TAKE Independent restaurants, groups and franchises

Restaurant Advertising: AI Video Ads for Restaurants and Social Media Marketing for Restaurants

A restaurant ad budget quietly funds two different businesses. One sells a visit, at full margin, inside a three mile radius. The other sells an order, and what that order is worth depends entirely on whether it came through a marketplace or through your own page. Most restaurant marketing advice never separates them, so the money goes to whichever one is easier to buy. This page separates them, prices both, and covers the review rule that took effect while nobody was reading. Or start now: paste your menu page URL on the right.

FTC review rule quoted verbatim No food videographer needed Where an agency wins, stated

Ad Recipe

TAKE

Pick a Creator

Hook Style

Free to start - no credit card required

The short answer Last updated August 2026

Restaurant advertising splits into visit advertising and order advertising, and the two are worth different amounts of money per unit sold. A visit is a dine-in cover at full contribution margin, bought inside a radius most people will actually drive, and it is sensitive to daypart, weather and day of week. An order is delivery or takeout, and its value depends on the channel: DoorDash publishes US partnership plans at 15, 25 and 30 percent commission on delivery with pickup at 6 percent, so a marketplace order can be worth substantially less to you than the identical order placed on your own site. That single fact should set your budgets, because you can rationally pay several times more to acquire a direct order than a marketplace one. Sitting on top of the math is a compliance layer that arrived recently and that almost no restaurant marketing guide mentions. The FTC rule at 16 CFR Part 465 makes it unlawful to give an incentive conditioned on a review expressing a particular sentiment, to publish insider reviews from staff or managers without disclosing the relationship, and to display a review wall on your own site that implies it shows most or all reviews while negatives are filtered out. Separately, 16 CFR 255.5 treats a comped meal as a material connection that a food influencer has to disclose. None of that is enforced by the ad platforms, so it lands on the operator.

Businesses in one ad budget

15 to 30%

Published marketplace commission

1080x1920

Reels, TikTok, Shorts

$49

UGCGen, per month flat

Read this first

Selling a visit and selling an order are two different businesses

Nearly every guide to restaurant advertising is organized by channel. A section on Instagram, a section on Google, a section on flyers and register tape. That structure hides the decision that actually moves money, which is what you are trying to sell in the first place. A cover and an order behave differently in almost every dimension that matters: the radius people will travel, the hours demand appears, the margin that survives, and how quickly the creative wears out. Holding one to the other's yardstick is how restaurants conclude that advertising does not work for them.

The clearest symptom is a restaurant that boosts posts about the restaurant generally, sees delivery orders tick up, and never notices that it paid to send business to a marketplace that takes a quarter of it. The second clearest is a restaurant with a Tuesday lunch problem running ads that say nothing about Tuesday or lunch.

	Visit advertising	Order advertising
What it buys	A cover in a seat, tonight or this weekend	A delivery or takeout order, often from home
Realistic radius	Roughly what someone will drive for dinner, usually a few miles	Whatever the delivery zone allows, frequently wider
Margin that survives	Full contribution margin, plus drinks and dessert attach	Depends entirely on channel, and the gap is large
What drives demand	Daypart, day of week, occasion, weather, who else is going	Convenience, bad weather, evenings in, craving specificity
Right creative	Room, atmosphere, people, the occasion the meal belongs to	The dish, close and specific, plus the ordering path
Right metric	Cost per incremental cover, checked against margin per cover	Cost per order, net of commission, by channel
Attribution	Weak, most people never click anything before walking in	Traceable when the order lands on your own domain
Creative burn rate	Fast, because a small local audience sees it repeatedly	Fast, and every menu item wants its own cut
Cheapest lever first	Advertise the soft daypart, not the restaurant in general	Move existing orders from marketplace to direct
How it scales	Caps at seats and at the population inside the radius	Caps at kitchen throughput during the delivery peak

Run the two as separate line items with separate targets. A single blended cost per result across both will always be dominated by whichever one is cheaper to buy, which is rarely the one worth more.

The number that sets your budget

A marketplace order and a direct order are not worth the same money

DoorDash publishes three US partnership plans and states the commission for each: Basic at 15 percent, Plus at 25 percent and Premier at 30 percent on delivery orders, with pickup at 6 percent on all three, and no monthly or signup fee on those published plans. Other marketplaces publish rates in broadly the same territory. Whatever tier you are on, the arithmetic is the same and it is unforgiving.

Take a $40 order. On a 25 percent plan the marketplace keeps $10 before you have bought a single ingredient. The identical $40 order placed on your own ordering page costs you card processing, which typically runs in low single digit percentages, so a couple of dollars. The difference between those two versions of the same order is most of the margin on the meal.

