Facebook & Instagram ads for restaurants: a no-waste guide

Most independent venues have the same Facebook ads story: they boosted a few posts, spent a couple of hundred dollars, got some likes from people three suburbs over, and quietly decided ads do not work for restaurants. Ads work. Boosting posts does not. This guide is the exact setup we use to turn a small daily budget into booked covers and direct orders, built for Australian venues in 2026.

Why most restaurant ads waste money

The problem is almost never the platform. It is what owners are asked to do by default. Instagram and Facebook nudge you toward the blue Boost button, which is designed to be easy, not effective. A boost optimises for the cheapest possible engagement, so Meta shows it to people who like and scroll but rarely act. You get a tidy little pile of likes and almost no bookings, then conclude ads do not work.

The second problem is the message. Most restaurant ads say some version of "we are open, come in." That is not a reason to change your Friday plans. People do not act on your existence; they act on a specific, timely reason. Fix those two things, the tool and the message, and the maths changes completely.

Everything below runs through Meta Ads Manager (free, at business.facebook.com), not the boost button. It is a little more setup once, and it is the difference between buying attention and buying customers. If you would rather not learn the tool at all, this is exactly what our restaurant Meta ads service handles end to end.

The only 3 campaigns worth running

Ignore the dozen objectives Meta offers. A single independent venue needs three jobs done, and they map to a simple funnel: get seen, get the warm ones to act, and recover the ones who almost did.

1. Reach the locals (top of funnel)

A low-cost campaign that puts your best video in front of people inside your catchment, optimised for reach or video views. The job here is cheap, repeated exposure so that your venue becomes familiar. Familiarity is what makes the next ad work. Keep this one always-on at a small daily budget.

2. Drive bookings or orders (the workhorse)

A conversion or traffic campaign that sends people to one specific action: your booking page, your reservation widget, or your direct ordering page. This is where you advertise an actual offer or event. This campaign carries the result, so most of your budget lives here. Send orders to a channel you own, not the delivery apps, for the reasons we cover in taking orders commission-free.

3. Retarget the warm crowd (the cheapest covers you will buy)

A retargeting campaign aimed only at people who already engaged: watched 50 percent of your video, visited your site, or tapped your profile. These people are most of the way to a decision, so a gentle nudge ("tables left this weekend") converts far cheaper than cold reach. This is the campaign almost no independent venue runs, and it is usually the most profitable one.

Advertise an offer, not your existence

The ad creative needs a reason that is specific and time-bound. Strong angles for a venue:

• An event or theme: a pasta night, a winemaker dinner, live music Friday, a new seasonal menu drop.
• A signature dish in motion: the cheese pull, the pour, the flame, the slice. Let the food do the selling.
• An added-value offer: a free glass with two courses midweek, not a blanket percentage off. The reasoning is in filling slow nights without wrecking your margin.
• A booking trigger: "Mother's Day bookings now open," "only Saturday sittings left." Scarcity and occasion both work.

Every ad needs one clear next step: Book a table, Order direct, See the menu. One ad, one offer, one action. If you are tempted to say three things, that is three ads.

Targeting a local venue

This is where owners overthink it. For a single location, the targeting is simple and the instinct to get clever usually backfires.

• Location: set a radius around your venue, usually 5 to 8km in a city, wider in a regional town. Drop a pin on your exact address rather than relying on the suburb. Make sure it is set to "people living in or recently in this location," not "travelling."
• Age and gender: only narrow these if you genuinely skew (for example a late-night bar). Otherwise leave them broad.
• Interests: here is the counter-intuitive part. On modern Meta, stacking interests like "fine dining" and "foodie" usually makes performance worse, because it shrinks the audience and fights the algorithm. For a tight local radius, a broad or Advantage+ audience lets Meta find the diners using its own signals. Test broad first.
• Exclude: add your staff and existing followers to a saved exclusion for cold campaigns so you are not paying to reach people who already know you.

The single most important targeting lever for a restaurant is the radius. Get that right and let the machine do the rest.

