Why Your Google Business Profile Is Not in the Top 3 (And What Moves the Map Pack Now)
Your Google Business Profile is done. You claimed it years ago, the hours are right, the photos are real, and there are a couple hundred reviews on it averaging better than four and a half stars.
And when someone a mile from your shop searches for the thing you do, the same three competitors show up in the map, you’re somewhere below them or not there at all, and the phone stays quiet.
If that’s you, your first guess is probably that you set something up wrong.
You probably didn’t.
You did what Google’s setup checklist told you to do, and the checklist is the price of admission, not the contest.
I’ve spent twenty-plus years looking at profiles in exactly that spot, and the fix is almost never a rebuild.
It’s a handful of fields and one page on your website that nobody told you decided the ranking.
A complete Google Business Profile gets you onto the map. It doesn’t get you into the top three. The Map Pack ranks businesses on a short list of signals the setup checklist never asks about: the exact primary category and the additional ones you never set, the page your profile links to, whether you’re open at the moment of the search, how recent your reviews are and what they say, whether anyone is defending the listing from edits and spam, and, in some industries, whether you’re paying for the row above the pack. Matching those signals to the search better than the three businesses above you is what moves you in.
That list hasn’t changed much in a decade.
What changed in 2026 is who’s reading it.
The same profile and the same page now feed the AI answers your customers have started asking, so every fix pays twice.
I’ll show you the numbers on that, then run the whole audit on a real business and client of ours, Gaetano’s Restaurant in Torrance, CA, then hand you six copy-and-paste prompts that let whichever AI assistant you already use run the same audit against the three competitors above you in about thirty minutes.


Key Takeaways:
- “Complete” is the entry fee, not the ranking. Google’s own help page names three things that decide local rank: relevance, distance, and prominence. The practitioners who test it have scored 187 individual factors, and seven of the top ten are fields on your profile or a page on your site that you can change today.
- The Map Pack still decides local buying searches, including in the AI era. In Whitespark’s 540-query study, the Map Pack appeared on 93% of local-intent searches and an AI Overview on 15%. In BrightLocal’s 200,085-search study of AI answers, Google Business Profile was the single most cited source, and 93% of the unique websites cited were individual business sites.
- Five things you were told to do don’t move the pack. Geo-tagged photos, keywords stuffed into your description and your review replies, posting volume, and a wider service area all sit near the bottom of the factor list. Some still help the person reading your profile. None of them moves your position.
- Six signals do: your primary and additional categories, the page your profile links to, your hours and services and pin, reviews that name the job and the place, defending the listing (including Google’s new four-day window on suggested edits), and knowing when to pay for Local Services Ads.
- The audit in this post is real. With their permission, I ran all six signals on a client’s profile, Gaetano’s Restaurant in Torrance, California: third in the pack from the middle of town, fifth two kilometers west, ninth for “restaurants torrance,” eighteenth for “catering torrance,” and fourth in ChatGPT’s answer, cited through a reservation site instead of its own. What we found and what we told them to fix is all here.
- Your website decides more of your Map Pack rank than you think. Google reads it to confirm your profile, and Google will overwrite your profile to match it. When the two disagree, the site usually wins.
- Run the audit yourself first. The six prompts at the end compare your categories, your connected page, your hours and services, your review language, your listing’s accuracy, and your Local Services Ads math against the three businesses above you. Fix the lowest score first.
Last verified: October 2026. The ranking-factor scores and survey figures in this guide are re-checked quarterly. See “What Changed in This Guide” at the end.
Why the Map Pack Still Decides, and Why AI Made It Worth More

There’s a stat going around that AI Overviews now appear on 68% of local business searches.
It’s true, and it’s the wrong number for you.
Whitespark, a local search software company, ran 540 queries across three US cities and six industries in 2025 and split them by intent.
On informational searches (“how much does a water heater cost”), an AI Overview showed 92% of the time.
On hybrid searches (“how much does a water heater cost in Torrance”), 97%.
On the searches where somebody is choosing a business to call (“water heater repair torrance,” “plumber near me”), an AI Overview appeared 15% of the time and the Map Pack appeared 93% of the time.
The 68% is an average of the three.
For the search that produces a phone call, the map is still most of the page.
Then there’s the other side of it.
BrightLocal, which sells local SEO and AI visibility tracking, ran 200,085 localized non-branded prompts across ChatGPT, Google’s AI Mode, and AI Overviews for 1,342 business locations this year and traced 1,967,762 citations back to their sources.
Google Business Profile was the single largest source of those citations at 28.5%, and 93% of the unique websites the AI tools cited were individual business sites, not directories.
The same study found Google Maps named a tracked business in 66% of searches while the AI tools named it 32% to 38% of the time, and that the AI tools barely agree with each other.
Google’s two AI surfaces share about 29% of the businesses they name, and ChatGPT overlaps with either one about 20% of the time, because ChatGPT leans on Bing and Yelp instead of Google’s profile data.
Both companies sell software that does better when you worry about this, so hold the exact percentages loosely.
The direction is hard to argue with, and it matches what Google’s own help page says it does: it builds your profile from what you enter, what it crawls on your website, licensed data, and what users contribute.
Read those two studies together and you have the point of this post.
The profile and the page it links to are now read by two systems.
Google’s local algorithm reads them to rank the pack.
Google’s AI reads them to answer the question.
ChatGPT reads your website and Yelp to answer the same question. There is no separate project called “AI visibility” for a business your size.
There’s one profile, one page, and one review stream, and fixing them for the pack fixes them for the rest.
I went deep on how a customer’s own AI assistant checks a business in How to Market Your Small Business When Personal AI Assistants Do the Shopping.
This post is the Google half of that.
One more number, because it tells you where the decision happens.
BrightLocal’s consumer survey this year found 45% of US consumers had asked an AI tool for a local business recommendation, up from 6% the year before.
Of the people who started a search in an AI tool, 18% felt ready to contact the business it recommended. 43% went and checked Google.
The AI builds the shortlist.
The pack is where your customer confirms it.
What Google Says Moves the Pack, and What the People Who Test It Say