The advertising consequence is the part restaurants miss. If a direct order is worth roughly $8 more to you than a marketplace order, then you can spend up to $8 more to acquire one and still be even. Almost nobody sets budgets that way. They advertise the restaurant, the ad drives an order through whichever channel the customer already had installed on their phone, and the restaurant pays twice: once for the ad, once for the commission.

None of this means marketplaces are a mistake. They are a genuinely effective discovery channel, they reach people who would never have found you, and the pickup rate is materially cheaper than the delivery rate. The mistake is paying to send customers you already reached into a channel that charges you for reaching them.

	Marketplace order	Direct order on your own site
Published take on delivery	15, 25 or 30 percent depending on DoorDash plan tier	Card processing only, typically low single digits
Published take on pickup	6 percent across the published plans	Card processing only
Who owns the customer data	The marketplace, so you cannot easily remarket	You, including email and order history
Discovery value	Real, this is the honest case for being listed	None, people have to already know you exist
What ads should do	Let it earn discovery on its own, do not fund it	This is where paid traffic belongs
Sensible acquisition budget	Low, the margin is already spent	Higher, up to the commission you are saving

DoorDash rates as published on its US merchant partnership plans, checked August 2026. Rates change and vary by market and agreement, so confirm your own tier before you build a budget on it. Processing costs are typical ranges, not a quote.

The part most guides skip

Incentivizing a review is legal. Incentivizing a five star review is not.

Reviews are the closest thing a restaurant has to user generated content, and the rules covering them tightened in a way that most restaurant marketing advice has not caught up with. The FTC's rule on consumer reviews and testimonials sits at 16 CFR Part 465, and four of its provisions land directly on things restaurants do every week without thinking about them.

Start with the one that catches the most people. Section 465.4 makes it an unfair or deceptive act or practice "for a business to provide compensation or other incentives in exchange for, or conditioned expressly or by implication on, the writing or creation of consumer reviews expressing a particular sentiment, whether positive or negative, regarding the product, service, or business that is the subject of the review."

Read that carefully, because the line is not where people assume. The rule does not ban giving a customer something for leaving a review. It bans tying the incentive to the sentiment. A card that says leave us an honest review and take a free coffee next visit is a different thing from a card that says leave us a five star review and take a free coffee. And the phrase "or by implication" reaches the version everyone thinks is clever: a sign showing five gold stars next to the offer says the quiet part loudly enough.

Next, section 465.5 covers insider reviews. It requires a clear and conspicuous disclosure of an officer's or manager's material relationship to the business when they write a review, unless that relationship is already obvious, and extends the same requirement to testimonials from employees and agents that the business disseminates. It also reaches a manager soliciting reviews from immediate relatives or from staff where the resulting review carries no disclosure. Family and friends opening week reviews are the single most common form of this, and they are usually written with no bad intent at all.

Then section 465.7(b), which is the one that surprises operators most. It reaches a business that materially misrepresents, expressly or by implication, that the consumer reviews displayed on a portion of its own website "represent most or all the reviews submitted to the website or platform when reviews are being suppressed based upon their ratings or their negative sentiment." In plain terms, a testimonials page on your own site that quietly shows only the good ones, while presenting itself as what customers say, is the exposure. The rule does carve out removals for reviews that are defamatory, contain personal information, are discriminatory, are clearly false or suspected fake, or are unrelated to what you sell. Section 465.7(a) separately reaches using unfounded legal threats or intimidation to get a review taken down, which is worth knowing before anyone sends an angry letter about a one star.

Finally, the influencer question. When a restaurant comps a meal for a local food creator, that free meal is a material connection under 16 CFR 255.5, and the disclosure duty attaches whether or not any cash changed hands. In video, a disclosure buried in a caption that a scrolling viewer never expands is not doing the work. It needs to be in the video, readable, and present while the endorsement is being made.

What the restaurant does	Where it lands	The safer version
Free dessert for a five star review	16 CFR 465.4, incentive conditioned on sentiment	Offer it for an honest review of any rating, and say so
Staff and family review the new location	16 CFR 465.5, insider reviews without disclosure	Require the relationship to be stated in the review itself
Testimonials page shows only the good ones	16 CFR 465.7(b), suppression by sentiment	Label it as selected highlights, do not imply it is all of them
Legal threat over a one star review	16 CFR 465.7(a), unfounded threats to force removal	Reply publicly and factually, or use the platform's process
Comped meal for a local food creator	16 CFR 255.5, undisclosed material connection	Disclosure on screen in the video, not only in the caption
Ad shows a dish you do not actually serve	Section 5 of the FTC Act, deceptive advertising	Use your own photography of the real dish and portion

This describes federal advertising rules in general terms and is not legal advice. State consumer protection statutes add their own requirements, and several are stricter than the federal floor. Have counsel review anything that runs at scale.

