What a restaurant owner should check on their Google listing every month

A Business Profile is not a thing you finish. It is a record that drifts, the same way par levels drift and the same way a prep list drifts. Nobody breaks it on purpose.

The difference is that when a prep list drifts you find out at four o'clock. When your listing drifts you find out from a review by somebody who drove across town on a Monday you were closed.

What actually changes on a listing without me touching it?

More than most owners realize, which is the entire reason this check exists.

Other people can edit it. Customers, competitors, and passersby can suggest changes to your hours, address, category, or even whether you are permanently closed. Some of those get applied.

Google updates things on its own. It reads your website and other sources and will adjust facts it believes are outdated.

Integrations overwrite your link fields. A reservation platform or delivery marketplace can populate the booking or ordering link the moment you connect it, and again the next time it reconnects. If a marketplace has quietly claimed your ordering link, you are paying commission on orders you could have taken directly, which is the situation described in what to do when a delivery app outranks your own website.

That field matters more than it looks. Whitespark's 2026 Local Search Ranking Factors survey, in which 47 local search experts scored 187 factors, added website URL forwarding to a different domain to its 2026 list of negative and suspension-risk factors. A link field pointed at somebody else's domain is not a neutral choice.

Customers add photos and answer questions. Strangers answer questions about your restaurant in public, often incorrectly, and their answer sits there with a vote count.

Duplicates appear. Aggregators and platforms create listings. Two listings for one restaurant splits your reviews and confuses everything.

None of that is malicious. It is what happens to a public record.

What belongs on the monthly check?

Eight items, in this order, because the first three cost you money the fastest.

Hours, and special hours for the next sixty days. Every holiday, every private buyout, every week the kitchen closes early. Enter them ahead, not the morning of. In the 2026 Whitespark survey, "business is open at time of search" entered at number five out of 187 factors. Darren Shaw describes the effect: "When your business is open you can be sitting pretty at #1 in the local results, and then 60 minutes before you close you'll start to drop off, and then when you're actually closed you'll be displaced by businesses that are open."

The three link fields. Website, menu, and the ordering or reservation link. Confirm each one goes where you want the money to come from.

Reviews since last month. Two questions. How many arrived, and did anyone go unanswered. Then look at the profiles ranking above you and count theirs. Recency of reviews ranks eleventh in the same survey and a sustained influx rather than bursts ranks fourteenth, so the pace matters more than the total. If they are collecting steadily and you are not, that gap widens quietly.

On responses, Google's own guidance is worth following rather than automating: "Instead of sending the same 'thank you' to everyone, focus on reviews where you can share a helpful update or answer a question."

Questions and messages. Answer anything a stranger answered, and clear any message sitting unread. A message asking about a party of fourteen that sat for six days is a booking somebody else took, which is the same failure as a slow response on a catering lead.

Photos. Add two or three. Then look at the first three photos a customer sees and confirm none of them is a dark flash photo somebody uploaded.

Suggested edits and updates by Google. Read them. Revert what is wrong.

Duplicates. Search your name and address. If a second listing exists, start the merge.

Categories, and the NAP that supports them. Primary category especially. If you added a bar program or a late kitchen, the category and the attributes may need to move with it, and dayparts have their own eligibility rules covered in ranking for the daypart rather than the cuisine. While you are there, confirm the name, address and phone in the text of your website match the profile exactly. That match scores fifteenth on the same list, and profile completeness sixteenth.

What should I look at quarterly instead?

The slower things, so the monthly check stays short enough to actually happen.

Attributes, in full. Service options, dietary options, accessibility, parking, atmosphere. These change with the seasons and the buildout, and reviewing them four times a year is enough.

Your description and your services list. Read them as a stranger. Delete anything no longer true.

Your position from several distances. Not your rank from behind the host stand, which is always flattering. Run your main searches from a mile out and three miles out, logged out.

Profile performance over a trend rather than a month. Calls, direction requests, and website clicks bounce around week to week and mean very little in isolation. The pattern across a quarter is the thing worth reading, and which numbers actually deserve your attention is the subject of the marketing metrics that predict cover count.

What numbers does the profile actually hand me?

Fewer than you would expect, and knowing the list keeps you from asking for something that does not exist.

Google's Business Profile performance reports publish interactions, the search terms people used, views, direction requests, calls, website clicks, messages, bookings, booking clicks, products, menus and offers.

There is no rank in there. No position. No impression-to-click pairing by position. Google Search Console has no local pack dimension either. So nobody, including your agency, can tell you what a move from position three to position one is worth in clicks, because the data required to calculate it is not published anywhere.