Google’s help page on local ranking gives you three words: relevance, distance, and prominence.
It also says, flatly, that there’s no way to request or pay for a better local ranking.
Both things are true, and neither tells you what to do on Tuesday.
For that I use Whitespark’s Local Search Ranking Factors, which Darren Shaw has run since taking it over in 2017.
The 2026 edition had 47 local search practitioners score 187 factors on a zero-to-five scale for their effect on Map Pack rank.
It’s a survey of people who optimize profiles for a living and test what moves, not a leak of Google’s code, and Shaw says so plainly.
It’s also the best map we have.
Here’s the top ten for the pack:
- Primary Google Business Profile category
- Proximity of your address to the person searching
- Keywords in your business name
- A physical address in the city being searched
- Being open at the time of the search
- High numerical Google rating
- Address showing on the profile (not hidden as a service-area business)
- Additional categories
- Number of Google reviews with text
- Proper placement of the map pin
Look at what’s on that list.
Two of them you can’t change honestly: where you are, and where the searcher is.
One of them you shouldn’t change, because stuffing keywords into your business name violates Google’s guidelines and gets reported (more on that in Signal 5).
The other seven are a category field, a toggle, your hours, your pin, and your reviews.
None of them is a line item on an agency proposal.
That’s the gap this post closes.
Further down the same list, the factors that climbed the most since the last edition are worth knowing: “proper hours set” rose 64 places into the top 25, “predefined services” went from near the bottom to #22, and “keywords in your services list” climbed about 75 places into the mid-30s.
Those are three fields that rarely get touched after setup day.
I mapped every placement on a Google results page, from AI Overviews to rich snippets, in How to Show Up on Google.
This post stays on one of them, because for a local business it’s the one that rings the phone.
Five Things You Were Told to Do That Don’t Move the Pack

Before the six that do, the five that don’t, because you’ve probably spent time on at least two of them. These were on every checklist for years.
They were reasonable things to try.
The 2026 survey ranks them among the lowest of 187 factors for pack rank, so you can stop.
Geo-tagging your photos. Editing location data into image files before you upload them scored dead last among all 187 factors. Real photos of your work, your team, and your building still matter, because the human reading your profile uses them to decide, and photo quality ranks #11 among the conversion factors. Take the photos. Skip the geo-tagging.
Stuffing keywords into your business description and your review replies. Both sit in the myths list. Your description is read by customers, not by anything that moves your rank. Your replies to reviews matter a great deal, but for a different reason: 89% of consumers expect an owner to reply, and half are put off by replies that read like a template. Reply like a person. Don’t reply like a keyword list.
Posting more. The number of Google Posts you publish and the keywords inside them both rank as myths for pack position. Posts are useful for a customer who’s already looking at your profile and wants to see you’re alive. They aren’t a ranking lever, so if you’ve been posting three times a week out of obligation, you can post when you have something to say.
Drawing a bigger service area. The service-area setting tells Google where you’ll travel. It doesn’t make you rank there. Your pack rank in a neighboring city is governed by factors #2 and #4, which is to say by where your address is. A wider circle doesn’t change that.
Schema markup as a pack fix. Marking up your site with structured data is good for rich results and for AI search (it scores in the top 45 of the AI visibility factors), but it’s in the myths list for the pack itself. If someone sold you schema as the reason you’d rank in the map, that isn’t what it does.
And one thing that isn’t on the survey because it isn’t a factor, it’s a violation: creating extra profiles at addresses where you don’t operate.
You’ll see it pitched online as a way to own more of the map.
Google removed more than 13 million fake Business Profiles in 2025.
A suspended real profile costs more than any second pin ever earned.
None of that is a knock on you for doing it. It’s a list of hours you get back.
The Business I’ll Audit in This Guide: Gaetano’s Restaurant, Torrance

Instead of a made-up example, I’m going to run the whole audit on a real business and show you everything we found.
Gaetano’s Restaurant is a family-owned Italian restaurant on Pacific Coast Highway in Torrance, California.
Steve and Dori Giuliano opened it in 1993 and named it for Steve’s grandfather, who opened Giuliano’s Delicatessen in Gardena in 1952, and Vince Giuliano runs it today.
We host and maintain gaetanos.com, so some of the fixes below are ours to make and the rest are things that Vince and his team need to keep an eye on.
I picked them on purpose, because theirs isn’t a broken profile. It’s a good one, which is the situation this post is about.
Here’s the starting state, pulled on October 24, 2026, from the live profile, Google Maps, the website, and a local rank tracker:
- Claimed profile. Primary category “Italian restaurant,” with “Bar,” “Catering food and drink supplier,” “Family restaurant,” and “Restaurant” as additional categories.
- 4.6 stars from 1,161 Google reviews, about forty of them in the last three months, and the owner replies to most of them under his own name. 1,145 photos, the newest from four days ago. Zero Google Posts.
- Hours 11 to 9 Sunday through Thursday and 11 to 10 Friday and Saturday, matching the website. Reserve and Order buttons live. Attributes filled in down to the high chairs and the parking lot.
- The website button points to the homepage, over http instead of https. The homepage title tag is “Best Italian Food in the South Bay | Gaetano’s Restaurant” and its H1 is “Home.” The menus are images.
- For “italian restaurant torrance” searched from the center of Torrance, they’re third, behind Scardino’s (4.5 stars, 1,013 reviews) and PASTA e PASTA by Allegro (4.5 stars, 48 reviews). On a three-by-three grid of searches two kilometers apart centered on the restaurant, they’re first at the restaurant, in the top three at four of the nine points, and fifth at all three points to the west.
- For “restaurants torrance,” ninth. For “catering torrance,” eighteenth, with the catering category already on the profile.
- Asked for the best Italian restaurant in Torrance for a family dinner, ChatGPT named Scardino’s first and cited Scardino’s own website. It named Gaetano’s fourth and cited a reservation site’s roundup, not gaetanos.com.
So the question for Gaetano’s isn’t “why aren’t we in the top three.”
From the middle of town, they are.
It’s why the top three stops about 1.5 miles from the door, why it disappears on the broader search and the catering search, and why the AI answer gets to them last.
That’s the version of this problem a good profile has, and it’s the version I see most.
Where a step looks different for a law firm, a real estate team, or a home services company, I’ll show that too.
If you want to follow along with your own business, open your profile in one tab and your top three competitors’ listings in another.
Signal 1: Your Primary Category Is Most of the Game, and the Additional Ones Are the Tiebreaker

