Google Ads for restaurants: the search you came here to buy is already answered, for free!

Restaurant PPC is mostly a question of what not to buy.

Every trade on this site has searches that look like customers and are not. A restaurant has that too, 215,000 a month of it, and most of it is people who want a shift, a supplier or something free. That part is ordinary and the negative list below handles it.

The unusual part is what happens with the search you do want. “Restaurants near me” is 5.6 million searches a month, and Google answers it with a map, for nothing, using the Business Profile you already own. An ad above that map is competing with your own free listing for the same person, and the person taps the map, because that is what people do. Most restaurant ad budgets die there, quietly, with a respectable click-through rate.

AdPilot: automated Google Ads for restaurants, catering and private dining

What is left is genuinely worth buying, and it is narrower than anybody wants it to be. Catering. Private dining and large parties. Holiday bookings, three months before the holiday. Your own delivery, instead of a marketplace taking a commission on every order. AdPilot connects to the account and works through it the way a consultant would if you were paying one out of the register. Each finding comes back as a card, in plain English. You approve it or you skip it. Nothing reaches your account until you do.

And you stay in control the whole way:

• You approve every change. The AI proposes; you decide. Every budget, bid, keyword, or ad edit arrives as a plain-language card you approve or skip with one tap. Nothing is ever applied on its own.
• A snapshot before every change. AdPilot saves exactly how things looked before anything is touched, so you always have a clean “before.”
• One-click revert. Don’t like a change once it’s live? Undo it in a single click, straight back to that snapshot.
• A full audit log. Who, what, when, the old value and the new value, every change recorded. Nothing hidden.
• You keep the keys. You connect with one secure Google sign-in, and it only ever makes changes you approve. The account stays yours start to finish - you never transfer ownership, and you can disconnect anytime.

Where restaurant ad budgets actually leak

Real US search volumes from August 2026. Most of these are people who want to work in a restaurant, supply one, or open one of their own.

The first row is the shape of the whole problem. “Server jobs near me” is 53,000 searches a month, it contains no word you would think to block, and a broad keyword on your restaurant's name plus its city reaches it more often than anybody expects.

US monthly search volume, Ahrefs, August 2026. Around 215,000 searches a month that will not book anything. Cross-checked against Google's own Keyword Planner, which groups close variants and reports several of these far higher.

Negative keywords for restaurants. Free, no email, no form.

Use this whether or not you ever become a customer. Ten minutes, no cost, and in this trade the job traffic alone usually pays for the effort in the first week.

Wants a shift

server jobs, waiter jobs, waitress jobs, chef jobs, line cook jobs, dishwasher jobs, restaurant manager jobs, bartender jobs, hostess jobs, hiring near me, now hiring, application form, food handlers card, servsafe, tip pooling rules, server salary, restaurant wages

Opening or supplying one

restaurant equipment, restaurant supply, restaurant supply store, commercial kitchen, walk in cooler, pos system, restaurant pos, restaurant software, how to open a restaurant, restaurant business plan, restaurant name ideas, food truck for sale, restaurant for sale, liquor license application

Free and discounted

free food, free meal, free birthday meal, coupons, coupon code, promo code, discount code, happy hour deals, half price, bogo, food bank, free samples

Delivery marketplaces

doordash driver, dasher, become a dasher, grubhub driver, uber eats driver, delivery driver jobs, doordash promo code, grubhub promo code, doordash sign up, doordash customer service

Reading, not booking

recipe, recipes, copycat recipe, how to make, nutrition facts, calories in, menu template, restaurant reviews reddit, michelin guide, food blog, cooking class

Chains and somebody else's brand

mcdonalds, chipotle, olive garden, cheesecake factory, applebees, chick fil a, taco bell, panera, buffalo wild wings, texas roadhouse