Draw the line here

Generate the creator. Do not generate the food.

This is the rule we hold ourselves to, and it is worth stating plainly because the alternative is tempting and it will eventually cost somebody a lot of money. A generated presenter talking about your restaurant is an ordinary piece of advertising creative, the same category as hiring an actor. A generated plate of food that does not correspond to anything your kitchen sends out is a misrepresentation of the product, and Section 5 of the FTC Act does not care how it was produced.

Food advertising has always lived close to this line. The classic food styling debates were about whether a burger in a photograph was assembled more carefully than one on a tray, and the answer landed on presentation being allowed while the product itself has to be real. Generated imagery moves the question from how carefully you stacked it to whether it exists at all, which is a much shorter distance to deception.

So the construction we recommend, and the one the tool is built for, is this: your dish photography stays yours. You paste a menu page URL or upload your own shots of the real plate, and the generator handles the presenter, the voiceover, the captions, the pacing and the aspect ratios. You get the volume of creative that paid social demands without inventing a menu item that never left the pass.

The same logic applies to portion and price. If the ad shows a plate that is visibly larger than what arrives, or leads with a price that only applies at a daypart the ad never mentions, the mechanism of the deception is the same. Match the ad to the ticket.

How it works

From a menu page to a week of ad creative

The constraint on restaurant paid social is almost never targeting. A three mile radius is a small audience, so the same people see the same ad again and again, frequency climbs within days, and performance decays long before the budget runs out. The fix is more creative, more often, which is exactly the thing a restaurant has no time or staff to produce.

Paste the menu page

Drop in a menu or dish page URL, or upload your own photographs of the real plate. Your food stays your food.

Pick a creator and an angle

Choose a presenter and the hook: the soft daypart, the direct ordering offer, the dish nobody knows you make.

Generate variants

Get multiple cuts with voiceover and burned-in captions, so you can rotate before frequency kills performance.

Export and run

Download 9:16, 1:1 and 16:9 at 1080p with full commercial rights, then point every ad at your own ordering page.

Who this is for

Restaurant advertising problems this actually solves

A soft daypart

Tuesday lunch is empty and Saturday dinner has a wait. Those need different ads, and one of them needs an offer. Ship a cut per daypart instead of one ad for the restaurant.

Too much marketplace, too little direct

Delivery volume looks healthy and margin does not. Run creative that advertises the direct ordering page specifically, and give people a reason to use it.

A new location opening

No reviews, no photos, no repeat traffic, and a hard date. You need volume of creative immediately, before there is any customer content to work with.

A menu nobody knows the depth of

People order the same three dishes because those are the three they have seen. Every item that deserves demand needs its own video, which is a production problem, not a marketing one.

Multi-unit groups and franchises

Every location wants local creative and none of them has a videographer. Produce per location and per market without scheduling a shoot in each city.

Agencies with restaurant clients

Restaurant retainers are small and creative demand is not. The economics only work when asset production stops being billed by the shoot day.

Where we are the wrong answer

When a food photographer or a local agency beats doing this yourself

Generated creative solves one bottleneck: producing enough video to keep a paid social account fed in a small local radius. It does not shoot your food. If your dish photography is bad, this will not fix it, and bad food photography is the most common reason restaurant ads underperform. A half day with a good food photographer, producing a library of real plates you own outright, is a better first purchase than any software including ours. Buy that first, then use it here.

A local agency also earns its retainer on things this does not touch. Media buying, your Google Business Profile, the reservation and ordering stack, the relationships with local press and creators, and knowing which neighborhoods actually convert in your city. If that seat is empty, filling it will move more revenue than asset volume will.

Real customer content still wins on trust. A regular filming themselves eating at your restaurant, unprompted, outperforms anything synthetic, and it always will. The argument here is not that customer video is weak. It is that you cannot schedule it, you cannot brief it, and you cannot get twelve pieces of it by Thursday because frequency is climbing. Collect it when it happens, with permission, and spend it where trust is being decided.

And if the food or the service is the problem, advertising will make things worse faster. Paid reach on a restaurant with a genuine execution issue buys you more first visits that do not become second visits, and more reviews you will then be tempted to handle in one of the ways section 465.7 covers. Fix the room first.

Restaurant advertising questions, answered

How much should a restaurant spend on advertising?