The primary category is the #1 factor on the list, and additional categories are #8.
Google uses the category to decide whether you’re even eligible for a search before it weighs anything else. P
ick a category that’s slightly off and you’re competing for the wrong pack.
Pick the right primary and skip the additional ones and you’re eligible for your main search and invisible for the rest.
Two things make this harder than it sounds.
First, the primary category is the one field where your competitors’ choice is public: it’s printed under the business name on every Maps listing.
Second, the category you’re hunting for may not exist. Google keeps a fixed list of a few thousand categories, and a category describes the business, not the job.
There’s no “emergency plumber” category, and in the US list there’s no water heater category either. The job goes in your services list, which is Signal 3.
Gaetano’s. The primary is “Italian restaurant,” and that’s right, because it’s the search that fills the dining room, and “Family restaurant,” “Bar,” and “Catering food and drink supplier” are each true and each defensible on the phone. This is the signal they’ve already done well, and the pack proves it: PASTA e PASTA sits second with one category and 48 reviews, so categories aren’t what’s holding Gaetano’s at third. The gap I’d still close is the one Scardino’s opened. Scardino’s runs seven: “Italian restaurant” plus “Bar & grill,” “Northern Italian restaurant,” “Southern Italian restaurant,” “Pizza restaurant,” “Wine bar,” and “Restaurant.” Gaetano’s has pizza on the menu and “pizza” in the summary Google writes from its reviews, so “Pizza restaurant” is a category they can defend. Private events are a real line of business, so an events category is worth checking for in the selector. The catering category stays even though they’re eighteenth for “catering torrance,” because the category is what makes them eligible for that pack and Signal 2 is what makes them prominent in it. Nothing else gets added. Seven categories isn’t the goal. Being eligible for every search they’d say yes to on the phone is.
Google’s own guidance is to use the fewest categories it takes to describe your core business.
Two or three additional categories you can defend in a sentence (“yes, we do catering”) beat eight you added for coverage.
A customer who calls for the thing your category promised and hears “we don’t do that” leaves the one-star review that costs you factor #6.
⚖️ Law firm: “Law firm” as a primary is the common miss. A personal injury practice wants “Personal injury attorney” as the primary and “Trial attorney” or “Law firm” as additional, because the person searching types the injury, not the word “firm.”
🏡 Real estate team: “Real estate agency” and “Real estate agent” are different categories with different packs. A team with a brokerage license and an office is an agency. An individual agent’s profile is an agent. Running both correctly is allowed. Running three agent profiles out of one desk isn’t.
🔧 Home services: “Plumber” as the primary, and “Drainage service” as an additional only if drain work is a line you market on its own. There’s no water heater category in the US list. That job goes in your services list by name, which is Signal 3.

#imTIPS: Pick your primary for the search worth the most revenue, not the most volume. Open Maps and search your top three services plus your city. Write down the primary category of each of the nine listings. You’ll usually find each pack is won by a different primary, which tells you exactly which search each competitor chose to own. Choose yours for the search where a new customer is worth the most to you, and cover the others with additional categories. If a competitor’s business name doesn’t match the name on its trucks or its license, hold that thought for Signal 5.
What you have now: a primary category chosen on purpose, two or three additional categories you can defend, and a list of the categories each competitor picked for each of your top searches.
Signal 2: The Page Your Profile Links To

This signal isn’t on the profile at all, which is why it gets missed.
Five of the top 25 pack factors are about the page your profile’s website button points to: your name, address, and phone matching the profile exactly in the page’s text (#15), keywords in that page’s title tag (#17), city and neighborhood relevance in the content (#18), keywords in the page’s headings (#20), and a dedicated page for each service (#24).
On the AI side, a dedicated page per service is the #2 factor for AI visibility and geographic relevance is #6.
And remember the BrightLocal finding: 93% of the unique sites AI tools cited were business websites.
This page does double duty.
It also does something quieter.
Google says it builds your profile partly from crawled web content, including your official website.
That means when your profile says one thing and your site says another, Google has a tiebreaker, and it’s usually your site.
I’ll come back to that in Signal 5, because it’s how profiles get edited without you noticing.
Gaetano’s. The profile links to the homepage, which is the right page for a single-location business, because that’s the page with the most links pointing at it. The homepage is also where most of this audit’s work turned out to be. What we found, and what we’re fixing, since the site is ours to maintain:
- The link itself. The website button points at
http://www.gaetanos.com/, which redirects once to the https version. One hop is minor. It’s also a thirty-second fix, so it goes first. - The title and the heading. The title tag reads “Best Italian Food in the South Bay | Gaetano’s Restaurant.” “South Bay” is how people here talk, and it stays. “Torrance” is the word in every search we measured, and it isn’t there. The H1 is “Home,” which is the theme printing the page’s name. The new title: “Italian Restaurant in Torrance, CA | Gaetano’s, Family-Owned Since 1993.” The new H1: “Family-Owned Italian in Torrance Since 1993.” Name, address, and phone are already in live text in the footer, with one small mismatch worth fixing while we’re in there: the site spells out “California” where the profile says “CA.” Character for character is the standard, because Google compares the two.
- The menus. The profile’s Menu button points to the menu page on the site, and every menu on it is a picture. The dinner menu is a single one-megabyte image whose alt text is “Dinner Menu.” Meanwhile Google’s own summary of the reviews names fresh pasta, pizza, lasagna, and chicken parmesan, and the review topics it extracted include butternut squash ravioli, pear and gorgonzola salad, short rib pasta, and eggplant parmigiana. Not one of those dishes appears as text anywhere on the site. The menus become real HTML with prices (the online ordering system already has every item as text), so Google and the assistants can read the dishes people are already searching for and reviewing.
- The structured data. The site’s schema still identifies the business under its old domain, gaetanosonline.com, and types it as an Organization rather than a Restaurant. We’re rewriting it as a Restaurant with the cuisine, the price range, the menu URL, the reservation link, the hours, and the founding year, so the machine-readable version of the site says what the profile says.
- The catering page. They’re eighteenth for “catering torrance” with the right category on the profile, which tells you the page isn’t carrying its share. The catering page lives at
/menu/catering/, runs 358 words, and names Torrance twice, and the address/catering/returns a 404. It gets its own page at/catering/, with the old address redirected to it, the delivery area, the occasions, photos of the trays, and the most recent reviews that mention catering. Private events gets the same treatment. Those two pages are what the profile’s catering category and the AI answers will have to cite.
That’s about two days of work for us, because we maintain the site.
On a smaller site it’s about a day, and it’s the most durable day in this guide.
If the site itself is the thing holding this up for you, the AI-Ready Website Playbook covers what those pages need to do for people, for Google, and for assistants, and the conversion leak analysis shows you where the page loses the call once someone lands on it.
⚖️ Law firm: One page per practice area per office, with the city in the title. A firm with “Practice Areas” as a single page is asking Google to guess.
🏡 Real estate team: Neighborhood and community pages are the service pages, and the profile should link to the team’s own page, not the brokerage’s national homepage.
🔧 Home services: One page per service you want to rank for (water heaters, drains, repipes), each with the city in the title and the first heading, a starting price where you’ll stand behind one, and the three most recent reviews that mention that job. The profile points at the homepage, which names the trade and the city in both.