Restaurant PPC: five things that matter more than your bids

• Do not pay for the search your free listing already wins. Google answers “restaurants near me” with a map, drawn from the Business Profile you own for nothing, and the map is where the taps go. Bidding on the same query means paying for a customer your free listing was already going to get. Fix the Business Profile first: hours, photographs, menu link, reservation link. That is the highest-return hour in restaurant marketing and it costs nothing.
• Buy the bookings a map cannot take. Catering, private dining, large parties, holiday bookings, rehearsal dinners. Those searches draw no map, so there is no free listing quietly competing with your ad, there is often no serious bidder either, and a single order is worth what fifty covers are.
• Advertise your own delivery, not the marketplace's. Marketplace commissions commonly run 15% to 30% of the order. An ad that sends somebody to your own ordering page turns a commissioned order into a full-margin one, and the arithmetic on that is usually far better than the click price suggests.
• Holiday party season is booked in September. Corporate holiday bookings are decided a quarter ahead, and almost every restaurant starts advertising them in late November when the calendar is already full. Run that campaign in September and October or do not bother.
• Price per head belongs in the ad. An office manager comparing four caterers wants a number, a deadline and whether you set up. Publishing a starting price per head filters out everybody below your band before they cost you a click and gets you shortlisted by everybody above it.

What one catering inquiry is worth to you

Four numbers you already know, and the one most restaurant accounts are run without. The click price is measured: “catering near me” is 104,000 searches a month at about $1.10 and “office catering near me” 450 at about $3.00, from the same pull as the table above, so the slider opens at $1.50. The order opens at $4,200, which is not one tray: it is the $700 average order the guide works from, taken six times by an office that keeps coming back. Put your own numbers in, and use a single order instead if your catering is mostly one-off. Nothing leaves your browser.

The whole build, written out so you can check it

Below is every move, in the order the agent makes them, with nothing held back for the paid version. An operator who works through it alone ends up with an account that produces catering orders and private bookings. We'd rather you were able to check the method than take our word for it.

Budget a Saturday for the build, and spend the first hour of it not in Google Ads at all. After that it's half an hour every other week, for as long as the ads run.

The same walkthrough, written for every other industry

1. 1

   Fix the free listing before you spend a dollar

   This is the only page on this site that opens by telling you not to advertise yet, and the reason is arithmetic rather than modesty.

   “Restaurants near me” is 5,610,000 searches a month in the US. “Restaurants open near me” is another 214,000. “Best restaurants near me” is 335,000. Google answers all three with a map, filled with businesses that paid nothing to be in it. One of them is you, or should be.

   So an ad on those searches is competing against your own free listing for the same tap. Ads on a local query do sit high on the page, sometimes above the map and sometimes inside it, and none of that helps you: whichever one the person taps, they were coming to a restaurant near them either way, and only one of the two costs money. You will get a respectable click-through rate and a report that looks fine, and the money is gone.

   The hour that actually pays is the Business Profile. Hours that are correct, including the holiday ones. Photographs of the food and the room, taken this year, not the ones from the opening. A menu link that works on a phone. A reservation or ordering link that goes to your own system. The attributes filled in, because people filter on outdoor seating and takeout and dietary options. Reviews answered, all of them, including the bad ones, in a normal voice.

   Then check what your listing looks like on a phone, on mobile data, at seven in the evening. That is the moment your customer is looking at it, and it is a different screen from the one on your laptop at eleven in the morning.

   Only when that is genuinely finished does a paid campaign make sense, and it will be aimed somewhere else entirely.

   • Hours correct, holiday hours set, and special hours added before every holiday rather than after.
   • Photographs from this year. Food, room, exterior at night, and the front door so people can find it.
   • Menu link and ordering or reservation link, both tested on a phone.
   • Attributes filled in: outdoor seating, takeout, delivery, dietary options, accessibility, parking.
   • Every review answered, including the one-star ones, in the voice you would use at the door.
   • Look at the whole thing on a phone at dinner time. That is when it gets read.
2. 2

   Sort the searches that a map cannot answer

   Now the keywords, and the sorting rule is unusual here. Do not ask whether the searcher is a customer. Ask whether Google is already answering them for free.

   Browse searches, the ones with a map above them, are out. “Restaurants near me”, “best restaurants near me”, “restaurants open near me”, “food delivery near me”. All genuine customers, all already served, none worth buying.

   Booking searches are in, and they are what this campaign is. Here are the measured ones. “Catering near me” is 104,000 searches a month at about $1.10 a click. “Wedding catering near me” is 2,900 at about $0.90. “Catering services near me” is 1,800 at $0.90. “Corporate catering near me” is 700 at about $2.50. “Brunch reservations near me” is 500 at $0.90. “Office catering near me” is 450 at about $3.00. “Private event space near me” is 400 at about $1.00. “Restaurant reservations near me” is 350 at $0.80. “Private dining room near me” is 150 at $0.80.

