How to decide where to eat, without the 20-minute debate
Seven methods that actually end it — and an honest note on where each one falls apart.
The hard part of choosing a restaurant with other people isn't a lack of options. It's that nobody wants to be the person who picked. Suggest a place and you own the outcome: if the wait is 40 minutes or the food is mediocre, that's on you. "I'm fine with anything" is almost never indifference. It's risk-avoidance.
That's why adding more options makes it worse, and why the group that can't pick a restaurant can usually pick one instantly once someone says "okay, we're doing the first place with a table." The problem was never information. It was that no one had agreed on how the decision would get made.
The one rule that fixes most of it: agree on the method before anyone names a restaurant. Once a name is on the table, every proposed method looks like an attempt to make that name win or lose, and you end up arguing about the rules instead of dinner.
Seven methods, ranked by how often they work
Method 1
Nominate, then vote
Everyone names one place — no discussion, no reacting to each other's picks. Then everyone votes. Highest count wins.
Works because nominating simultaneously stops the anchoring problem, where the first suggestion quietly becomes the thing everyone argues about. And a vote spreads the blame across the group, so no one person owns a bad night.
Falls apart when the vote ties, which happens constantly in groups of four or fewer. Have a tiebreaker decided in advance or you've just moved the argument.
Method 2
Veto rounds — everyone gets one "no"
Put five or six places on a list. Going around the group, each person strikes exactly one. Whatever survives is where you're eating.
Works because it's built for constraints rather than preferences. The person who can't eat shellfish, the one who went to that place for lunch, the one who's genuinely broke this week — each gets to remove their dealbreaker without having to explain or defend it in front of everyone.
Falls apart when the group is large. Six people with six vetoes will strike a six-item list down to nothing. Rule of thumb: the list needs at least two more options than there are people.
Method 3
Head-to-head elimination
Bracket the options in pairs. Vote on one pair at a time — this place or that place — and the winner advances until one is left.
Works because people are far better at comparing two things than ranking six. A six-way choice invites paralysis; "Thai or burgers?" gets an answer in about two seconds, and stacking those answers gets you a real winner.
Falls apart when you run it by hand. Doing it verbally around a table takes longer than the original argument, and the seeding — who faces whom first — genuinely changes the outcome, which people notice and object to.
Method 4
Random pick — wheel, dice, or a coin
Write the shortlist down and let chance choose. A spinning wheel, a rolled die, a coin flip between two finalists.
Works because it removes responsibility completely, which is the actual thing stopping the group. Nobody picked, so nobody can be blamed, so nobody has to hedge. It's also the fastest method on this list by a wide margin.
Falls apart when someone doesn't like the result — and then you discover the group never really agreed to be bound by it. Randomness only works if everyone commits before the spin. It also ignores that preferences differ in strength: a place one person actively dislikes and four are neutral on has exactly the same odds as everyone's favourite.
Method 5
Rotating dictator
One person chooses, unilaterally, with no vote. Next time you eat together, it's someone else's turn.
Works because it's the only method here with zero overhead, and over a few months it's arguably the fairest — everyone gets their favourite occasionally, instead of the group permanently converging on the inoffensive middle that nobody loves.
Falls apart when the group isn't a recurring one. It only balances out over repeated dinners, so it does nothing for a one-off with people you rarely see. It also collapses if anyone quietly skips their turn to avoid the pressure — which is the same fear that started all this.
Method 6
Constraints first, options second
Before naming anywhere, agree the boundaries out loud: budget, maximum travel, how long you've got, any dietary non-negotiables. Only then start suggesting.
Works because most restaurant arguments are actually unspoken disagreements about price or distance wearing a costume. Someone rejecting three suggestions "for no reason" is usually someone who doesn't want to say they can't afford the second one. Stating the constraints first lets that be a rule instead of an awkward veto.
Falls apart when it's the whole method. Constraints narrow the field, they don't choose — you still need one of the methods above to finish the job.
Method 7
The two-option rule
One person — whoever cares least — picks exactly two places. The group chooses between them. That's the entire choice.
Works because it splits the work in two. Shortlisting is the tiring part, choosing is the easy part, and this hands each to whoever minds it less. The person who picked the two doesn't own the outcome, because they didn't make the final call.
Falls apart when the shortlister has an agenda. Two options where one is obviously the "real" suggestion isn't a choice, and people can tell.
Which one should your group use?
Pick on group size and how often you eat together, not on which sounds fairest in the abstract:
| Your situation | Use this | Why |
|---|---|---|
| Two people | Two-option rule | Voting with two people just ties. Someone has to narrow it. |
| Three to five, deciding now | Nominate and vote, head-to-head to break ties | Small enough that everyone's nomination survives to the vote. |
| Six or more | Constraints first, then vote | Large groups have more hidden dealbreakers; surface them early. |
| Anyone with hard restrictions | Veto rounds | Removes a dealbreaker without making anyone explain it. |
| The same group every week | Rotating dictator | Fairest over time, and takes zero minutes each week. |
| Nobody has an opinion, at all | Random pick | When preferences are genuinely flat, speed beats accuracy. |
| Everyone has an opinion, loudly | Head-to-head elimination | Forces strong preferences to compete directly instead of shout. |
Four ways groups get stuck
The polite deadlock. Everyone defers to everyone else and nothing moves. This is the most common failure and it's not indecision — it's three people each waiting for someone to take the blame first. Any method that assigns the decision to a process rather than a person breaks it immediately.
The loudest voice. One person's preference wins because arguing costs more than eating somewhere you're lukewarm about. Nobody objects, so it looks like consensus, and it repeats until people quietly stop suggesting things. Simultaneous nomination and a silent vote both fix this, because neither lets anyone see what others chose before choosing.
Decision fatigue. By 7pm on a weekday, a group that's already made a hundred small choices will reject good options for no reason. This is why the pre-agreed method matters more than the options do. Deciding how you'll decide is one small choice; comparing eleven restaurants is a hundred.
The serial vetoer. One person says no to everything without proposing anything. Usually they have a constraint they haven't said out loud — price, distance, an ex who works there. Ask for the constraint, not the objection, or give them a formal veto so it has a limit of one.
A short version, if you're standing in a doorway right now
- Say the constraints out loud first: budget, distance, time.
- Everyone names one place at the same time. No commentary.
- Vote. If it ties, take the top two head-to-head, and if that ties, flip a coin — agreed to before the flip.
- Go. The decision is worth less than the twenty minutes you'd spend improving it.
Or let something else be the bad guy
MealDuel runs the nominate → vote → head-to-head sequence for you. You start a session in the iPhone app and share the link; everyone else joins in their browser — no install, no account — picks a restaurant, votes, and a duel settles any tie. The answer comes from the process, not from whoever spoke first.
Already got a join code? Join in your browser.