#imTIPS: Count the redirects on your own website button. Open your profile in a private browser window, tap the website button, and watch the address bar. If it changes more than once before it settles, your link is riding a redirect chain, and after two or three redesigns some of your best old links are barely passing anything. Paste the final URL into the profile and point every old URL straight at its final page. While you’re there, ignore anyone who tells you the title tag has to be 60 characters. The survey ranks title-tag length near the very bottom of the list. Put the words that earn the click first and the extra service terms after.
What you have now: a profile that links straight to a page whose title, heading, and text name your service and your city, with your name, address, and phone matching the profile exactly, and one page per service worth ranking for.
Signal 3: Hours, Services, and the Pin

Three mechanical signals, all inside the profile, all climbing.
Being open at the time of the search is #5.
Proper placement of the map pin is #10.
Properly set hours is #21, up 64 places from the last edition.
Google’s predefined services list is #22, and keywords in your services list climbed about 75 places.
The open-at-time-of-search signal deserves a plain explanation because it’s new and it’s blunt.
When your profile says you’re closed, Google would rather show someone who’s open.
You can watch it happen: a business sitting first in the afternoon starts dropping about an hour before its posted closing time and, once it’s closed, is displaced by competitors whose hours run later.
Shaw describes exactly that behavior in the survey notes, along with the warning that matters: don’t set hours you don’t keep for a business where people show up at the door.
A locked storefront with “Open” on the profile earns the review that costs you factor #6 and violates Google’s guidelines besides.
If a real person answers your phone around the clock and you’ll actually send someone, 24-hour hours are honest. If the answering service takes a message for the morning, they aren’t.
Hours are also the thing customers most often can’t find.

63% of consumers said they’d stopped being a customer of a local business because they couldn’t find an answer they needed online, and among the answers they went looking for, current or holiday hours was second only to pricing.
Source: SOCi 2026 Local Discovery Index
SOCi sells software to multi-location brands, so take the exact percentage as directional.
The direction is that wrong hours cost you twice: once in the pack, once at the door.
Gaetano’s. This is the signal where the audit came back mostly clean, and it’s worth seeing what clean looks like. The hours on the profile match the hours on the website and on Yelp, so a stranger’s suggested edit has nothing to lean on (that matters in Signal 5). The attributes are filled in: outdoor seating, delivery, takeout, dine-in, happy hour, kids’ menu, high chairs, reservations recommended, free parking lot. The Reserve button goes to OpenTable and the Order button to the online ordering system, which for a restaurant is the services list. The pin sits on the building inside the shopping center at Pacific Coast Highway and Crenshaw, not out on the highway. Three things we’re checking with Vince:
- “Brunch” shows under dining options. The restaurant opens at 11. If nobody’s serving brunch, that attribute came from a visitor and it comes off, because an attribute that promises something you don’t do earns the review that costs you factor #6.
- The open-at-time-of-search window is theirs to win on weekdays. Scardino’s keeps the same hours Gaetano’s does. PASTA e PASTA, the second listing, is closed Monday and Tuesday and closes between lunch and dinner the rest of the week. So between 2 and 5 p.m. most days, and all day Monday and Tuesday, the listing above Gaetano’s is closed, and the open-now signal should favor the one that’s open, as long as the hours it claims are honest. If the bar stays open after the kitchen closes, “More hours” lets you publish kitchen hours and bar hours separately, and the kitchen hours are the ones to keep conservative.
- Holiday hours through the end of the year, set now, because December is when a “closed” that should have read “open” costs a restaurant the most.
Address matters here too, in two ways you can’t shortcut.
A physical address in the city being searched is #4; a Torrance address doesn’t rank in Redondo Beach’s pack no matter what the service-area circle says, and you can watch distance work inside one city too.
About a mile and a half west of Gaetano’s, on the same search, they drop to fifth and a different Italian restaurant is first at all three western points.
And “address showing on the profile” is #7, which means the profiles that hide their address (service-area businesses working from home) tend to lose ground.
Shaw suspects that one is partly a bug where Google loses the pin and re-places it somewhere wrong, and I think he’s right, but the practical advice is the same: if you’re allowed to show your address, show it, and if you aren’t, check where Google put your pin after you hid it.
⚖️ Law firm: A 24/7 intake line is honest hours if a human answers and can set a consultation. Services are practice areas and specific case types, in plain words (“car accident lawyer,” not “motor vehicle torts”).
🏡 Real estate team: “Hours by appointment” plus the hours someone really answers the phone. Services are the things you do, not the neighborhoods.
🔧 Home services: If a dispatcher answers at 2 a.m. and a tech rolls, 24-hour hours are honest. If the answering service takes a message for the morning, they aren’t. List every service you sell in the words a homeowner types (“tankless water heater installation,” “sewer camera inspection”), and zoom in on your pin. Profiles set up from a phone often have it sitting on the street centerline instead of the building.