   Add those up and you get roughly 111,250 searches a month that a restaurant can actually buy. Set that against 215,000 in the drains table and the picture is clear enough: in this trade there is more job traffic than bookable traffic, and the bookable part is dominated by one word, which is catering.

   Notice the prices. The generic catering term is a dollar. The corporate one is two and a half, and Google's own top-of-page range on it runs $11 to $63. That gap is the whole opportunity. The expensive term is expensive because the order behind it is worth twenty times a table for two, and most independent restaurants never bid on it at all.

   The last group is your own delivery, and it is the one people forget. Somebody typing your restaurant's name plus “delivery” is about to open a marketplace app and hand over a commission on an order that was already yours.

   • Ask it out loud on every term: does this search draw a map with your listing already in it? If yes, do not buy it.
   • Catering words: catering near me, corporate catering, office catering, wedding catering, drop off catering, catering menu prices.
   • Event words: private dining, private room, private event space, holiday party venue, rehearsal dinner, large group.
   • Own-delivery words: your restaurant name plus delivery, plus order online, plus takeout.
   • Not your searchers: jobs, supply, equipment, recipes, coupons, marketplace brand names.
   • Note the shape of every one of those. They're phrases, not single words, and step five explains what happens the day you shorten one.

   Eight restaurant searches read side by side: four you bid on, four you do not

   • catering near me - 104,000/mo, ~$1.10 - BUY. The volume term, and it is cheap for what it is.
   • corporate catering near me - 700/mo, ~$2.50, Google's range $11 to $63 - BUY. Small, pricey, and the best order you will take all month.
   • private dining room near me - 150/mo, ~$0.80 - BUY. Tiny nationally, real in your city, and hardly anybody bids on it.
   • wedding catering near me - 2,900/mo, ~$0.90 - BUY if you do them. One booking pays for a year of the campaign.
   • restaurants near me - 5,610,000/mo - the map answers it and your listing is already in the map, free. Not a keyword at any price.
   • server jobs near me - 53,000/mo - wants a shift. Block the phrase, and block "restaurant jobs" separately.
   • free food near me - 34,000/mo - 34,000 people a month, and not one of them is paying. Never bare "free" though: your own ad may offer a free tasting.
   • restaurant supply store - 26,000/mo - somebody buying napkins in bulk. Block the phrase.
3. 3

   A catering page with a price on it, not your menu

   An office manager who typed “corporate catering nashville” eleven seconds ago is comparing four places before a meeting at two. On most restaurant sites the catering link goes to a PDF of the dinner menu and a contact form.

   That person needs three things and they are always the same three. What does it cost per head, what is the ordering deadline, and do you deliver and set up. Answer all three above the fold and you will be shortlisted by people who never call the other three restaurants.

   Publish a starting price per head. A range is fine. “From $14 a head for drop-off, from $22 for staffed” is enough. The people below your band leave without costing you a phone call, and the ones above it now trust you, because you are the only one of four who answered.

   Put the practical constraints where they can be read: minimum order, how far you deliver, how much notice you need, what happens with dietary requirements. Every one of those is a reason somebody would otherwise call to disqualify you.

   Then make the ask small and fast. A quote request with five fields and a promise to reply within the hour beats a contact form every time, because this person is deciding today. If you can genuinely reply in an hour during business hours, say so and then do it. Of the four places this person contacts, the one that answers first is usually the one that quotes.

   Private dining gets its own page with the same discipline: room capacity seated and standing, minimum spend, what the set menu options cost, and photographs of the actual room rather than the dining room at large.

   • The first heading names the service and the city: Corporate Catering in Nashville.
   • Price per head visible without scrolling. A range is fine. A PDF is not.
   • Ordering deadline, delivery radius, minimum order, and whether you set up.
   • Photographs of catering trays and the private room, not of a plated dinner special.
   • Five form fields at most: name, company, date, headcount, phone.
   • A reply time you can actually keep, written on the page. Then keep it.

