Before they messaged you, your guest already opened three apps. They know the taco place has 4.6 stars and 2,200 reviews. They know the coffee shop two blocks over is "trendy" according to a stranger who visited once in 2019. They have more restaurant data in their pocket than any guidebook could ever compete with.
So when a host writes "Try the amazing pizza place on Main Street, it's a local favorite!" they've told the guest nothing. The phone already knew there was a pizza place. The phone already has opinions about it. What the guest actually wants is the thing the phone can't give them: a person who lives nearby telling them which three options matter and why.
The rating tells you nothing about fit
A 4.7-star brunch spot could be loud, cash-only, and have a forty-minute wait on weekends. A 4.1-star diner three blocks away could be quiet, fast, and exactly what a family with a cranky toddler needs at 8am. Stars average out everyone's experience into a single number. They don't tell you which place is right for the person standing in your unit right now.
This is the gap you can fill, because you know things no aggregate rating captures:
- Which restaurant takes walk-ins after 2pm even when it says it's booked
- Which coffee shop has two tables in the back that nobody uses before 10am
- That the "best" ramen place everyone reviews well has a 45-minute wait on Fridays, but the one next door is just as good and empty
- That the grocery store three minutes away is overpriced, but the one eight minutes away has everything for half the cost
None of that shows up in a star rating. All of it is useful the moment someone is standing in your kitchen wondering where to eat.
Write it like you're answering a text from a friend
Compare these two:
"Great Italian restaurant nearby, highly recommended!"
"Luigi's on 4th, two blocks past the gas station. Get the carbonara, skip the tiramisu (it's frozen, not made in-house). Go before 6:30 or after 8:30 or you'll wait outside. Cash tip only, they add a card fee otherwise."
The second one isn't fancier writing. It's just specific. It tells the guest exactly where to go, what to order, when to show up, and one thing to know before they get there. A guest reading that message feels like they got an actual tip, not a pleasantry.
If you can't write that level of detail about a place, that's useful information too. It usually means you haven't been there in a while, or you're recommending it because it's famous, not because you know it. Cut it or go check it out yourself.
The "don't bother" list is worth as much as the "go here" list
Guests rarely get this from anyone. Every list they find online is a "best of" list, because nobody gets paid to write "skip this." But you know the overpriced tourist trap two blocks from your listing that gets slammed with five-star reviews from people who had nothing to compare it to. You know the "scenic overlook" that's actually a parking lot with a view blocked by trees half the year.
A single sentence like this saves someone an afternoon: "Skip the walking tour that leaves from the fountain, it's overpriced and rushed. The free one from the library at 10am covers the same ground better."
You don't need a long list. Two or three honest warnings do more for a guest's trust in you than twenty generic recommendations.
Match the recommendation to who's actually staying
A recommendation that works for a couple on a weekend trip is often wrong for a family of five, and a recommendation for a business traveler alone for one night is different again. If you know who's booked, you can tailor what you send instead of handing everyone the same list.
For a family: "The playground at the end of the street has shade and a working water fountain. The one by the main square looks nicer in photos but there's no bathroom nearby."
For a couple without kids: "The wine bar on the corner has a patio that's quiet after 9pm. Ask for the back table, not the ones facing the street."
For someone traveling for work: "There's a 24-hour pharmacy four minutes north if you need anything late, and the diner next to it does a decent breakfast before 7am if you have an early flight."
You're not writing three separate guidebooks. You're just noticing who's in front of you and picking the three recommendations out of your list that actually apply.
Keep the list short enough that you've actually checked it
A guidebook with forty restaurant recommendations is a sign the host copied a city guide, not that they live there. Nobody has forty go-to spots within walking distance. Five or six solid ones, each with a real reason attached, beats forty generic entries that could apply to any city in the country.
If a recommendation has been sitting in your welcome guide for two years and you haven't eaten there recently, either go check it's still good or take it out. Restaurants close, change owners, get worse. A stale recommendation is worse than no recommendation, because it spends your credibility on something you didn't actually verify.
The pattern worth remembering
Anything a search engine can already tell a guest isn't worth repeating. The name of a place, its star rating, its hours — they have all of that before they land. What they don't have is judgment: which of the five well-reviewed restaurants is actually good, which one to avoid despite the reviews, and which one fits the specific people in your unit this week. That judgment is the only thing you're offering that an app can't. It's worth more than any star rating you could point them to, and it's the one thing nobody can get except from someone who's actually been there.