#imTIPS: Search your main term from your phone at 7:30 tonight. If you vanish and a competitor you don’t recognize appears, you just watched the open-at-time-of-search signal reorder the pack. Then decide honestly whether your phone is actually answered then. If it is, your hours are wrong on the profile. If it isn’t, the fix is a person or a service that answers, and that’s a staffing decision, not an SEO one.
What you have now: hours that match when a human really answers, a services list that covers everything you sell in the words customers use, and a pin on your building.
Signal 4: Reviews That Name the Job and the Place

Reviews are four of the top fourteen factors: your numerical rating (#6), the number of reviews with text (#9), how recent they are (#11), and whether they arrive steadily instead of in bursts (#14).
Keywords in the review text are #36. On the conversion side, rating is the #1 factor and positive sentiment in the text is #2.
And in BrightLocal’s 1,002-person consumer survey this year, 74% said they only pay attention to reviews from the last three months, 47% won’t use a business with fewer than 20 reviews, and 82% read the AI-generated summaries Google now writes above your reviews.
That last number is the one I want you to sit with.
The summary at the top of your reviews is written by software reading your reviews.
If two hundred of them say “great service,” the summary says customers praise the friendly service.
If forty of them say “butternut squash ravioli on the patio,” the summary says that, and so does the AI answer when someone asks for Italian with a patio in Torrance.
Google also pulls those phrases out as “justifications” under your pack listing.
The words in your reviews are the words you rank for and the words you get summarized by.
Gaetano’s. This is the signal they’d win a trophy for, and I want to show you what good looks like before I show you the two gaps. 1,161 reviews at 4.6. About forty in the last ninety days, so roughly thirteen a month, which keeps the “last three months” reader satisfied without anyone running a campaign. Vince replies in his own voice, signs “Vince,” names the server, and names the dish (“Grandma Dot’s meatballs”), which is exactly what a reply is for. The summary Google writes reads: “Diners like this restaurant’s delicious Italian food, including fresh pasta, pizza, and generous portions of lasagna and chicken parmesan. They also highlight the reasonable prices for both food and wine, and the warm, inviting atmosphere.” The topics it pulled out: patio (68 reviews), butternut squash ravioli, pear and gorgonzola salad, bread pudding, happy hour, eggplant parmigiana, spumoni, short rib pasta. The reviews are already telling the ranking system what the restaurant is known for. Two gaps:
- The one-star from two months ago has no reply. A couple celebrating their 47th anniversary, a lukewarm chicken parm, and a server who didn’t come back. It’s the review the next person reads, because it’s the newest low one. Several star-only ratings from the same stretch are unanswered too. The one-star gets a reply in Vince’s voice: sorry, here’s what changed, come back and let us make it right. The star-only ratings get one line each, because 89% of consumers expect a reply and half are put off by one that reads like a template. The lower ratings also repeat two things, slow service on a Monday and the air conditioning, and the summary software will surface a complaint the moment it arrives often enough to count.
- Not one topic is a place. Sixty-eight reviews mention the patio. Almost none mention Torrance, the South Bay, or a neighborhood, and “anniversary” doesn’t come up often enough to make the list even though the one-star was about one. So the summary can’t connect the restaurant to a place or an occasion, and neither can the AI answer for “anniversary dinner in Torrance.” The ask changes, not the script: a line on the check and from the server, “a quick Google review helps the next family in the South Bay find us.” Google’s policy draws the line at telling people what to write. Telling them who the review helps isn’t over it.
And one thing to keep doing.
OpenTable shows 4.8 stars from 2,735 reviews and Yelp 4.4 from about 1,650, and when ChatGPT named Gaetano’s, it cited an OpenTable list, not gaetanos.com.
ChatGPT barely reads Google’s profile data; it reads Yelp and Bing.
The reviews on third-party sites are doing the AI work right now, and the authority of the sites your reviews sit on is the #5 AI visibility factor.
Keep them healthy while the website catches up.
If your reviews look nothing like that, old and generic, the fix is mechanical.
Ask within the hour, from the person who did the work, with one line that invites a review without saying what to write. Expect about one in ten, and take four a week every week over forty in one week, because Google and your customers both read a sudden pile of reviews as a purchase.
Reply to every one in your own words.
For the deeper version, including timing and wording that stays inside Google’s rules, we wrote How to Get More Reviews.
⚖️ Law firm: Ask at the close of the matter, from the attorney, and never in a way that invites case details. “They handled my case in Torrance” is enough. Your bar’s rules on testimonials come before anything in this section.
🏡 Real estate team: Ask at closing, in person, then text the link that night. Happy buyers name the neighborhood on their own.
🔧 Home services: The tech texts the review link from his own phone before he’s out of the driveway: “Thanks for having us out today. If you’ve got a minute, a quick Google review helps the next homeowner in your area find us.” Customers prompted that way tend to name the neighborhood on their own.

#imTIPS: Read your own AI review summary once a month. It’s at the top of your reviews on Maps. It’s a plain-language report of what your customers are telling the ranking system about you, and it tells you which job, which neighborhood, and which complaint are showing up often enough to be summarized. If the summary doesn’t mention the service you want to be known for, you now know what to invite customers to talk about.
What you have now: a review request that goes out within the hour from the person who did the work, a realistic weekly target, replies in your own words, and a monthly read of the summary software writes about you.
Signal 5: Defending the Listing From Edits, Blue Text, and Names That Aren’t Real