   The catering page, written out line by line

   • H1: Corporate Catering in Nashville
   • Under it: From $14 a head drop-off, from $22 staffed. Order by 4pm for next-day delivery.
   • Three lines high on the page: we deliver within 12 miles, minimum order 10 people, we set up and come back for the trays.
   • Menus: three packages with the price per head next to each one. Not a PDF.
   • Dietary: vegan, gluten free and nut free options on every package, marked on the tray.
   • The ask: Get a quote. Name, company, date, headcount, phone. We reply within the hour, 9 to 5.
   • Proof block: the companies you cater for regularly, with permission, and how long you have been doing it.
   • Never on this page: a link to the dinner menu as the main call to action, and the phrase “bespoke culinary experiences”.

4. None of this is difficult. It is a long run of small decisions, and each one quietly costs money when it goes the wrong way. The agent makes them for you and shows you every single one before it touches the account.

   Start free - no credit cardFree for 7 days. No card, no call, cancel whenever.
5. 4

   Three ad groups, and a calendar that runs the budget

   A campaign is a budget and a goal. For a restaurant that is one campaign with three ad groups: catering, private events, and your own delivery if you run it.

   There is no browse ad group. That is the whole argument of this page and it is worth saying once more in the place where somebody would build it.

   Now the arithmetic. Catering runs about $1.10 a click and the corporate terms nearer $2.50, so call it $1.50 blended. Twenty clicks a day is $30 a day, a $912 monthly ceiling, about 600 clicks. If one in twenty-five asks for a quote, that is 24 quote requests. If one in three books, eight catering orders. At an average order of $700 that is $5,600 of revenue against $912 of spend, and catering food costs are better than dine-in because you are not paying for a server to stand next to it.

   That is a good trade, and it is a completely different business from the one where the same $912 went on “restaurants near me”.

   Then the calendar, which matters more here than in any other trade on this site. Corporate holiday parties are booked in September and October. Wedding catering gets asked about nine to twelve months out. Graduation is May, and Mother's Day fills a dining room faster than anything else on the calendar and gets decided the week before. A budget set once in January is describing a market that has moved four times by December.

   Practically that means two things. Use ad scheduling so you are not paying for clicks at two in the morning unless you take orders then. And put a note in the calendar for early September that says raise the holiday budget, because you will not remember and by the time you do the good rooms are gone.

   Then leave the bidding alone. Google says the learning period runs up to three weeks, or one to two conversion cycles, whichever lands later. Separately, it says not to judge a target cost per action until you have thirty conversions in thirty days, and at 24 quote requests a month that is about six weeks. Start on Maximize clicks while nothing is tracked, move to Maximize conversions with no target the day it is, and never set a target cost per quote on a campaign with no history.

   • One campaign. Three ad groups: catering, private events, own delivery. No browse group.
   • Work backwards from the click price to the budget. Twenty clicks a day is a sensible floor.
   • Ad scheduling on, matched to when you can actually answer.
   • Holiday party budget goes up in early September. Put it in the calendar now.
   • Phrase match everywhere on day one. Broad stays off until something is tracked and converting.
   • Maximize clicks only while nothing is tracked. Maximize conversions, no target, the day it is.

   The budget, worked out on paper first

   • Blended click price about $1.50, from the measured terms in step two.
   • Twenty clicks a day is $30/day, a $912 monthly ceiling, roughly 600 clicks. Google can spend twice the daily figure on any one day and never more than 30.4x it in a month.
   • At one quote request per 25 clicks, that is 24 quote requests a month.
   • At one booking per three quotes, eight catering orders.
   • At an average order of $700 that is $5,600 of revenue against $912 of ad spend.
   • Compare the same $912 spent on "restaurants near me": a good click-through rate, some covers you would probably have had anyway, and no way to tell the difference.
   • September: move an extra $15/day into the private events group for eight weeks. Holiday parties are decided before Halloween.
   • Thirty conversions at 24 a month is about six weeks of learning. Don't touch the bid strategy in week two.
6. 5

   Block the job hunt, and never block the word menu

   Negatives go in before the campaign serves an impression. In this trade the biggest single category is people looking for work, and it is bigger than most operators guess: 131,800 searches a month across servers, cooks, chefs, managers and the certification they need to start.

   Start there. Server jobs, waiter jobs, waitress jobs, chef jobs, line cook jobs, dishwasher, bartender, hostess, restaurant manager jobs. Then the certification words, which nobody thinks to block because they do not look like job searches. “Food handlers card” alone is 24,000 searches a month, and servsafe sits next to it unmeasured.