Published small business guidance commonly lands between 3 and 8 percent of revenue, and restaurant specific advice often cites the higher end for a new location and the lower end for an established neighborhood spot. Treat those as estimates. The figure that decides whether the spend was smart is cost per incremental cover, judged against your contribution margin per cover rather than against the check average.

What is the best way to advertise a restaurant?

It depends on whether you are short of visits or short of orders, because those are different problems. If tables sit empty at known times, buy a tight local radius on paid social and Google, and advertise the daypart rather than the restaurant. If delivery volume is the gap, the highest value move is usually pulling orders off a marketplace and onto your own ordering page, where you keep the commission.

How much does DoorDash charge restaurants?

DoorDash publishes three US partnership plans that charge commission per order: Basic at 15 percent, Plus at 25 percent and Premier at 30 percent on delivery, with pickup at 6 percent across all three. There are no monthly or signup fees on those published plans. The practical consequence is that a marketplace order and a direct order are not worth the same amount, so they should not carry the same ad budget.

Can restaurants offer discounts for reviews?

Offering something for a review is not automatically unlawful. Conditioning it on the sentiment of the review is. The FTC rule at 16 CFR 465.4 reaches compensation or incentives provided in exchange for, or conditioned expressly or by implication on, a review expressing a particular sentiment. Free dessert for any honest review can be defensible. Free dessert for a five star review is the violation.

Can a restaurant delete bad reviews?

You can curate reviews you host on your own website, but not in a way that misleads. 16 CFR 465.7(b) reaches a business that materially misrepresents that the reviews shown on its own site represent most or all reviews submitted, when negative ones are being suppressed. The rule does allow removing reviews that are defamatory, contain personal information, are unrelated to your food or service, or are suspected fakes.

Do restaurants have to disclose paid influencer meals?

Yes, and a comped meal counts. Under 16 CFR 255.5 a material connection is any relationship that might materially affect the weight a viewer gives an endorsement, and free food is exactly that. Payment is not required for the disclosure duty to attach. The disclosure has to be clear and conspicuous in the video itself, not only in a caption that a viewer scrolling a feed never opens.

Can restaurants use AI generated food videos in ads?

For the presenter and the storytelling, generally yes. For the food itself, be careful. An ad that shows a dish you do not serve, or that materially misrepresents portion, ingredients or presentation, is deceptive under Section 5 of the FTC Act regardless of how it was produced. The safe construction is to generate the creator and the script while the dish on screen comes from your own photography.

Do Facebook and Instagram ads work for restaurants?

They work well for restaurants because the product is visual and the buying decision is impulsive and local. The constraint is rarely targeting and almost always creative volume. A three mile radius audience is small, so the same people see your ads repeatedly and frequency burns a creative out in days rather than months. Restaurants that succeed on paid social ship new video constantly.

How do I get more customers in my restaurant?

Fix the specific hole rather than advertising the restaurant in general. Identify which dayparts are soft, then build offers and creative aimed only at those hours, because a Tuesday lunch problem and a Saturday dinner problem need different messages. Then make sure your Google Business Profile, menu photos and ordering link are current, since most local ad clicks land there before anyone decides.

What is a good marketing budget for a new restaurant?

New locations usually need a front loaded budget, because you are buying awareness that does not exist yet rather than capturing demand that does. Published guidance often suggests a higher percentage of projected revenue in the first six to twelve months, then a step down once repeat traffic builds. Budget separately for the opening push and for the ongoing baseline, and do not judge one by the other's payback period.

How do restaurants get more direct orders instead of DoorDash orders?

Advertise the direct channel specifically rather than advertising the restaurant and hoping. Because a direct order keeps the commission that a marketplace order gives away, you can profitably pay far more to acquire one. Practical moves include putting the direct ordering link in every ad, adding an insert to marketplace bags, and offering something on the direct channel that the marketplace listing does not carry.

Can restaurant staff write reviews for their own restaurant?

Not without disclosing the relationship. 16 CFR 465.5 addresses insider reviews and testimonials, requiring clear and conspicuous disclosure of an officer's, manager's, employee's or agent's material relationship to the business where that relationship is not otherwise clear. It also reaches managers soliciting reviews from immediate relatives or staff when the resulting review carries no disclosure.

Keep the local feed fed

Paste your menu page URL or upload your own dish photos, pick a presenter, and get a UGC-style ad with voiceover and burned-in captions in 9:16, 1:1 or 16:9. No shoot day, no videographer, no waiting on a creator to reply. Free to start with your account, no credit card needed.

Try the generator free See pricing

Keep reading

Adjacent vertical