Your profile isn’t a brochure you published once.
It’s a public record other people can edit, and Google’s policy on that got sharper this year.
Here’s the rule, from Google’s help page, as of September 2026: when a user suggests an edit to your profile, Google may notify you, and “once notified, you have 4 days to accept or reject the suggestion.
If you don’t respond within 4 days, Google may automatically publish the update if it is supported by other publicly available information, such as your business website.”
Changes Google has made on its own show in blue text in your profile editor.
In 2025 alone, Google processed more than 80 million suggested updates to hours and contact details, and in April 2026 it said it was starting to email verified, active owners before important edits go live.
That alert only helps if it reaches a person who reads it.
Read the conditional clause again.
Google publishes the suggested edit when your website backs it up.
That’s Signal 2 coming back around: if your site still shows the old hours, the old phone number, or the old suite, a stranger’s suggestion that matches your stale site becomes your profile in four days, and you find out when the phone stops ringing on Saturdays.
The other half of defending the listing is the competitors who aren’t playing straight.
Keywords in the business name is the #3 pack factor, which is exactly why you’ll see listings named something like “24-7 Emergency Plumber [City]” when the truck says something else.
Google’s guidelines say the name on the profile must be the real-world name on your signage and your stationery, and that adding keywords, cities, or taglines is a violation.
Reporting it works.
“Removal of spam listings through spam fighting” is the #19 pack factor, because when a padded name gets corrected, its rank drops and yours rises at no cost to you.
Gaetano’s. The profile and the website agree on the name, the address, the phone, and the hours, so a suggested edit has nothing to lean on. Four things came up anyway:
- Google’s own description is wrong. The one-line description Google writes under the buttons says “Family-owned spot since 1999.” The restaurant opened in 1993. That line isn’t editable, but it’s built from what Google finds on the web, and the profile has no owner-written description competing with it. Vince writes the “From the business” description with 1993 in the first sentence, the homepage heading says 1993 (Signal 2), and the structured data carries the founding year. Google rewrites its line when the sources agree.
- A stale copy of the hours is out there. A listings aggregator’s directory page shows Gaetano’s open 9 to 6 Monday through Friday and closed on weekends. That’s a stranger’s suggested edit waiting to happen, and “other publicly available information” is exactly what the four-day rule lets Google lean on. The copies get corrected. That’s directory cleanup, not profile work, and a listings service does it in an afternoon.
- Two names for one business. The profile says “Gaetano’s Restaurant.” The website’s structured data says “Gaetano’s Italian Restaurant,” left over from the old domain. Google compares the two. The fix rides along with the schema rewrite in Signal 2, and the lesson is that your name should appear one way everywhere, matching the sign.
- Nothing to report. The two listings above them are real businesses under their real names, so there’s no business redressal complaint to file this time. For “catering torrance,” the first listing is named “Torrance Catering,” with 16 reviews. A business can legitimately be named after its city, so before anyone files anything, the check is whether that’s the name on its license and its van. If it is, it stays. If it isn’t, it’s a ten-minute form, and it’s the only line in this guide that improves your rank by changing someone else’s profile.
And the thing that didn’t come up: zero Google Posts, third in the pack.
If posting moved the pack, the business with none wouldn’t be sitting there. Notifications go on and route to Vince’s phone, so the next suggested edit is a four-day decision instead of a surprise.
⚖️ Law firm: “[City] Injury Lawyers” as a profile name for a firm whose sign says something else is the same violation, and legal is one of the most heavily spammed categories on the map. Report it with the bar listing attached.
🏡 Real estate team: The common problem is duplicates, not spam: an old agent profile, a team profile, and a brokerage profile at the same address, splitting reviews three ways. Merge or close the ones that aren’t the business.
🔧 Home services: Keyword-stuffed names are thickest here, and the evidence is usually in the listing’s own photos: the van says one name, the profile says another. Report it with the state license lookup attached.

#imTIPS: Give the profile an owner and a ten-minute Monday. One named person, business-info notifications on, and a standing ten minutes every week to open the editor, look for blue text, check the suggested-edit queue, and search your top three terms to see who’s new in the pack. Four days goes by fast, and the person who notices a bad edit on day five is the one explaining to you why the Saturday calls stopped.
What you have now: notifications routed to a person who checks them, a website that agrees with the profile so suggested edits can’t overrule you, and a report filed on any competitor above you whose name isn’t their name.
Signal 6: When to Pay for the Row Above the Pack

In home services, legal, and a growing list of other categories, there’s a row above the Map Pack: Local Services Ads, the listings with the green check that say “Google Guaranteed” or “Google Screened.”
You pay per lead (a call or a message), not per click.
You don’t pick keywords; Google decides when you show based on your service category and location.
Shaw’s survey ranked the factors that move Local Services Ads placement: budget and bidding first, reviews second (with recency mattering as much as count), service selection third, response time fourth, verification fifth.
He also notes something that should get your attention if you’re in a competitive category: personal injury clients who rank well in both the pack and organic still bring in up to 70% of their cases through Local Services Ads.
That’s one agency’s observation, not an industry number, but it’s the reason the row exists.
Whether it’s worth it for you is arithmetic, not opinion, and it’s the same arithmetic I walked through in How to Build a Small Business Marketing Plan From the Customers You Actually Need: what a new customer is worth in the first year, what share of that you can spend to get one, and what that means a lead can cost.
Gaetano’s. Restaurants aren’t eligible for Local Services Ads, so the row above the pack for Gaetano’s is regular Google Ads plus the reservation button Google already shows. The arithmetic is the same; only the lead source changes. For dinner searches, I’d spend nothing. They’re already in the pack from the center of town, and paying for a search you already win is the expensive version of this mistake. Catering is different. A catering order is worth many dinners for two, the organic position is eighteenth, and the page work in Signal 2 takes a quarter or two to pay. So the paid test worth running is a capped Google Ads campaign on “catering torrance” and the neighboring cities, pointed at the new catering page, run during the hours someone answers the catering line, with a ceiling worked out from what a catering order is worth to them and how many inquiries turn into orders. Those numbers are Vince’s, and not mine to publish. The decision rule is the one in the last prompt at the end of this post: what a customer is worth, the share you’ll spend to get one, your close rate, your ceiling.
Three rules we hold clients to on Local Services Ads, in order of how often they get skipped.
Answer live, every time, because a missed call is a paid lead you lost and a ranking signal against you. Dispute every lead that’s spam, outside your area, or for a service you don’t offer, because Google credits them back, but only if you flag them.
And never let the ads replace the pack work in this guide, because the row above the pack rents your spot and the pack earns it.
⚖️ Law firm: Local Services Ads are often the most expensive and most productive lead source a practice has. The math still applies; the first-year value is just larger.
🏡 Real estate team: “Google Screened” exists for agents in many markets. Response time matters more here than anywhere, because the lead is also calling two other agents.
🔧 Home services: The arithmetic with real numbers. A new plumbing customer worth about $1,100 in the first year, a quarter of that to acquire one, and a 45% booking rate on answered calls puts your ceiling near $125 a lead. If Google’s estimated cost per lead clears it, run the ads only during the hours a human answers, because Local Services Ads can charge for a call that goes to voicemail.