   Then the supply side: equipment, supply store, commercial kitchen, walk in cooler, pos system. Then the people opening one: how to open a restaurant, business plan, name ideas, restaurant for sale. Then the free traffic: coupons, promo codes, free birthday meal. Restaurant Week is the one to think about rather than paste. Ten thousand searches a month, and every one of them is a drain if you are not participating and a booking if you are.

   Now the marketplaces, and be careful here, because the obvious move is wrong. Do not block bare “doordash”, “grubhub” or “uber eats”. A phrase negative on “doordash” also blocks “larkspur kitchen doordash”, and that search is somebody thirty seconds from handing a marketplace a commission on an order that was already yours. It is the cheapest margin you will buy all year and your own delivery ad group exists to catch it. What you block instead is the traffic around the brand that has nothing to do with you: drivers looking for shifts, people hunting a promo code, and anybody trying to reach the marketplace's own support line.

   Negatives have a reach problem nobody warns you about. They match what you typed and nothing near it. No plural, no synonym. “Job” does not reach “jobs”. “Recipe” does not reach “recipes”. Put both forms in and read the list back hunting for the word you entered once.

   Now the words that must never go in bare, and this trade has an unusual number of them. Never bare “menu”, because “catering menu prices” and “private dining menu” are people ready to book. Never bare “party”, because “holiday party venue” and “birthday party restaurant” are two of the best searches you will see. Never bare “order”, because your own delivery group depends on it. Never bare “near me”, which is in nearly every term you are buying. Never bare “free”, because your own ad may offer a free tasting to a catering prospect. Never bare “catering”, which sounds obvious and gets typed into a negative list by somebody trying to block “catering jobs”. And never bare “doordash”, “grubhub” or “uber eats”, for the reason two paragraphs up.

   Pick the narrowest form that stops the waste. Negative phrase blocks your words in that order. Negative broad blocks a search containing all of them in any order, which is a hammer. Phrase is nearly always right here, because almost every trap in this trade is a safe word standing next to a dangerous one.

   • Campaign level, in before the first impression, not after the first bad report.
   • The job list first. It is the biggest category and the least obvious one.
   • Food handlers card and servsafe go on the list. 24,000 a month and they do not look like job searches.
   • Singular and plural on every entry. Read the list back hunting for the word you typed once.
   • Phrase, not broad, unless you can say out loud what broad would also kill.
   • Nine words stay bare-free: menu, party, order, near me, free, catering, doordash, grubhub, uber eats.

   What goes in before the first impression

   • server jobs, waiter jobs, waitress jobs, chef jobs, line cook jobs, prep cook jobs, dishwasher jobs, busser jobs, bartender jobs, hostess jobs, restaurant manager jobs, hiring near me, now hiring, job application, server salary, restaurant wages, tip pooling rules
   • food handlers card, food handlers permit, servsafe, servsafe practice test, alcohol server certification, food safety course
   • restaurant equipment, restaurant supply, restaurant supply store, commercial kitchen, walk in cooler, pos system, restaurant pos, restaurant software, uniform supplier, linen service
   • how to open a restaurant, restaurant business plan, restaurant name ideas, restaurant for sale, food truck for sale, liquor license application, restaurant startup costs
   • coupons, coupon code, promo code, discount code, free food, free meal, free birthday meal, half price, bogo, happy hour deals
   • doordash driver, dasher, become a dasher, grubhub driver, uber eats driver, delivery driver jobs, doordash promo code, grubhub promo code, doordash sign up, doordash customer service
   • recipe, recipes, copycat recipe, how to make, nutrition facts, calories in, menu template, food blog, cooking class, michelin guide
   • Careful: never bare menu, never bare party, never bare order, never bare near me, never bare free, never bare catering, never bare doordash, never bare grubhub, never bare uber eats. Every one of them sits inside a search you are paying to win - "catering menu prices", "holiday party venue near me", "order online for pickup", "private dining room near me", "larkspur kitchen doordash".
   • That last one is the whole reason the marketplace list above blocks drivers and promo codes instead of brand names. Somebody typing your restaurant plus doordash is about to hand over a commission on an order that was already yours, and that click is the cheapest margin you will buy all year.

7. Everything past this point is the work that never finishes: reading search terms, adding negatives, checking what the budget did overnight. That is exactly the part the agent takes off your hands.

