Google Ads for Restaurants: The 2026 Owner's Guide

Google Ads works for restaurants because it captures the single highest-intent moment in local dining: someone standing on the sidewalk or sitting on the couch typing "best tacos near me." Per LocaliQ's 2026 search advertising benchmarks, Restaurants and Food is one of the cheapest categories on the platform, with an average cost per click around $2.05 against a $5.42 all-industry average, and an average cost per lead around $30.27 against roughly $66.69 across all industries.

The catch is that Google's product surface has changed enough that most advice written before 2025 will steer you into the wrong campaign types and the wrong settings. This guide covers what a restaurant actually uses in 2026, the budget math, the structure and settings we install, the specific ways self-managing owners lose money, and short variants for home services businesses and clinics, whose rules differ in ways that matter right now.

Which Google Ads campaign types should a restaurant use in 2026?

Three, and only three. First, Search campaigns for branded and non-branded local intent, optionally with the AI Max enhancement layer. Second, Performance Max with your location and Google Business Profile feed attached. Third, Demand Gen, which in 2026 gained Google Maps as a standalone placement, including Maps-only campaigns.

Some historical context, because old blog posts still confuse owners: the dedicated "Local campaigns" type no longer exists. Google migrated Local campaigns into Performance Max in August and September 2022. If a guide tells you to launch a Local campaign, it predates the current product.

Here is how the three surviving pieces divide the work:

• Search captures explicit intent: "sushi near me," "restaurants on Lake Travis," "your restaurant's name." This is where most of your budget belongs, because the person has already told Google what they want.
• Performance Max extends your reach across Google's surfaces (Maps, Display, YouTube, Gmail) using your location feed and Google Business Profile. For a restaurant it works as a complement to Search, not a replacement. One watch-out: Performance Max can quietly absorb branded searches your Search campaign would have won more cheaply, so review the placement and search-category reports once it is live.
• Demand Gen with the Maps placement is the newest option. Google's Demand Gen documentation covers the campaign type, and the Maps placement rollout means you can now put visual ads in front of people browsing the map around your neighborhood, including Maps-only campaigns. One caveat, which is practitioner observation rather than Google guidance: Demand Gen is audience-first and has no conventional search-terms report, so you cannot audit it query-by-query the way you can a Search campaign. Treat it as a discovery layer you judge on store-level results.

What is AI Max, and should you turn it on?

AI Max is an AI enhancement layer for Search campaigns that Google announced on May 6, 2025 and rolled out globally in the third quarter of 2025. It expands matching and creative automatically, and Google is deprecating legacy broad-match settings and Dynamic Search Ads into it, so some version of this future is not optional.

Google reports a typical conversion lift of about 14 percent, and up to 27 percent for campaigns built mostly on exact and phrase match. Those are Google's own figures for its own product, so weigh them accordingly. Our practical position: AI Max is worth testing once your conversion tracking is trustworthy, because every AI layer amplifies whatever signal you feed it, good or garbage.

What budget does a restaurant actually need?

The practitioner math is straightforward: roughly $750 to $1,500 per month buys on the order of 350 to 750 clicks at restaurant-level CPCs, which is enough volume for Google's smart bidding to function. Below that, the algorithm never accumulates enough conversion data to optimize, and you are effectively paying full price for a system running at half intelligence.

A worked example at the benchmark averages:

Monthly budget	Clicks at ~$2.05 CPC	Leads at ~$30.27 CPL
$750	~365	~25
$1,000	~490	~33
$1,500	~730	~50

A "lead" here means a tracked action with real intent: a call over 60 seconds, a reservation, an online order, a direction request. One caution before you compare this figure across platforms: the roughly $30 cost per lead here comes from LocaliQ's search advertising dataset and measures search-intent actions, while the much lower costs per lead published for Facebook come from a different dataset measuring lead-form submissions, so the two numbers are not directly comparable.

Now run the unit economics against your own menu, and run them honestly. Suppose your average party is 2.5 covers at a $32 check, so roughly $80 per booked table. At the $1,000 budget and benchmark averages, 33 leads seated at 50 percent yield about 16 tables and $1,280 in gross revenue against $1,000 of media. That is not breakeven: at restaurant contribution margins, $1,280 of revenue covers only a fraction of the spend, so the first visit loses money on its own. The case for the channel rests on repeat business: a new guest who returns twice a quarter is worth several times the first check, which is why we insist on measuring cost per lead against lifetime value, never against the first ticket. Run the same arithmetic with your own check average, margin, and return rate before you commit, and revisit it monthly once real numbers replace the benchmarks.

For evidence this works in the category at real budgets, the agency 39 Celsius published a restaurant case study reporting $19,000 in incremental sales in 30 days. It is an agency's own case study, so treat it as a directional proof point rather than a promise, but the mechanism it describes matches what we see: concentrated local search budget, tight geography, clean tracking.

How should the account be structured?

The structure we install, which reflects broad practitioner best practice:

1. Branded Search campaign. Your restaurant's name and misspellings. Cheap, high-converting, and it keeps aggregators and rivals from owning your own name.
2. Non-branded local Search campaign. The money campaign: cuisine and occasion terms plus local modifiers ("thai food lakeway," "patio dining near me," "private dining austin").
3. Optional competitor campaign. Bidding on nearby competitors' names. Higher cost, lower conversion; only once the first two campaigns are saturated.
4. Performance Max with the location feed and Google Business Profile connected.
5. Demand Gen Maps as a test layer once Search is stable.

