are calm, and write it at the top of your list: beans on toast, eggs on rice, the soup in the freezer. When the tool comes back with nothing, you do not reopen the question, you cook the emergency dish. Deciding what to do when you cannot decide is a job for a Sunday, not for half past five.
Two more awkward cases. If the ingredients simply are not in the house, drop the third question entirely and pick on time and energy alone, then buy nothing on the way home except what that dish needs. And if two of you want different things, do not merge the lists: one person answers the three questions and picks, and the other one picks tomorrow. Alternating beats negotiating at half past five, every time.
The rule that makes it work is that you stop at three. Not a longlist, not a browse, not one more scroll to check nothing better exists. Something better always exists and it is not going to be cooked tonight.
Setting it up once so you stop having this conversation
Doing this from scratch every night still costs a minute a night. Three small bits of setup make most of the minute go away permanently.
- Write down the dishes you can already cook. Once, in one place, with rough minutes next to each. Most people have between ten and twenty and have never seen them in a list. Seeing them written down is the single biggest change here.
- Have three you never have to think about. The fifteen-minute cupboard ones. Their whole job is to exist on the evening when nothing else is possible, and they are the difference between dinner and a delivery app.
- Decide once a week, not once a night. Not a full plan. Just three named nights, which turns four decisions into one. We built a generator for exactly that, and it uses your dishes rather than inventing recipes.
Where we come into it
Every free meal planner on the internet will hand you a week of somebody else's recipes. That is a different product and it solves a different problem, which is a shortage of ideas that you do not have.
dEssence holds the things you saved: the links, the screenshots of recipes, your notes about which one everybody actually eats. You ask Tracy the way you would say it out loud, along the lines of what can I make in twenty minutes with the mince, and the answer comes back out of your own saved things rather than from a database.
The honest limits: it does not know what is in your fridge, it does not write recipes, and it cannot see what you cooked yesterday unless you told it. The tool on this page works the same way on purpose. It only ever hands back what you typed in, because a suggestion you have never cooked is not a fast decision, it is a new project.
Speed comes from ruling things out, not from finding more ideas.
Time, energy, one ingredient. In that order, and stop at three options.
How the decider picks, and what it cannot know
It only knows what you typed. There is no recipe database here, no catalogue and no suggestions from outside your own list. If a dish is not in the box it cannot come out of the tool, which is deliberate: the fast decision is always between things you have made before.
The line format. Everything after the name is optional and separated by bars: chilli | 40 | medium | mince | 2 pans. A bare number is minutes, easy, medium or hard is the effort, a number followed by pans is the washing up, and anything else is treated as ingredients for matching. A plain chilli 40 still works. Lines with no minutes stay in the running and are marked as having no time written down, because guessing a duration on your behalf would be inventing information.
How it ranks. The tool removes dishes that take longer than the time you have. It then prefers dishes marked easy when your energy is low, allows the hard ones when the evening is good, and pushes anything mentioning the ingredient you need to use straight to the top. If you did not write an effort level, it falls back to length as a rough stand-in, which is exactly what it is: shorter is not always easier, and only you know which of your dishes is fiddly. Fewer pans wins a tie, then the remaining ties are broken at random, which is why the reshuffle button gives you a different three.
Exactly three, or fewer, honestly. If your constraints only leave two, it shows two and says which constraint to relax rather than padding the list out with something that does not fit.
Numbers on this page. The recipe figures come from a OnePoll survey for HelloFresh of 2,000 Americans in January 2023, via StudyFinds. The choice-overload evidence and its caveat are set out on our page about list size, sourced to a summary of the jam study and a summary of the replication picture.
Nothing leaves the page. No account, no upload, no storage, and closing the tab throws away your list.