What your Google listing looks like to a machine, not to you
An AI assistant named three companies for a job in your town and skipped you, even though your reviews are better. The reason is in the half of your Google profile a person never reads: the machine does.
By Stacey Tallitsch | October 22, 2026
A homeowner two streets over had a problem last week. Water heater, breaker panel, a roof leak, it does not matter which. She did not open Google and scroll. She picked up her phone, asked the assistant who she should call, and it gave her three names. Yours was not one of them. You have been in that town 15 years. You have more reviews than at least one of the companies it named, and better ones. You pulled up your own profile to see what went wrong and it looked fine to you. Good photos. Five stars. Hours correct. So you closed it, annoyed, and decided the machine was broken.
The machine was not broken. You were reading the half of your profile a person sees. The machine was reading the other half.
Your listing has two faces
Every Google Business Profile is two things at once. One is the page a human lands on: the photos, the star rating, the posts, the little tour of your work. You have spent time there because it is the part you can see, and it is the part that feels like marketing. The other face is a block of structured fields a person barely registers and a machine reads first. Your primary category. Your additional categories. Your defined service area. Your attributes. The exact spelling of your name, address, and phone, and whether that spelling matches everywhere else you appear on the web.
When an assistant decides who to name, it starts with the second face, not the first. Before it can weigh whether you are good, it has to understand what you are and where you work. That understanding comes out of the structured fields, not the photos. If those fields are vague, the machine does not know enough about you to put you on the short list, and everything you did on the first face never gets counted.
What the fields are actually worth
This is not a hunch. Whitespark studied 1.8 million Google Business Profiles across 4,209 categories and published the results through Search Engine Land, and the numbers are blunt about which fields move you.
Take the primary category, the single field that tells the machine what you do. Businesses that chose a specific primary category landed in the top 10 of local results about 36% more often than businesses that chose a generic one. Same work, different label, meaningfully different visibility.
The additional categories matter almost as much, and the trade numbers are stark. An electrician profile with well-chosen additional categories averaged rank 38.2. The same kind of profile with only a primary category averaged 47.5. And the profiles that had reached for the lazy catch-all category, "Service establishment," averaged 63.5 — worse than naming nothing useful at all. Roofing told the same story: 47.1 with good additional categories, 52.1 with none, 70.3 with the catch-all. The machine reads "Service establishment" as "this business could not tell me what it does," and treats it accordingly.
Then there is completeness. Whitespark scored profiles zero to five on whether they had a website, a description, hours, photos, and a claimed status. Profiles that scored zero averaged rank 62 and showed up in the top 10 four percent of the time. Profiles that scored five averaged 43 and hit the top 10 13% of the time. Filling in the boring fields roughly tripled a business's chance of being seen. Home-services profiles were already the most complete of any industry in the study, averaging 4.4 of five, which means the owners who are still at a three or a four are losing to neighbors who simply finished the form.
Correlation is not proof, and Darren Shaw, who ran the study, says so plainly. Owners who set specific categories tend to do other things right too. But the direction is not in question, and the fix costs nothing but attention.
The half you have been ignoring
Here is the part that stings. For two years the advice you have been given points almost entirely at the first face. Get more reviews. Shoot better photos. Post weekly. All real, all worth doing — I have written about choosing the profile photos that actually earn the next call, because the human who lands on your page does decide fast.
But you can do every bit of that and still be invisible to the machine, because the machine never got far enough to care about your photos. A plumber filed under a generic category, with a service area that is a single pin instead of the towns he actually drives to, is a plumber the assistant cannot confidently place. His reviews are excellent. They are also unread, because the machine could not establish what he is and where he works, and it will not recommend a business it cannot place.
There is a quieter version of the same failure. The machine cross-checks the name, address, and phone number on your profile against every other place you show up on the web — the old directory listing you forgot about, the number printed on the truck wrap's landing page, the spelling a supplier used. When those disagree, the machine does not know which business is the real one, so it trusts all of them a little less. A single stale phone number from a line you dropped two years ago can hold you off a list you otherwise belong on, and you will never see it, because on your own screen the profile looks right. This is the same reason you can rank first when you search for your own company and stay invisible to everyone else: you are looking at a version of the result the machine does not use.
The stakes on this are not drifting slowly either. Roughly 68% of United States Google searches now end without anyone clicking through to a website (SparkToro and Similarweb clickstream data, 2026). The homeowner is getting her answer inside the result, from the structured data, and leaving. When that answer is a spoken one, there is no short list at all — a voice assistant usually says a single name. Being read correctly by the machine is quietly becoming the whole contest, and the thing being read is the face you have not looked at. The answer a customer hears about you is assembled off your own website entirely, which is a problem I have taken apart in its own right.
None of this is about where things are headed. This is what a machine sees when it looks at your business today, this afternoon, if you let it.
What to do before you close the tab
Open your Google Business Profile right now, and do not look at the photos. Read your primary category out loud. Is it the exact thing you do, or a soft, general version of it? Change it to the precise one. Look at your additional categories and add the specific ones that fit the work you actually want — you get up to nine, and "Service establishment" is not one of them. Check your service area. If it is a single location pin, replace it with the real list of towns your trucks go to. Then do the one test that tells you the truth: open an AI assistant, type the exact question a homeowner two streets over would ask — "who should I call for a water heater in [your town]" — and see whether you come up. If you do not, you have just found 15 minutes of work worth more than a month of ad spend, and there is nobody to buy anything from to do it.
— Stacey Tallitsch, Stronghold CMO
About the Author
Stacey Tallitsch builds marketing and answering systems for home-services and trades businesses at Stronghold CMO, part of Talisman Capital, Inc. He is a 30-year technology veteran and the author of 21 books, with more than 30,000 students across his course catalog.
- LinkedIn: https://linkedin.com/in/stacey-tallitsch-729b6336a
- Books on Amazon: https://amazon.com/author/staceytallitsch
- Courses on Udemy: https://udemy.com/user/stacey-tallitsch
Quick reference
Why does an AI assistant recommend my competitor and not me when I have better reviews? Because the machine reads your reviews second, not first. It starts with your primary category and service area to work out what you do and where you do it. If those fields are generic or vague, it cannot confidently place you, so it leaves you off the list and your reviews never get weighed. Fix the structured fields before you chase more stars.
What is the single most important field on my Google Business Profile for getting found? Your primary category. In a study of 1.8 million profiles, businesses with a specific primary category appeared in the top 10 of local results about 36% more often than businesses using a generic one. Set it to the exact thing you do, not the general version.
How do I check what a machine actually sees when it looks at my business? Open your Google Business Profile and read the primary category, the service area, and the hours — ignore the photos, since those are for people. Then open an AI assistant and type the question a nearby homeowner would ask about your trade and town, and see whether you are named. It takes about 15 minutes and costs nothing.