You type "where should I eat near me" into your phone, and the phone answers immediately. Fifty places, all of them open, most of them fine.
Twenty minutes later you're still there, three reviews deep on a Thai place you've already decided against, hungrier than when you started.
The search was never the hard part. Your phone solved that instantly. The hard part is that it handed you fifty options and no way to cut them down, which is a worse position than having three.
Why the answer takes so long to arrive
Fifty is past the useful limit. Barry Schwartz's paradox of choice covers this: beyond a certain number, extra options make people slower to choose and less happy with what they picked. Three restaurants is a decision. Fifty is a research project.
You're doing it hungry. Which is when you're worst at it. Blood sugar drops, patience goes, and the longer it takes the worse both get.
With other people it compounds. Four sets of preferences, at least one dietary constraint, and everyone performing flexibility so hard that nobody will name a place.
Six ways to actually land on somewhere
Roughly fastest first.
1. Let a wheel decide
The quickest route from hungry to eating. Give a restaurant wheel your suburb, let it pull the well-rated places nearby, spin it once, and go.
It's fastest because it skips the part you're bad at. You never evaluate fifty options — you accept one, which takes no effort at all.
The low-tech version works too: search on Maps, zoom until a dozen pins are showing, close your eyes and tap. Same principle, worse information.
Either way, the result only counts if you honour it.
2. Sort by rating and take from the top
More control, still quick.
Open Maps, search restaurants, filter to highest rated, pick from the first three or four. Don't scroll past that. Anything above 4.5 stars with a few hundred reviews is going to feed you well, and the difference between the 4.6 and the 4.7 is noise you cannot taste.
3. Ask people who live there
Local knowledge beats ranking algorithms, particularly for places that are excellent and bad at marketing.
Suburb Facebook groups, the city subreddit, a neighbourhood chat, the Instagram location tag. All good sources.
The catch is latency. This is a great way to build a list for next month and a useless way to solve dinner tonight while standing on a corner.
4. Browse a delivery app without ordering
Even when you're eating out, delivery apps are a decent restaurant directory. They show you what's nearby, what it costs, and — crucially — the actual menu.
One caveat: ratings there measure the whole delivery experience. A three-star review is often about a cold bag and a late driver, not the kitchen. Read a couple before you write somewhere off.
5. Walk in one direction
Old-fashioned and still good in any dense area. Pick a direction, start walking, stop at the first place that appeals.
You get information no app has — how full it is, whether it smells right, whether the people inside look pleased to be there. That's a genuinely better signal than a star rating.
Useless in the suburbs, or if it's raining, or if you have twenty minutes.
6. Build a shortlist once and reuse it forever
The long game, and the one that eventually makes the rest unnecessary.
Pick eight or ten places near home or work that you know are good. Save them in a note or a Maps list. When you're hungry, choose from those and nowhere else.
You're now picking from ten pre-approved options instead of fifty unknown ones, and every choice after this one is faster. Revisit it every few months — drop the disappointments, add the discoveries.
Define "near me" before you search
Most bad results come from the phrase itself. "Near" isn't a distance, it's a time budget, and it changes hourly.
- Walking, short break — five to ten minutes, 500m to 1km
- Driving — ten to fifteen minutes, around 5km
- Worth the trip — 20 to 30 minutes, for when it's an occasion
Set the radius to the situation. A 30-minute break and a 5km search will reliably send you somewhere you can't get to.
Two things sharpen the results: turn location services on, so you're being matched to where you actually are rather than the middle of your city, and search by suburb name rather than "near me". Naming the suburb gives you tighter results and tends to surface local write-ups that a generic search never will. That's how our restaurant picker works — you name the area, it looks there.
Groups
Adding people doesn't divide the difficulty. It multiplies it.
Never ask the open question. "Where should we eat?" produces "I don't know, where do you want to eat?" indefinitely. Offer two options instead. People answer a choice; they stall on a blank.
Or take one suggestion each and spin. Everyone names a place, all of them go on the wheel, one spin decides. Everyone's represented, nobody chose, and there's no one to be annoyed at if it's average. Agree to honour the result before spinning — this falls apart the moment "best of three" enters the conversation.
Or rotate. If you eat together regularly, take turns choosing. Monday is yours, Tuesday is theirs. Fair, and it retires the debate permanently.
Three ways people make this harder
Reading too many reviews. Read fifty and you'll find a reason to avoid everywhere, because everywhere has a bad night on record. Read the five most recent and move.
Optimising. You're choosing lunch, not a school. Good ratings, food you like, decided in two minutes.
Confusing this with a different problem. If you genuinely don't know what you feel like, no amount of searching will help — that's an appetite problem, not a directory problem, and it has its own set of fixes.
The short version
Your phone can find fifty restaurants near you in half a second. That has never been the constraint.
The constraint is that you'll take twenty minutes to pick one, and that the difference between the best and worst of those fifty is one adequate meal either way. Pick a method above, apply it, eat.
If the pattern feels familiar beyond food, we wrote about that too — which decisions actually repay the effort, and which ones are just quietly taxing you.
Or skip all of it and let the wheel pick.