#imTIPS: Set your ad hours by when you can answer, not by the week. If your phone is covered from 7 a.m. to 7 p.m., run the ads then and only then. A lead at 10 p.m. that goes to voicemail costs you the lead fee, the customer, and a point against your response time. Turning the ads off when nobody’s there is the cheapest optimization in the account.
What you have now: a cost-per-lead ceiling you worked out from your own numbers, a decision to run Local Services Ads or not that you can defend, and, if you run them, ad hours that match your phone coverage.
The Completed Audit, Side by Side

Here’s the Gaetano’s audit on one screen, what we found and what we recommended, so you can see the whole thing before you run your own.
| Signal | What we found | What we recommended |
|---|---|---|
| Categories | Italian restaurant, plus Bar, Catering food and drink supplier, Family restaurant, Restaurant. The listing above them has one category; the listing at the top has seven | Keep all five. Add Pizza restaurant (pizza is on the menu and in Google’s review summary). Check the selector for an events category to cover private events |
| Connected page | Website button on http; title names the South Bay but not Torrance; H1 “Home”; address spells out “California” where the profile says “CA”; menus are images; schema still under the old domain and typed Organization; /catering/ returns a 404; eighteenth for “catering torrance” | https link; new title and H1 with Torrance and 1993; address formatted as on the profile; HTML menus with prices; Restaurant schema with menu, reservations, hours, founding year; dedicated catering and private-events pages with the old address redirected |
| Hours, services, pin | Hours match the site and Yelp; attributes complete; Reserve and Order buttons live; pin on the building; “Brunch” attribute unconfirmed | Confirm or remove Brunch; publish kitchen and bar hours separately if they differ; holiday hours through year end; watch the 2 to 5 p.m. and Monday and Tuesday windows when the second listing is closed |
| Reviews | 4.6 from 1,161; about 40 in 90 days; owner replies by name; summary names the dishes; one-star from two months ago and several star-only ratings unanswered; no place or occasion among the topics | Reply to the one-star and the star-only ratings; change the ask to “helps the next family in the South Bay find us”; keep OpenTable and Yelp healthy, because that’s where ChatGPT found them |
| Defending the listing | Profile and site agree; Google’s own description says “since 1999” (it’s 1993); an aggregator shows 9 to 6 weekday hours; two versions of the name; zero Posts and still third; no stuffed names above them | Owner description with 1993; directory cleanup; one name everywhere; notifications on; nothing to report this time |
| Paying for the row above | Not eligible for Local Services Ads; already in the pack for dinner searches from the center of town; eighteenth for catering | No paid spend on dinner searches; a capped Google Ads test on catering pointed at the new page, with a ceiling from their own order value and close rate |
Total owner time: about two hours in the profile and the reply queue, then ten minutes a week. The site work is ours, about two days.
What to expect, and when.
In our experience the category, hours, and pin changes show up in the pack within days to a few weeks, because they change what you’re eligible for.
The page work and the review stream take a quarter or two to move the needle, because they change how prominent you are, and prominence is earned.
We pulled Gaetano’s pack positions, the nine-point grid, and the review counts on October 24, 2026.
We’ll run the same grid again in ninety days and add what moved to the changelog at the bottom of this post.
Do the same for yourself: write down the date you made the changes and your position for your top three searches from three spots (your office, two miles out, five miles out), so when it moves you know what moved it.
Six Prompts to Run the Same Audit on Your Business

