Google Ads Campaign Types for Restaurants

It is a setup problem that starts at the moment someone picks a campaign type without first matching it to a conversion action tied to physical visits.

Google Ads Campaign Type Selection and Whether Spend Traces to Seats or Disappears into Clicks

Nothing in that dashboard tells the owner why. Search campaigns run mostly on a pay-per-click basis, and other formats like Display, YouTube, and Performance Max can also charge per impression or per view, so every one of those clicks and views costs real money whether or not it ends with someone walking through the door. Google reports clicks, impressions, and click-through rate by default, and none of those numbers confirm that a seat got filled. A restaurant that never connects its ad account to a conversion tied to a physical visit is, by definition, measuring something other than the thing it cares about and then optimizing its budget toward that wrong thing. The campaign-type decision is where this either gets fixed or gets baked in, because each type reaches a different point in a diner's decision process, at a different price, with different signals available to confirm whether anyone actually showed up. Choosing the wrong type, or the right type pointed at the wrong goal, makes it structurally impossible to trace spend back to revenue. That is not a copywriting problem and it is not a budget-size problem. It gets decided at setup, before a single ad has run.

What each Google Ads campaign type actually does, and where it sits in the path to an in-person visit

Search campaigns put text ads at the top of the results page when someone types in a relevant query. These catch people in active decision mode: "breakfast near me," "best pizza in [city]," "happy hour downtown," often chosen within the hour they're searched. Search sees fewer impressions than other formats but carries the sharpest intent of anything Google Ads offers; it remains the most common starting point for restaurants trying to reach someone who is deciding where to eat right now. A growing slice of that intent now arrives by voice, with people asking an assistant for "the best sushi restaurant nearby" instead of typing it.

Display campaigns run banner and image ads across Google's network of partner sites and apps. These reach people who are not actively searching for anywhere to eat, which makes them a reasonable tool for brand awareness and for showing off a dining room or a plate of food, but a poor one for anyone trying to track an actual visit. Display builds familiarity rather than capturing the decision-stage intent that Search does, and it works better as a retargeting layer aimed at people who already know the restaurant exists than as a way to generate new diners from a cold start.

Maps ads, surfaced through Local campaigns or Maps placement, show up directly inside Google Maps when someone searches for nearby places to eat. They put directions and contact information right inside the map interface, which makes them one of the only formats built explicitly to move someone from phone screen to front door. Anyone searching Maps for dinner options is, almost by definition, closer to walking in than someone glancing at a banner ad.

Video campaigns run on YouTube and partner placements. They're suited to showing off atmosphere, a tasting menu, or an event, which puts them firmly in the awareness-and-consideration part of the funnel rather than the high-intent capture that Search or Maps provide. Tracking a video view through to an actual visit takes extra setup, and video alone rarely drives walk-in traffic without another campaign type doing the heavier lifting.

Performance Max, or PMax, is a format. Google built it alongside Search as part of what the company calls an "Ads Power Pair," designed to run across every Google channel from a single campaign. One PMax campaign can serve ads across Search, Shopping, Maps, YouTube, Display, Discover, and Gmail at the same time, with Google's own algorithm deciding how to split the budget across those placements. For a restaurant specifically, Maps tends to soak up the majority of PMax's impressions, which makes the format especially useful for generating direction requests and store visits rather than, say, newsletter signups. Practitioners who build these campaigns for a living call PMax a "workhorse" once the conversion actions and location data behind it are set up correctly. PMax drives a large share of all conversions recorded across Google Ads, and one documented case, a Maps-weighted PMax campaign for Hundred Mile Brewing, produced thousands of store visits at a low cost per visit. A separate case produced hundreds of new visits at a low cost per visit for another restaurant chain, with every dollar spent traced back to an actual transaction rather than a click.

Search and Performance Max as a Two-Engine System for Restaurants

Restaurant advertisers who use both Search and PMax aren't hedging their bets. Each format covers a portion of the customer journey that the other one structurally cannot reach. Search is built to catch someone at the exact moment they're deciding where to eat, but it can only catch people who are already typing the search. It has no reach into the much larger group of people who are near the restaurant, hungry, and simply haven't thought to look yet.

That's the gap PMax is built to fill. It spreads across Maps, YouTube, Discover, and Gmail, and uses Google's machine learning to find people likely to visit before they've run any search at all. Its Maps-heavy impression share means it is functioning as a persistent local-awareness layer alongside Search's intent-capture role. Google describes Search and Performance Max as the "Ads Power Pair," designed to improve campaign efficiency across Google channels by leveraging first-party data and machine learning together.

Search lets a restaurant pick its own keywords and set its own bids, which allows for real precision: targeting by time of day to separate a lunch crowd from a dinner crowd, adjusting for device, layering in demographic targeting where it's useful. PMax doesn't offer that same dial-turning. It delegates the optimization decision to the algorithm. The only lever an operator really has left is the quality of the conversion signal fed into it. Whatever PMax is told to count as success becomes the thing its AI chases, so a campaign optimizing for website clicks will end up serving a noticeably different mix of placements than one optimizing for store visits or direction requests. That single configuration choice, not the campaign type itself, decides what Performance Max actually spends its budget chasing, which sets up the question the rest of this piece is really about: what should a restaurant be telling Google Ads to optimize for in the first place?

