Meta Ads for Restaurants: What Actually Gets People Through the Door?
Creative, targeting, offers, retargeting, and why boosting Instagram posts isn't a real advertising strategy.
Every restaurant owner has done it: a plate of food comes out looking incredible, the phone comes out, the photo goes up, and then that little blue "Boost Post" button starts winking at you. It feels like advertising. It costs money, it has a targeting screen, and it gets you a few more likes than usual. But if you've been boosting posts for months and your dining room isn't any fuller, that's not bad luck. It's the tool. Boosting was never built to fill seats. It was built to make engagement numbers go up.
Real Meta advertising, built in Ads Manager and not from the app, is a different animal, and it's the difference between "we tried Instagram ads once" and a channel that reliably brings in new guests every week. Here's what actually works, broken into the four things you have to get right: creative, targeting, offers, and retargeting.
Why boosting a post isn't a strategy
Boosting feels like advertising because Meta wants it to. It's the path of least resistance, and it's designed to spend your money quickly and painlessly. But it's missing almost everything that makes an ad campaign actually perform.
A boosted post can only optimize for a handful of shallow goals (more profile visits, more comments, more reach) because those are the only objectives the boost button offers. It has no access to real conversion campaigns like "get people to view the menu and tap directions" or "get people to claim an offer." It's also stuck in a narrow set of placements, usually just the main feed, while a proper campaign can run across Stories, Reels, and Explore, which is where a lot of your actual audience is scrolling. And because boosting skips Meta's full auction and optimization system, it tends to cost more per result than a well-built campaign, not less: you're paying for reach, not for outcomes.
The biggest gap is targeting. Boosting gives you broad, generic audience sliders. Real campaigns let you build custom audiences from people who've visited your website, viewed your menu, watched a chunk of your Reels, or visited your Instagram profile, and then layer lookalike audiences and precise local geo-targeting on top of that. That's the entire game, and boosting doesn't play it.
None of this means the content you'd boost is wasted. It just belongs inside a campaign, not standing alone. Take the same photo or video and run it as an actual ad, and you're suddenly using Meta's system the way it's meant to be used.
Creative: the plate has two seconds to earn attention
Creative is the single biggest lever in restaurant advertising, more than targeting or budget. People scroll fast, and food content lives or dies in the first couple of seconds.
The clearest pattern across restaurants that advertise well is that raw, phone-shot content consistently outperforms polished, agency-produced video. A slightly imperfect video of a burger being plated, sauce being poured, or a cocktail being shaken reads as real, and real gets watched longer than something that looks like a commercial. Cut to the food fast, ideally within the first two seconds, and keep the whole thing to 10-20 seconds. Anything longer and you're relying on a level of attention nobody is giving a food ad on Instagram.
A few formats that consistently earn their keep:
Short video of food being made or plated, shot handheld, on the actual line, not staged.
Carousels that walk through a handful of dishes or a full experience (starter, entrée, dessert, drink), so people can swipe at their own pace.
A single strong photo paired with a real customer quote or review, which does double duty as social proof and as creative.
Skip the stock-photo aesthetic and skip over-produced brand videos with music swells and slow motion; they read as ads, and people are extremely good at tuning ads out. The goal is content that looks like it belongs in someone's feed, not next to it.
Targeting: tight geography beats broad demographics
For a physical restaurant, the single most important targeting lever isn't age or income. It's distance. Nobody drives forty-five minutes for a Tuesday dinner because an ad looked good. Set your radius based on how dense your area is: something like 3-4 kilometers in a dense urban core, and 5-8 kilometers in the suburbs, using "people living in or recently in this location" rather than just "people who live here," since that also catches people currently nearby: visitors, commuters, people at work.
From there, layer on light demographic targeting (a broad adult range covers most restaurants; resist the urge to over-narrow by age) and a handful of genuine dining-related interests. But don't stop at cold audiences. The real value shows up when you build custom audiences out of people who already know you: recent website visitors, people who've engaged with your Instagram or Facebook profile, and people who've watched a meaningful chunk of your videos. These warm groups convert dramatically better than strangers, because they've already raised their hand.
Offers: specificity beats "come try us"
A generic offer like "10% off your visit" or a caption like "our signature dish!" doesn't give anyone a reason to act today instead of eventually. Specific, concrete offers do. Name the actual dish: "slow-braised short rib, back on the menu this week" does more work than "check out our specials." Attach a real deadline when you can ("available through the end of the month") rather than a vague "for a limited time," because urgency only works when it's believable.
The best-performing offers for restaurants tend to solve a specific occasion rather than trying to appeal to everyone: a weeknight family deal, a happy hour built around a slow afternoon shift, a new menu item worth trying once, a private event or catering push during a season when that's relevant. An offer aimed at a moment is easier to make a decision about than an offer aimed at "anyone, anytime."
Whatever the offer, make claiming it effortless. Ads that use Meta's built-in lead forms (the ones that open inside the app with a pre-filled name and email) convert meaningfully better than ads that send people to an outside reservation or ordering page, simply because every extra tap or page load loses people.
Retargeting: the return visit is where the money is
Most people who see your ad, or even visit your website, won't do anything the first time. That's normal; most of the internet works that way. Retargeting is how you catch the people who were close, not the people who were never interested. Someone who watched most of a Reel, browsed your menu page, or visited your profile has already shown real intent; a second, more direct nudge toward booking or ordering tends to land far better than another awareness ad ever could.
Structure retargeting by how recently someone engaged. People from the last week or two are your warmest group and deserve your most direct offer. People from the last month are still worth a nudge. People from a few months back are better served by a broader "here's what's new" reminder than a hard sell. And once someone's already booked or ordered, take them out of the retargeting pool. Showing an offer to someone who just paid full price is how you train people to wait for a deal next time.
As a rule of thumb, this remarketing layer usually only needs a modest slice of the overall budget, something like 15-25%, with the bulk still going toward reaching new people, but it tends to be the most efficient money you spend, because you're talking to people who already know who you are.
Putting it together
None of these four pieces works especially well in isolation. Great creative aimed at the wrong radius wastes reach. Precise targeting behind a generic "10% off" offer gets ignored. A great offer with no retargeting layer only ever gets one shot at each person. The restaurants that get real, repeatable results from Meta treat it the way it's built to be used: real campaigns, built in Ads Manager, with a bit of budget split across reaching new people nearby, drawing them in with a genuine offer, and following up with the ones who were already close.
Boosting a good post isn't wrong, exactly. It's just not enough on its own to be called a strategy. The good news is that the leap from "boosting occasionally" to "running real campaigns" isn't complicated. It just takes building the campaign in the right place.