You don’t need me to run this.
You need thirty minutes, your profile open in one tab, your top three competitors open in another, and whichever AI assistant you already use.
Each prompt below is one line, so you can copy it as-is and paste your details into the brackets.
They don’t name a tool because they work in any of them.
Prompt 1: The category gap.
Here are the three businesses in the Google Maps results when I search "[your main service] [your city]": [names]. For each one, find the primary category shown under its name on its Google listing and the services it lists. Then compare them with mine (primary category: [yours]; additional categories: [yours]; services: [yours]) and tell me which category or service gaps most likely explain why they're eligible for that search and I'm not, and which changes I can defend as true about my business.
What you’ll learn: which search each competitor chose to own, and whether your empty additional-category list is the reason you’re missing from a pack you should be in.
Prompt 2: The connected page.
Read the page my Google Business Profile links to, [URL], and tell me whether its title tag, first heading, and first paragraph name the exact service and city a customer would type, whether my business name, address, and phone number appear in text exactly as "[name, address, phone as shown on your profile]", how many redirects the URL passes through, and the first three changes you'd make so this page is the obvious match for "[service] in [city]".
What you’ll learn: whether the page Google uses to confirm your profile actually confirms it, and whether your website button is riding a redirect chain.
Prompt 3: Hours and services.
Here are the hours on my Google Business Profile: [hours]. Here are the hours a real person actually answers my phone: [hours]. Here are the services listed on my profile: [list]. Here are the services I actually sell: [list]. Tell me where the two sets of hours disagree, which services I sell that are missing from the profile, how a customer would word each missing service in a search, and whether any listed hours would mislead someone who showed up at my door.
What you’ll learn: every after-hours search you’ve been handing to a competitor, and the services the profile has never told Google you sell.
Prompt 4: What your reviews are teaching the ranking system.
Here are my twenty most recent Google reviews: [paste]. Count how many name a specific service and how many name my city or neighborhood, tell me the dates of the newest five, summarize what a stranger would conclude I'm known for, and write one plain two-sentence text message I can send a customer after a job that invites a review without telling them what to say.
What you’ll learn: the summary software is already writing about you, how stale your reviews look to someone who only trusts the last three months, and a request you can send today.
Prompt 5: The listing that isn’t yours anymore.
Here is how my business currently appears on Google Maps: [name, categories, address, hours, phone, website]. Here is how it appears on my website: [paste the same fields]. List every field that doesn't match character for character. Then, for the three businesses above me for "[service] [city]", [names], tell me which ones use a business name on Google that includes words that don't appear to be part of their legal or signage name, and draft a short factual note I could submit with a redressal complaint.
What you’ll learn: the mismatches a stranger’s suggested edit could turn into your profile in four days, and which competitor above you is ranking on a name that isn’t theirs.
Prompt 6: The paid-row decision.
I run a [type of business] in [city]. A new customer is worth about $[X] to me in the first year, I'm willing to spend about [X]% of that to acquire one, I book about [X]% of the inquiries I answer, and a human answers my phone from [start time] to [end time]. Tell me whether my category is eligible for Google Local Services Ads, work out the most I can pay per lead (through Local Services Ads if eligible, through Google Ads if not) and still come out ahead, tell me what I'd need to fix before turning ads on, and tell me which ad hours I should run.
What you’ll learn: your cost-per-lead ceiling in your own numbers, so Google’s estimate becomes a yes or a no instead of a guess, whichever paid row your industry gets.
Run all six, score yourself honestly, and fix the lowest one first. Then put a note in your calendar to run them again in ninety days.
Four Numbers to Check Monthly
Once the six signals are fixed, this doesn’t need an hour a week.
It needs four numbers once a month, the same way I set it up in How to Measure Marketing ROI in Just 10 Minutes a Month.
- Calls, direction requests, and website clicks from the profile. The Performance tab inside your Business Profile reports all three. If the fixes worked, calls move first.
- Your pack position for your top three searches from three spots. Search from your office, from two miles out, and from five miles out, on your phone, during your busiest hour. A local rank-grid tool automates this if you’d rather, but three manual searches tell you most of what you need.
- Reviews added in the last thirty days, and the date of the newest one. Whatever weekly target you set, if the newest one is older than two weeks, the request isn’t going out.
- Cost per booked job from Local Services Ads or Google Ads, if you run them. Leads divided into jobs, then dollars divided into jobs. If it’s drifting above your ceiling, dispute the junk leads first and tighten the ad hours second.
Put the four in the scoreboard from your marketing plan if you built one.
If you didn’t, a note on your phone with the date and four numbers is enough.
The point is the trend, not the spreadsheet.
Who Does the Work: You, Your Team, or a Partner
Everything in Signals 1, 3, 4, and 5 can be done by you or whoever answers your phone, in a couple of hours, and I’d rather you did it yourself than paid anyone to, because you’ll understand your own profile afterward in a way no report gives you.
Signal 2 needs whoever maintains your website for a day. Signal 6 needs your calculator.
Where it makes sense to bring us in is the system around it: the service pages built to be cited and to convert, the review request wired into your job-completion workflow so it goes out without anyone remembering, the alerts and the Monday check assigned to someone whose job it is, and the Local Services Ads account built and disputed properly.
That’s the local search and website work inside our AI marketing systems, and it’s what our Local Search Marketing and AI-Ready Optimization services exist for.
We can set it up and hand it back to you, run it with you, or give you an hour of guidance and let you keep the controls.
If you’re deciding which, we wrote an honest version of that choice in Should You DIY Your Marketing With AI or Hire an Agency?.
If you run the six prompts and want a second set of eyes on the results, book thirty minutes with me and bring your lowest score.
Frequently Asked Questions About Google Business Profile and the Map Pack
Usually because of one of six things: your primary or additional categories don’t match the search, the page your profile links to doesn’t name your service and city, your hours say you’re closed when the person is searching, your reviews are old or generic, a suggested edit or a competitor’s spam has changed what Google shows, or you’re in a category where Local Services Ads sit above the pack. A complete profile makes you eligible. Those six signals decide the order.
Category, hours, and pin corrections usually show within days to a few weeks because they change which searches you’re eligible for. Page improvements and a steady review stream usually take one to two quarters, because they build prominence, which Google grants over time. Proximity never changes: a business in a neighboring city won’t rank in your pack no matter what else you fix.
Not in the Map Pack. In the 2026 Local Search Ranking Factors survey, the number of Google Posts and the keywords in them both rank among the lowest of 187 factors for pack position, and the restaurant audited in this post sits third in its pack with zero posts. Posts help a customer who’s already looking at your profile decide to call. Post when you have something worth saying, not on a schedule for ranking’s sake.
Google’s AI tools lean on it heavily: in BrightLocal’s 2026 study of 200,085 local AI searches, Google Business Profile was the top citation source at 28.5% of 1.97 million citations. ChatGPT barely uses it and leans on Bing and Yelp instead. Across all the tools, 93% of the unique websites cited were individual business sites, so your profile and the page it links to feed both.
If Google’s guidelines require you to hide it (customers don’t come to your location), hide it, then check where Google placed your pin afterward, because profiles that hide the address sometimes lose rank when the pin moves. If customers do come to your location, show the address. “Address showing on the profile” is the #7 Map Pack factor.
For “call now” services like plumbing, HVAC, electrical, locksmiths, and many legal practices, often yes, if a person answers the phone live and you dispute junk leads. Work out your ceiling first: what a new customer is worth in year one, the share you’ll spend to get one, and your booking rate on calls. If Google’s estimated cost per lead is under that ceiling, test with a capped budget during the hours your phone is covered. Restaurants and most retail aren’t eligible.
Keeping This Post Current
The ranking-factor scores in this guide come from Whitespark’s 2026 Local Search Ranking Factors, the consumer numbers from BrightLocal’s 2026 surveys and studies, and the policy rules from Google’s own help pages.
The Gaetano’s positions, counts, and competitor details were pulled on October 24, 2026, and get re-run with each check.
We re-check all of it quarterly, and sooner if Google changes the rules on edits, verification, or reviews.
Changes get logged below with a date.
If something here is wrong or out of date, I’d rather hear it from you than have you find out in the pack: reply here.
Last verified: October 2026.
What Changed in This Guide
- 2026-10-28: First version. Factor ranks from Whitespark’s 2026 edition (published November 2025). BrightLocal’s 200,085-search AI visibility study (September 2026) and 2026 Local Consumer Review Survey. Google’s four-day suggested-edit rule as published September 2026. Gaetano’s Restaurant audit pulled October 24, 2026 (profile, pack positions, nine-point grid, reviews, competitor listings, ChatGPT check). Prompt set v1: six audit prompts.
The Top Three Were Never About the Profile Being Complete
I’ll end where I started.
Your profile isn’t in the top three because it’s incomplete.
It’s not in the top three because a complete profile is where everyone starts, and the order after that is decided by a category field, a page on your website, your hours, your reviews, whether anyone’s watching the listing, and in some industries whether you pay for the row above it.
Those were the signals before AI started reading the same profile, and they’re the signals now.
AI just made the same fixes worth more.
Gaetano’s is in the top three from the middle of Torrance with zero posts, no tricks, and a thirty-three-year-old reputation, because the fundamentals are done.
The rest of its map is a page and a few fields away.
Run the six prompts this week. Fix your lowest score first.
Then check four numbers once a month and let the rest compound.
The three businesses above you did some version of this, probably without ever seeing the factor list.
You can do it on purpose.

Want a second set of eyes on your audit?
Run the six prompts on your own profile, or bring your lowest score to a 30-minute review and we’ll look at what’s keeping you out of the top three and what to fix first.