How Conversion Actions Determine What the Campaign Optimizes Toward

The single most consequential setting in any restaurant's Google Ads account is the conversion action, not the campaign type. It's the conversion action, the specific event the system is told to chase, and the default options Google suggests tend to push operators toward numbers that sound good but confirm nothing about whether a diner showed up. Left alone, Google Ads will recommend optimizing for clicks, form fills, or website visits, all of which are easy to measure and all of which fall short of proving anyone actually walked in. A click on a menu page is not a reservation. A reservation is not a visit. The only event that actually fills a seat is the one at the host stand, and none of Google's default suggestions get anywhere near it.

Follow the chain through: a campaign told to optimize for clicks will, over time, get better and better at finding people who click. There's no guarantee that group overlaps with the group of people who actually show up to eat. The algorithm is doing what it was told to do, which is the whole problem, because what it was told to do was the wrong thing.

A few conversion actions sit much closer to an actual visit. Direction requests, triggered when someone asks for directions to the restaurant from a Maps ad, serve as a strong stand-in for real intent to show up, and they can be tracked directly as a conversion, landing meaningfully closer to a walk-in than any website click ever will. A completed reservation, if the restaurant runs an online booking system, can fire a conversion tag off the confirmation page itself, closing the loop between the ad and an actual future seat. Phone calls generated through call extensions, provided they run past a minimum duration, can also count as conversions, which captures the diner who calls to book instead of clicking through online. Google's own native store-visit tracking uses mobile location data to estimate how many people who saw or clicked an ad later showed up in person. It's directional rather than exact, and it requires meeting certain volume thresholds to even activate, but among everything the platform offers natively, it comes closest to measuring the thing a restaurant actually cares about.

Here is where the chain either holds or breaks: feed the campaign the wrong signal, and it learns to find the wrong audience. Feed it clicks. It gets efficient at generating clicks. The targeting drifts away from people likely to visit, cost per actual customer climbs, and at some point a campaign that used to perform fine simply stops working, with nothing in the dashboard explaining why. First, compare what the platform reports against the restaurant's own point-of-sale records: if the ad platform is claiming substantially more revenue than the POS actually took in, more than one platform is probably claiming credit for the same visit. Second, watch for branded clicks, people searching the restaurant's own name who were coming in regardless of whether an ad existed. Google still bills for those clicks, and they inflate the conversion count without representing any new demand at all.

Separating Private-Event Campaigns from Walk-In Campaigns from the Start

One mistake recurs often enough to deserve its own section: running private-event lead generation in the same campaign as walk-in traffic. Practitioners have converged on a two-engine strategy because Search and PMax cover different parts of the customer journey that neither handles alone.

A store-visit conversion might cost under a dollar to generate. Putting both conversion types into the same campaign leaves the algorithm with no way to tell them apart. It optimizes toward whichever one fires more often, almost always the cheaper, more frequent walk-in signal rather than the rarer, more valuable event lead. The campaign ends up underbidding on the leads that matter most and overspending on walk-in clicks that were never going to turn into a 100-person holiday party anyway.

The fix is structural rather than a bid adjustment. An event-inquiry campaign should target phrases like "private dining," "event space," or "restaurant buyout," with a submitted inquiry form as the conversion action. A walk-in or reservation campaign should target phrases like "dinner near me" or "book a table," with a completed reservation or a direction request as the conversion action. Keeping these separate doesn't just protect each campaign's optimization, it also makes every future budget decision more legible, because the restaurant can finally see what it costs to generate a walk-in versus what it costs to generate an event lead, instead of one blended number that tells it nothing about either.

The foundational setup that must exist before any campaign type can attribute spend to in-store visits

None of the preceding sections matter if the infrastructure sitting outside the ad platform isn't in place first. A campaign can be built with the right type, the right conversion action, and clean separation between goals, and still fail to connect a single dollar to a visit if the groundwork underneath it is missing.

That groundwork starts with Google Business Profile. A verified profile is required before location assets can appear in ads at all, and before Maps placements can function correctly. Google Business Profile, which used to go by Google My Business, is what links the ad account to an actual verified address. Without it, Maps ads and location extensions simply have nothing to attach to. Keeping the name, address, phone number, and hours consistent across every directory matters here too, since conflicting location data gets in the way of the store-visit tracking the whole system depends on.

The website and landing page matter just as much. Whatever a Search ad promises has to be the first thing a visitor sees on the page it leads to. A mismatch between what the ad says and what the landing page shows raises the bounce rate and drags down Quality Score, and a lower Quality Score raises the cost of every future click. Mobile matters here in a very literal sense: most restaurant ad clicks happen on a phone, and a page that loads slowly or renders badly on mobile loses the visitor before any conversion has a chance to fire. None of this works without conversion tags installed on the actual confirmation pages, whether that's a completed reservation or a submitted form, because without that tag, the algorithm has no way to know a conversion happened.

Location extensions that show the restaurant's address and a map link right inside the ad raise the odds that someone nearby clicks through and shows up. Call extensions, promotion extensions, and sitelinks pointing toward the menu or the reservation page all add useful information before anyone even clicks, and without that small addition, the rest of the campaign structure described in this piece never gets the chance to prove itself.