   Start free - no credit cardFree for 7 days. No card, no call, cancel whenever.
8. 6

   Twelve headlines, and three of them are logistics

   Google gives you room for 15 headlines at 30 characters and 4 descriptions at 90, and it won't save the ad under 3 and 2. Fill 10 to 12 and three, and only with lines that say something the others don't. Put every line through a character counter rather than eyeballing it.

   You're not shipping twelve headlines. You're shipping every trio Google can build out of twelve, in an order nobody shows you, and each trio bills at full price. So pull the three that say least about the logistics and read them as if they were the whole ad.

   The failure mode in this trade is a menu description. “Exceptional Cuisine”, “Farm To Table”, “A Memorable Experience” is a finished ad for a wedding photographer. It says nothing about price, headcount, deadline or delivery, which are the only four things the person searching cares about at that moment.

   This is the one trade on the site where the boring headlines outperform the appetizing ones. The person buying office lunch for thirty is not being seduced. They are solving a problem before a meeting.

   One job per headline. The subject repeats, the job never does. “Corporate Catering, Nashville” is the search term, “Office Lunch From $14 A Head” is the price, “Order By 4pm For Next Day” is the deadline, “We Deliver And Set Up” is the objection. Keep two or three lines under twenty characters, because a set where everything runs 29 or 30 gets trimmed to two slots and you lose a third of the ad.

   If you run your own delivery, one headline says so plainly. “No Platform Commission” is not for the customer's benefit, it is for yours, and it turns out a fair number of people prefer ordering direct once somebody tells them the option exists.

   • 10 to 12 headlines and two or three descriptions, counted in a character counter rather than guessed.
   • Price per head, ordering deadline and delivery are all in there somewhere.
   • Read the three that say least about logistics as though they were the whole ad.
   • One quality claim, one slot. Fresh, authentic and handcrafted are the same headline three times.
   • Two or three headlines under 20 characters, so Google can fit a third on the line.
   • If you make a dietary claim in a headline, the page has to back it and the kitchen has to deliver it.

   Nine drafts, counted, and the two that had to be rewritten

   • Corporate Catering, Nashville (29) - the search term, exactly as it gets typed.
   • Office Lunch From $14 A Head (28) - the price, and the cheapest filter you own.
   • Order By 4pm For Next Day (25) - the deadline, which is the second question every time.
   • We Deliver And Set Up (21) - the objection, answered before the click.
   • Menus For 10 To 200 (19) - the capacity question, short enough to always fit.
   • WEAK: Exceptional Local Cuisine (25) - a menu description. Says nothing the searcher needs.
   • FIXED: Vegan And GF Options (20)
   • WEAK: Unforgettable Events (20) - true of a wedding photographer, a DJ and a party bus.
   • FIXED: Private Room Seats 40 (21)
9. 7

   Count the quote, and the half hour that keeps it honest

   Two numbers decide whether any of this works and Google has neither. The first is what a booking is worth. An average catering order of $700 is not the number, because a company that orders lunch once orders it again. If a corporate account comes back six times a year, the first order is worth $4,200 of revenue, and that is the figure the bidding should be working from.

   Divide by how many quote requests it takes to win one. At one in three, a quote request is worth $1,400 on the repeat number, or $233 on the single-order one. Pick the one you can defend and put it in the conversion value field.

   Make the results countable. If people call, Google will build the call conversion for you: a forwarding number shows in the ad and any call past a length you set counts. Sixty seconds is too short for a catering call and about right for a reservation, so if this campaign is mostly catering, set it higher and accept fewer, better rows. If they fill in a quote form, the conversion fires on the thank-you page, which is two snippets Google hands you. Before you go near a developer, check auto-tagging is on and look for conversions Google already imported from Analytics and left switched off.

   There is one measurement problem specific to this trade and it is worth naming rather than pretending away. A lot of restaurant business is somebody seeing the ad, remembering the name, and walking in on Saturday. Nothing in Google Ads will ever show you that, and store visit reporting only appears on accounts with enough volume to qualify. So do not judge a catering campaign on covers. Judge it on quote requests and catering orders, which are countable, and treat the walk-ins as a bonus you cannot invoice.

   Day one and two you look at two things. Is it spending, and is the ad serving. Day seven, open the search terms report and add negatives, and in this trade week one will be full of job searches you did not predict. Leave bids and budgets alone.