Creative that stops the scroll

The creative matters more than every targeting setting combined. The feed is a thumb-stopping contest, and food is one of the few things that reliably wins it. Rules that hold up:

• Vertical video first. Shoot 9:16 for Reels and Stories. A 5 to 15 second clip of food in motion beats a polished static image almost every time.
• Win the first 1.5 seconds. Open on the most appetising frame, no slow intros or logos. If they do not stop in the first beat, nothing else matters.
• Caption everything. Most feeds play on mute. Burn captions into the video so the message lands without sound.
• Real over polished. Phone footage from your actual kitchen and dining room outperforms stock and over-produced agency video. It looks native, so people trust it. User-generated clips from happy diners are gold; with permission, run them as ads.
• Refresh before fatigue. When a winning ad's results start sliding after a few weeks, that is creative fatigue. Have the next clip ready. This is why a steady content habit and ads feed each other, which we cover in Instagram and Reels for restaurants.

Budget & the cost-per-cover maths

You do not need a big budget. You need a profitable one. Here is how to think about it with illustrative numbers; plug in your own.

Say you run the bookings campaign at 15 dollars a day, about 450 a month. Over a month it delivers 60 booked covers you can attribute to the ad. That is a cost per booked cover of 7.50 dollars. If your average spend per head is 55 dollars at a 65 percent gross margin, each cover returns roughly 35.75 dollars in gross profit. You spent 7.50 to make 35.75. That campaign should be scaled, not questioned.

To set your ceiling, work backwards: decide the most you would pay to acquire one cover (or one new regular, who is worth far more than one visit), and keep any campaign that comes in under it. Scale winners by lifting the budget around 20 percent every few days so you do not reset Meta's learning. Cut anything that cannot get under your ceiling after a fair test.

Tracking what actually worked

An ad you cannot measure is a donation. Three things to set up once:

• Meta Pixel + Conversions API. Install the Pixel on your site (most booking and ordering platforms have a one-click field for the ID). Pair it with the Conversions API so you still capture results when browsers block cookies. This is what lets Meta optimise toward bookings instead of clicks.
• Tagged links (UTMs). Put tracking tags on every ad's destination URL so your booking system and Google Analytics show which covers came from which ad. Without this you are guessing.
• Offline reality check. The Pixel will undercount, because plenty of people see the ad and then phone or walk in. Ask new bookings "how did you hear about us," and watch your overall covers on the weeks you advertise versus the weeks you do not. The trend tells the truth even when attribution is fuzzy.

7 mistakes that drain the budget

• Boosting posts instead of running Ads Manager campaigns. The number one money-burner.
• No offer or reason in the ad. "We are open" is not a campaign.
• Stacking interests on a tiny local audience, which starves the algorithm.
• Static images only. In a food category, video almost always wins.
• Judging on likes and reach instead of covers and orders.
• Set and forget. Not refreshing creative, so winning ads quietly fatigue and costs climb.
• No retargeting. Leaving the cheapest covers on the table by never following up warm viewers.

Avoid those seven and you are already ahead of almost every independent venue running ads. For where paid ads sit alongside local search, social and your owned list, see the complete guide to restaurant marketing in Australia.

FAQ

How much should a restaurant spend on Facebook ads?

Start small and prove it works before you scale. For most independent venues, 10 to 20 dollars a day (roughly 300 to 600 a month) is enough to learn which offers and creative pull bookings. Once you can see a campaign returning more in covers than it costs, increase the budget on that winner by about 20 percent at a time. Do not start at 50 dollars a day on an untested campaign.

Do Facebook and Instagram ads actually work for restaurants?

Yes, when they advertise a real reason to come and send people to a booking or order page you control. They fail when they boost a generic post to a broad audience and get judged on likes. The venues that win treat ads as a way to put a specific offer in front of nearby diners, then measure the result in covers and orders, not engagement.

Should I boost posts or use Ads Manager?

Use Ads Manager. The boost button is a stripped-back version that hides the targeting, placements, optimisation goal and conversion tracking that make ads profitable. Boosting optimises for cheap engagement, which is why it delivers likes instead of customers. Everything in this guide runs through Ads Manager.

What is a good cost per result for a restaurant ad?

It depends on your average spend, but the only number that matters is cost per booked cover or per order versus what that customer is worth. If a campaign costs 6 dollars per booking and your average table is worth 90 dollars, that is a strong result. Work backwards from your margin to set a ceiling, then keep the campaigns that beat it and cut the ones that do not.