Two things also changed underneath those reports. Sterling Sky documents that Google removed Photo Insights and the old direct versus discovery breakdown when it moved from GMB Insights to the current reports. And views is now a deduplicated count of unique visitors. Google's definition: "The metric counts the number of unique visitors who viewed your profile... A person can only be counted once a day."

Should I compare my profile to an industry benchmark?

No, and this is the part of the monthly check where most owners get talked into a number that cannot be true.

There is exactly one large-n public study of Google profile performance. BrightLocal's Google My Business Insights Study covered 45,264 local businesses across 36 industries in the US, UK, Australia and Canada, using data from September 2017 to December 2018. Its medians per listing per month: 1,009 total searches, 1,260 total views, 59 total actions, 33 website clicks, and a view-to-action conversion rate of 4.68%.

Three problems, and the third disqualifies it.

BrightLocal sells local SEO software and drew the sample from its own audit tool users, businesses that already cared about local SEO. It skews high. It has never been refreshed, so the July 2019 edition is still current and the data window is eight years old. And several of the metrics it is built on no longer exist. Direct versus discovery is not reportable at all now, and "views" means something different than it did in 2018. That 4.68% cannot be reproduced on a modern account, so it cannot be a benchmark for yours.

The obvious replacements do not survive checking either. Birdeye's State of Google Business Profile publishes an action mix of website visits 47%, direction requests 38%, phone calls 15%, with no sample size and no methodology anywhere on the page. WebFX discloses only "proprietary industry benchmarks" from in-house data. Neither is citable with a number attached.

Benchmark your profile against its own trailing 90 days and against the named restaurants ranking above you. There is no published average worth the risk.

How do I know whether the check is working?

Compare a 90-day window against the previous 90-day window, in actions, using your own report.

Say you fix your hours, your ordering link and your review pace in March, and you pull the numbers in July:

• Calls: 128 in the prior 90 days, 151 in the following 90 days.
• Direction requests: 402, then 466.
• Website clicks: 291, then 356.
• Total actions: 821, then 973. A gain of 152 actions, or roughly 50 a month.

Now convert it with your own booking rate rather than a published one. If roughly one in ten profile actions becomes a table, that is 5 tables a month. At your own average check and party size from your POS, you have a dollar figure you can defend.

Two honest limits. This is a before and after comparison, not a controlled test, so season and weather sit inside your result. And Sterling Sky's field data shows clicks-to-call from Business Profiles declining even for profiles whose rankings held, based on 179 profiles across 34 US law firms tracked over two years. Flat calls in a declining channel is not the same as no progress. Read the actions together.

Rating movement belongs in the same picture. Uberall, which sells listings and reputation management, analyzed 64,000 business locations across four countries and associated a rise from 3.5 to 3.7 stars with 120% conversion growth, defining conversion as calls, direction requests and website clicks combined. Relative, not absolute, enterprise-skewed, and with no monthly volumes attached. Use it as a reason to work the reviews, not as a forecast.

Who should actually own this?

Not you, and this is where most of these checklists die.

An owner will do it for two months and then hit a bad week. So put it on a manager, on a fixed day, with the eight items written down and a place to record what was found. It should live wherever your other recurring procedures live, not in a note on someone's phone.

The reason procedures like this fail is not laziness. It is that the task is invisible when done and invisible when skipped, so nothing signals the miss. A check nobody can tell you skipped is a check that will get skipped. The fix is the same one that keeps any operating procedure alive past week two, which is the argument in why SOPs get ignored by week two.

Give it a date, a name, and a one line record. That is enough to survive a bad month.

What to do this week

Run the eight item check once yourself, slowly, and write down what you find. That first pass usually takes longer than fifteen minutes because there is a backlog.

Enter special hours for the next sixty days while you are in there.

Answer every review and message sitting open, including the old ones.

Then write the eight items into whatever system your managers already use, assign a day of the month, and put your name on the first three months so somebody is watching the handoff.

Then export your last 90 days of profile performance and save it. Next month's check is far more useful when you have a real baseline instead of a benchmark somebody published in 2019.

Questions about your monthly check?

Email me at eric@seod.com with your restaurant name and city. I will run the eight item check on your profile once, fill in your current values, and send it back as a filled in list so you have a starting baseline to check against next month. You keep the list either way.

I ran restaurants for sixteen years and I answer these myself. If your profile is already clean, the list will just say so, and that is a useful thing to know.

Or keep reading more on restaurant marketing.