Nobody is hungry for your announcement, a Hybrid Storytellers feature

Nobody Is Hungry for Your Announcement

November 20, 2026•8 min read

Most restaurant social media is a bulletin board. Tonight's special. Holiday hours. A photo of the patio with "Come see us!" underneath.

Every one of those posts is written from the owner's side of the counter. It answers the question "what do we want people to know?" and nobody scrolling at 4:45 on a Thursday is asking that question. They are asking a different one: where should we eat?

A post that answers their question gets a table filled. A post that answers yours gets a polite like from your sister-in-law.

This is how we think about it at Hybrid Storytellers. It is the process behind every caption, reel and photo we publish for the restaurants we work with.

A post has to earn four yeses

A guest does not go from scrolling to sitting in your dining room in one step. They say yes four times, in order, and a post that skips one loses them there.

  • Stop. "Is this worth two seconds?" Earned by motion, steam, a cheese pull, a face, a first line that starts mid-thought.

  • Believe. "Is this place actually good at this?" Earned by one specific reason: the 14-hour brisket, the tortillas pressed to order, the owner's mother's recipe.

  • Worth it. "Is it worth leaving the house and spending the money?" Earned by the full picture of the night: who they would bring, what it costs, how easy it is.

  • Tonight. "Why now instead of someday?" Earned by a real reason with a real date: Thursday only, 40 portions, patio closes after this weekend.

Most restaurant posts try to earn only the fourth yes. "Join us tonight!" is a request to act from someone who has not been stopped, convinced or tempted yet.

Not every post needs all four. A reel of the grill at full flame might only earn the stop and the belief, and that is a good day's work. But every post should know which yeses it is going after.

Your guests already wrote your best captions

The owner says "elevated comfort food with locally sourced ingredients." The guest says "I think about that mac and cheese at least once a week."

Only one of those sentences makes anyone hungry.

So before we write a word for a client, we read. Google reviews, Instagram comments, DMs, tagged photos, the notes servers pass along. We are looking for three things:

  • The words they use. If twelve reviews call it "the crispy chicken thing," that is its name on social, whatever the menu says.

  • The moment they describe. "We came for my dad's birthday." "First place we go when we land." Those are the occasions the restaurant already owns.

  • The doubt they had. "I almost didn't come because parking looked bad." Every doubt a happy guest got past is one a future guest still has.

The job is to hand people their own reasons back, in their own language, with a better photo. We are not there to invent a clever angle.

One post, one guest

"Everyone within ten miles" is a delivery radius. It tells you nothing about who you are writing to.

Every restaurant serves several distinct groups, and each one decides differently:

  • The couple who needs a date night that feels like an occasion but not a production

  • The office manager ordering lunch for fifteen by 10 a.m.

  • The family that needs everyone fed in under an hour

  • The regular who wants to know what is new

  • The group planning a birthday, a shower or a rehearsal dinner

A post written for all five speaks to none of them. The same short rib is "the thing to order on your anniversary" to one and "feeds your whole team, delivered hot" to another.

We pick one guest per post. Across a week, each group gets its turn.

Five questions before any caption

Writing is the last and fastest step. Bad posts come from writing first and thinking second. Before anything gets drafted, by a person or by AI, we answer five questions.

  1. What are the facts? The dish, the price, the hours, the date, what is in the photo. Nothing goes out that the kitchen cannot back up.

  2. Who is this for? One guest from the list above.

  3. What do they already want? Their words, from reviews and comments.

  4. What is the one reason to believe? A single specific detail about how this kitchen does it. One, not five.

  5. What is the real reason to act now? If there is none, the post does not ask for action. It earns attention and stops there.

This matters more now that AI can write a passable caption in four seconds. A model with no brief guesses, and it guesses the same thing for every restaurant in America. "Indulge in our mouthwatering..." comes from an empty brief. The brief is where the work is.

The process on one post

Take a made-up example: a neighborhood Italian place we will call Rosa's. It makes pasta by hand every morning, and Tuesdays are slow.

The bulletin-board version:

Pasta night is back! Join us every Tuesday for fresh handmade pasta. Reservations recommended.

Nothing in it is false, and nothing in it would stop anyone. It could belong to any Italian restaurant in the country.

The brief:

  • Facts: Pasta is rolled and cut by hand each morning starting at 7. Tuesday pasta dishes are $16. The kitchen makes about 60 portions.

  • Guest: Couples nearby who want a weeknight out that does not feel like a big spend.

  • What they already say: Reviews keep mentioning watching the pasta being made through the kitchen window, and that it is "nothing like the boxed stuff."

  • Reason to believe: The 7 a.m. start, and the person doing it.

  • Reason to act: Tuesday only, and when the morning's batch is gone, it is gone.

The post: a reel that opens on hands cutting pappardelle, flour in the air, no music for the first second.

This was flour and eggs at 7 this morning. By 9 tonight it's gone.

Rosa makes about 60 portions every Tuesday and doesn't make more. $16 a plate. Bring someone you like.

The second version is the same restaurant, the same night and the same price. It earns all four yeses in 40 words because every line came from the brief.

Copy what is proven, bet on what is new

A restaurant's feed cannot be all experiments, and it cannot be all repeats. We sort every post in a week by how much risk it carries.

  • Safe (about half the week). A format and angle that has already worked for this restaurant.

  • Stretch (about a third). A proven format with one new element: new guest, new hook, new dish.

  • Bet (one or two posts). Something this restaurant has never tried: a new format, a trend, a staff-led series.

The safe posts keep the feed dependable. The bets are how we find next quarter's safe posts.

The rule for bets is to change one or two things at a time. A new format, with a new angle, for a new audience, on a new dish, teaches nothing whether it wins or loses, because we cannot tell which change did it.

Likes are not the scoreboard

A process with no feedback loop is a ritual. Every month we look at what each post did, and we judge it by the yes it was built to earn.

  • Stop: three-second views, and how much of the reel people watched.

  • Believe: saves, and comments that name the dish.

  • Worth it: shares and tags, like "we need to go here" sent to another person.

  • Tonight: reservation and order clicks, calls, direction taps, guests who say they saw it.

A share is the number we care about most. Nobody eats out alone on purpose, so a post sent to a spouse or a group chat is a table being planned in front of you.

Then we do three things. Posts that worked move into the safe pile. Posts that failed get one more try with a single change. And we ask the staff what guests mentioned, because the best data a restaurant has is still a server hearing "we saw the video."

What we will not do

A restaurant lives on repeat visits from people who live nearby. Any tactic that wins one visit and costs trust is a bad trade, so a few things are off the table:

  • Invented urgency. "Limited time" means there is an end date. "Almost sold out" means it is.

  • Photos the kitchen cannot match. If the plate in the post is not the plate on the table, the review writes itself.

  • Claims nobody can check. "Best in the city" earns nothing. "Smoked for 14 hours" earns belief.

  • Trends that do not fit. If the owner would wince watching it, guests will too.

The short version

A post is there to help someone who is already hungry decide where to go. An announcement of the restaurant does not do that.

Know which guest you are talking to. Use their words. Give one specific reason to believe and one true reason to come now. Write last. Then check what happened and do more of what worked.

None of it is complicated, but it takes discipline every week, and most restaurants do not have the time for that. That is the job we do at Hybrid Storytellers.

Book a free strategy call and we will walk through your last month of posts together.

Back to Blog