   The real read is day fourteen to thirty. Cost per quote request by ad group. Search terms again, this time for waste and for gold: anything that produced a quote goes back in as an exact match keyword with its own ad. That harvest is the job and it never stops being the job. Then one decision per ad group at month two or three: feed it, fix it, or stop it. What costs the most here was never the build. It is the two weeks you skipped in December, which is precisely when the budget was doing the most.

   • Auto-tagging on. One switch, no downside, and everything else leans on it.
   • Look for conversions imported from Analytics and left disabled. Switch them on.
   • Call conversion with a minimum length that matches a real catering call, not sixty seconds.
   • Conversion value = what a catering account is worth over a year, divided by quotes per won order. Not the single ticket.
   • Judge it on quote requests and orders. Covers are real and not countable, and pretending otherwise leads somewhere bad.
   • Google counts from the day you switch it on and never fills in the past, so turn tracking on before you turn the ads on.

   Fourteen days of search terms, and what each line is worth

   • "corporate catering nashville" - 11 clicks, 3 quotes -> add as exact match, with its own ad.
   • "office lunch delivery nashville" - 8 clicks, 2 quotes -> add as exact. Nobody else is bidding on it.
   • "private dining room downtown nashville" - 5 clicks, 1 quote -> add as exact, and give it its own landing page.
   • "holiday party venue nashville" - 14 clicks, 2 quotes -> add as exact, and note that this is September traffic. Raise the budget before October.
   • "server jobs nashville" - 31 clicks, 0 -> negative phrase "server jobs". Thirty-one clicks at $1.50 is $47, and it was the biggest single line in the report.
   • "food handlers card tn" - 19 clicks, 0 -> negative phrase "food handlers". It does not look like a job search, which is exactly why it survived the first list.
   • "restaurant supply nashville" - 12 clicks, 0 -> negative phrase "restaurant supply".
   • "catering jobs nashville" - 9 clicks, 0 -> negative phrase "catering jobs". Never bare "catering", which is the word the whole campaign runs on.
   • "free lunch downtown nashville" - 7 clicks, 0 -> negative phrase "free lunch". Not bare "free", because your own ad offers a free tasting.
   • "restaurant week nashville menu" - 6 clicks, 0 -> negative phrase "restaurant week". A city promotion, and the searcher is picking from a list you are not on.
   • Every one of those got past a list already blocking waiter jobs, doordash and coupons. That is what the half hour is for. The list is never finished.
   • And the report is not the whole month. Google withholds low-volume searches behind a privacy threshold, so your rows will always add up to less than your spend.

One finished ad for a corporate catering ad group

One ad group, corporate catering in Nashville, for a restaurant that doesn't exist. Twelve headlines, every one counted, and one description. The card shows the first three, because three is the most any searcher ever gets; the other nine sit underneath as what they are, spare parts Google draws from. Two descriptions is the least Google will save and four is the most, and it shows one or two depending on the space it has, so the one below is written to carry the ad on its own.

• Not one headline describes the food.. Price, deadline, delivery, capacity, dietary options, invoicing. Not one word about the food, and it takes chefs a while to accept that. The person buying lunch for thirty before a two o'clock meeting is solving a problem, not choosing a restaurant for their anniversary.
• “Same Invoice Every Time” is doing more work than it looks.. Office managers order catering repeatedly and the thing that loses a supplier is not the food, it is a bill that arrives different every month and has to be queried. Nobody advertises billing consistency, which is exactly why it stands out in a row of ads about culinary passion.
• The weakest trio still books a caterer.. Pick the weakest three deliberately, because any three you happen to sample will pass. Here the ones that say least about logistics are “Vegan And GF Options”, “Private Room Seats 40” and “Book Your Holiday Party”. A stranger reading only those knows there is catering with dietary options, a private room with a capacity, and that they can book an event. Compare “Exceptional Cuisine | Farm To Table | A Memorable Experience”, which is a finished ad for a wedding photographer.
• “No Platform Commission” is aimed at your margin, not their appetite.. Marketplace commissions commonly run 15% to 30% of the order. A fair number of customers will order direct when somebody tells them the option exists, and almost no restaurant tells them. Check the wording against your own marketplace agreements before you run it, because some of them have opinions about how you may advertise against them.
• The lengths run 19 to 29 characters on purpose.. Google shows up to three headlines, but only when they fit the space it has. A set where everything is 29 or 30 characters routinely gets trimmed to two, and you lose a third of the ad you wrote. “Menus For 10 To 200” at 19 is in there to be the one that always fits.
• The sitelinks are the free half of the ad.. Four of them roughly double the height your ad takes on the page and they cost nothing extra. These answer the second question a buyer has after the first one: what is on the menu and what does it cost, is there a room, will you come this far, and how fast can I get a number.
• One description on screen, two before Google will save it.. The floor is two and the ceiling is four, and it shows one or two depending on the room the page has, so write the first to stand alone. The one above carries the service, the city, the price, the deadline and the delivery. Write a second that carries what it leaves out, either the capacity range or the dietary options, and a third only if it makes a claim the first two do not.