Three settings decisions matter more than the structure itself:

• Location setting: "Presence," not "Presence or Interest." The default includes people merely interested in your area, which for a restaurant means paying for clicks from people planning a trip three months out, or from nowhere near you at all.
• Dayparting around service hours and decision windows. Bid up in the 11am and 4 to 7pm windows when people decide where to eat; bid down or off in the dead hours after close. Nothing wastes a restaurant budget quite like winning the auction for "dinner near me" at 1am. An Austin restaurant should also think in terms of the local calendar: F1 weekend, SXSW, ACL, and UT home games all spike search volume and out-of-town intent, and those visitors search generically ("best barbecue austin") because they have no local habits yet. Budget those weeks deliberately, raise bids on non-branded terms, and accept a wider radius for that window only, using the same lead times and geo moves we lay out in our Austin event advertising playbook.
• Geography that matches your actual draw. A 5 to 10 mile radius for everyday dining; wider only if you are genuinely a destination.

How do you track conversions that happen offline?

This is where most restaurant accounts fail, because the conversion is a table filled, not a form submitted. The conversion actions worth configuring:

• Phone calls over 60 seconds (call reporting on ads and on the site number), filtering out wrong numbers and one-ring hangups.
• Direction requests from ads and profile.
• Menu views as a soft signal.
• Online orders and reservations as primary conversions, imported from your ordering or booking platform.
• Enhanced conversions, plus offline conversion upload for phone reservations, so the bidding system learns from bookings your book-keeper sees but your website never does.

And connect your Google Business Profile to the ads account, with your primary category set precisely ("Thai restaurant," not "Restaurant"). The profile powers your location assets, Maps presence, and a surprising share of how Google decides you are relevant to "near me" queries.

Checklist: the minimum viable tracking setup

• Call tracking on, 60-second threshold set
• Reservation or ordering platform conversions imported
• Enhanced conversions enabled
• Offline upload path for phone bookings
• Google Business Profile linked, primary category exact
• Direction requests and menu views tracked as secondary actions

Where do self-managing owners waste money?

Everywhere Google's defaults favor Google. The recurring offenders, which Search Engine Land's review of default settings documents in detail:

1. Search partners left on. The default opts you into a network of non-Google sites with weaker intent. Turn it off and watch nothing bad happen.
2. Display expansion on Search campaigns. Your budget quietly leaks onto banner placements you never chose.
3. Auto-apply recommendations. Google will helpfully rewrite your account toward broader matching and higher spend while you sleep. Turn it off; apply recommendations yourself, one at a time, on purpose.
4. Broad match with no negative keywords. Broad match plus smart bidding can work, but only fenced by negatives.
5. Sending clicks to the homepage. A person searching "book private dining" should land on the private dining page. Every extra click before the action is paid attrition.
6. No negative keyword list at all. Every restaurant account needs one from day one: "jobs," "hiring," "recipes," "calories," and the names of cities you do not serve. Ten minutes spent on negatives is some of the cheapest waste reduction available in a self-managed account; check the search terms report monthly and keep adding.

What if you run a home services business instead?

The restaurant playbook mostly transfers, with one urgent difference: Local Services Ads are migrating into Performance Max, phased from August 2026, per Google's own migration documentation. US home services and storefront verticals go first, service-area businesses follow in late 2026, and non-US plus remaining categories in 2027.

What changes in practice: campaigns move to pay-per-lead goals inside Performance Max and you pay only for valid leads, manual max cost-per-lead bidding and BBB callouts go away, weekly budgets convert to daily budgets (divided by seven), and leads move into a new Lead Manager. If LSAs are a meaningful lead source for you, audit your converted weekly-to-daily budget now, learn the Lead Manager before the switch is forced, and expect a re-learning period; do not schedule the migration month as your make-or-break season.

What if you run a clinic or health practice?

Healthcare is a restricted category on both major platforms, and the restrictions bind harder than most practice owners realize.

On Google, healthcare falls under the Sensitive Interest Category rules: no remarketing, no lookalike style audiences, and no custom segments built on health conditions, formalized in the May 2025 personalized-ads policy update. A July 2025 carve-out allows targeting licensed healthcare professionals, which helps B2B health marketing but not patient acquisition.

Meta is comparably strict: since January 2025, ad accounts categorized as health lost lower-funnel optimization events and on-site custom audiences, a change documented in Foley Hoag's January 2025 legal analysis and by measurement vendors such as Freshpaint and Triple Whale. The practical consequence on both platforms is the same: clinics compete on Search intent, geography, first-party creative, and landing-page quality, because the retargeting shortcuts are gone. Budget expectations should be set accordingly.

What we'd do first

Taking over a restaurant's Google Ads account, our first two weeks:

1. Fix tracking before touching spend. The checklist above, complete, because every later decision depends on it.
2. Kill the defaults. Search partners off, Display expansion off, auto-apply off, "Presence" only.
3. Rebuild to branded plus non-branded local Search, with a negative list installed on day one.
4. Set budget to the math, $750 to $1,500 per month concentrated on Search, dayparted to decision windows.
5. Only then test the AI layers: AI Max on Search, then Performance Max, then Demand Gen Maps, one at a time, each judged against tracked conversions.

Google captures the demand that already exists; Meta creates demand that does not exist yet, and the two work best as one system, which is why we wrote a companion guide on Meta ads in the age of Andromeda. Both belong inside a single accountable marketing program.

If you would rather see exactly where your account leaks before spending another month on it, request our free business audit and we will walk you through it line by line.