Most hosts build a recommendation list by opening a review site, sorting by rating, and copying the top ten. The result is a list that duplicates something the guest already has on their phone, and that answers a question they were not asking.
The question a guest is actually asking is narrower than "where is good". It is: where should we go, from here, tonight, with these people, given how tired we are. Build the list around that and it becomes useful.
Start with a radius, not a rating
Draw a circle with a twelve-minute walk as the radius. Everything inside it is a candidate. Everything outside it needs a reason.
Twelve minutes is not arbitrary. It is roughly the distance at which people stop deciding and start committing — beyond it, a guest is making a plan rather than wandering out. Most stays contain far more of the first kind of decision than the second.
A merely good restaurant four minutes away will be visited. An excellent one twenty-five minutes away, with a change of transport, will be admired and skipped. If your list is all excellence and no proximity, it will be read and not used.
This is also why walking time beats distance as a unit. "900 metres" makes a guest do arithmetic. "Eleven minutes, and the last bit is uphill" does not.
The five slots
Rather than ranking places, fill roles. Almost every stay needs these five, and a list that fills them is more useful than a list twice its length.
1. Coffee and something to eat, within five minutes. The first morning. Guests wake up in an unfamiliar kitchen with no milk. This is the single most-used entry in any guide I have written, and it is almost never the best coffee in the city — it is the nearest good one that opens early.
2. Dinner without planning. Somewhere that takes walk-ins, is open most nights, and will not ruin the evening. This is the workhorse. It does not need to be remarkable. It needs to be reliable, and you need to say so.
3. The one worth the trip. Book-ahead, special, the reason someone remembers the town. Include the booking lead time, because a recommendation the guest discovers too late to use is a small cruelty.
4. The rainy afternoon. Indoors, lasts two hours, works for the group sizes you host. Museums, baths, a cinema, a covered market, a long lunch. Every destination has bad weather and almost no guidebook plans for it.
5. The thing only a local knows. The tidal pool, the viewpoint behind the car park, the Thursday market, the bakery that sells yesterday's bread half-price at four. One or two of these is what makes a guide feel like a gift.
Fill those five and you have covered most of a three-night stay. Add a second entry per slot, and you are done.
Why the best restaurant in town is often the wrong answer
The highest-rated place in a city is generally the one that is hardest to get into, most expensive, and most likely to disappoint against expectation. Sending a guest there transfers risk to them: they book weeks ahead, spend a lot, and if it is merely good they blame the person who recommended it.
Recommend it, by all means — but as the "worth the trip" slot, with the constraint stated plainly. The reliability slots are where your credibility is actually earned.
Places you have not been
You do not need to have eaten everywhere to recommend it, but you do need to be honest about which is which. Two phrasings solve this:
I haven't eaten here, but three separate guests have come back and told me to add it.
Locals queue here on Sundays, which is usually all I need to know.
Both are more persuasive than a fabricated personal anecdote, and neither costs you anything if the guest is unimpressed.
The exclusions people forget
Check the practical constraints before adding an entry, because these are what generate complaints:
- Card or cash. Still a live issue in a surprising number of places.
- Accessibility. Steps at the entrance, a basement dining room, no accessible toilet.
- Children. Plenty of good places are effectively adults-only after seven.
- Dogs. If you accept pets, your list should mark which places do.
- Dietary options. If a place has one vegetarian dish and it is a side salad, say so.
- Closing day. The most common cause of a wasted walk. Mondays and Sundays, mostly.
Re-sort by guest, not by category
The last move is the most useful one. Take your finished list and ask: if a family with a seven-year-old arrived tonight in the rain, could they find the right three entries in under a minute?
If the answer is no, the list is organised for you rather than for them. Group by occasion, not by cuisine. "Breakfast", "easy dinner", "worth booking", "if it rains", "our favourites" beats "Italian, Thai, Cafes, Bars" every time, because nobody stands in a hallway thinking "I am in the mood for the Thai category".
A note on distance and photos
Two details make a list feel finished: a photo of each place, and how far it is on foot. Guests recognise a frontage far faster than they parse a name, and a walking time turns a list into a plan. Whether you assemble that by hand or with a tool that fetches it for you, do not ship a list without it.