The shape a single restaurant on a real budget actually runs

One campaign, three ad groups, and deliberately no browse group. Keywords start on phrase match. Exact match is earned later by the searches that actually produce quote requests. The budget below is the ordinary-month version; step four explains what happens to it in September.

Why it's shaped like this. There is no browse ad group, and that absence is the single most valuable thing on this page. "Restaurants near me" and its relatives are answered by a map pack sitting above the ads, using a Business Profile that costs nothing, so a browse group is a restaurant paying to outrank itself. Catering carries the highest bids and takes the largest share of one campaign budget, because that is the only lever there is - Google has no ad group budget - and because catering is where the volume and the order size both are: 104,000 searches a month on the generic term alone, and an average order worth what fifty covers are. Private events looks small in a keyword tool and is not, because those searches barely register nationally while being entirely real in one city, and hardly anybody bids on them. Own delivery sits in the main campaign rather than the brand one because it is a margin play rather than a defensive one: somebody typing your name plus delivery is thirty seconds from opening a marketplace app, and that click converts a commissioned order into a full-margin one. There is no ad group for reservations. "Restaurant reservations near me" is 350 searches a month nationally and your own listing already carries a booking link, so it is not worth the split. Your own name sits alone because brand clicks are cheap and convert better than anything else you run, so blended into the main campaign they drag the average down and hide what a catering quote really costs to buy. One thing the shape above does not solve: seasonality runs this trade more than any other on the site, and no structure fixes a budget nobody moved. September for holiday parties, January for wedding catering, and the week before Mother's Day for everything else.

Google Ads management for restaurants, shown as cards you approve one at a time.

What you actually pay a restaurant marketing agency for is the hands-on part. Keeping the budget off the searches your free listing already wins, finding the catering and private dining terms nobody else bids on, digging through search terms for job hunters, moving the money into holiday parties in September. AdPilot’s AI does that work, and explains the thinking behind each move.

Examples of what the cards look like - not results from your account

Holiday parties are booked in September. Not in the December report.

You keep control, and you stop being dependent.

Most operators never learn how their own Google Ads account works, which matters here because restaurant margins do not absorb a wasted month the way other businesses can. AdPilot is built the other way round. Every card tells you:

• What it changes - for example, pause the browse ad group that spent $340 on searches your free map listing already answered.
• Why it helps - the reasoning, not just the instruction.
• What it is based on - the numbers in your own account.

Approve one and a snapshot is saved first. Then a before and after dashboard shows what happened to cost per quote request, clicks and spend, in plain words. After a month or two you know which levers actually produce bookings.

AdPilot is new - and we’d rather earn your trust than fake it.

We could plaster this page with five-star quotes and a “trusted by 10,000 businesses” banner. We won’t - because AdPilot is new, and inventing proof would be lying to you. Instead, here’s what actually protects you:

• You risk nothing to try it. 7 days free, no credit card. If it isn’t useful, you walk away having spent nothing.
• Your account is never at risk. Nothing is ever applied without your one-tap approval, a snapshot is saved before every edit, every change has one-click revert, and a full audit log records it. The worst case is a suggestion you don’t like - so you just don’t approve it.
• Guardrails you can’t trip by accident. AdPilot never deletes campaigns and never touches shared budgets. Some doors simply don’t exist.
• Built by an operator, not a VC-funded black box. Software made to give small businesses the Google Ads work an agency charges a retainer for - without agency prices or the loss of control.

Judge us on the work, the math, and the safeguards - on your own numbers, not our adjectives.