The number that matters is not your average. It is your worst pair.
I have measured chains whose average similarity looked fine at 51 percent, while their worst two pages sat at 94. The average was reassuring and useless.
Google does not judge your location pages as a mean. It sees pages. When two of them are the same document with a different address, it draws a conclusion about how the whole set was made.
So before anything else, find your two most similar store pages and read that number. Everything below assumes you have it.
Where The Thresholds Sit
Under 45 percent is healthy for pages about the same kind of subject. Some overlap is normal and correct. Every store has hours and a phone number, and the legal notice at the bottom is the same by design.
Between 45 and 70 is the templated band. Nothing is broken and nothing is urgent. But there is enough shared wording that a reviewer could fairly call the set one document, reused. Most chains sit here without knowing it.
Above 70 is where I stop calling it a quality issue and start calling it exposure. The pages are one document with different addresses.
Google does not publish these numbers, and I am not presenting them as policy. They are where problems start to appear on accounts I have worked on. That is a weaker kind of evidence, so use the bands to set priorities, not as a line to sit just under.
What Google’s Own Policies Say
Two of Google’s spam policies describe this pattern closely, and it helps to read them in Google’s words rather than an agency’s.
Doorway abuse. One example Google gives is having pages “targeted at specific regions or cities that funnel users to one page”. A page for each city, each saying the same thing, is the textbook case.
Scaled content abuse. Google defines this as many pages generated mainly to manipulate rankings, not to help users. Its examples include using AI tools to make many pages without adding value for users.
A real store with a real address is not a doorway. The page becomes a problem when it tells a customer of that store nothing they could not read on any other store’s page.
There is also a quieter cost that arrives long before any penalty. When store pages are near copies, Google can treat them as one page in different clothes and pick one to show. That is often why two of your stores fight over one keyword and neither moves.
Why The Score Is Worse Than The Pages Look
This is the part that surprises people. It is worth understanding before you argue with the number.
A similarity check strips the wrapper first. The header, navigation, hero band, sidebar and footer all come out. They are the same everywhere by design, and leaving them in would make every page on the web look alike. What remains is body content.
On a dispensary site, the body content is often much thinner than the page looks.
Three things make it thinner still. If the menu arrives through an iframe or JavaScript, the products never reach the HTML, so they cannot count toward what makes the page unique. A third-party reviews widget is shared markup too.
Then there is the age gate. If your age gate shows Google a different page from the one a customer sees, the crawler may be comparing two copies of the gate.
Strip all of that away. A page that felt substantial can come down to an address, an hours table and one paragraph of welcome copy. If that paragraph has the same shape across your stores, your real overlap is close to total.
Why Rewriting Usually Fails
Almost every chain I have worked with has already tried to fix this. Most tried the same way: they gave the pages to a writer and asked for a rewrite.
It rarely moves the number much, and the reason is worth sitting with.
A rewrite starts from the existing page. The writer reads what is there, changes the words, and improves the flow. What comes back reads better and says the same thing.
The vocabulary is new. The substance is not. A check that compares runs of words sees straight through it, and so does a human reviewer.
Here is the test I use. It costs nothing.
Take any paragraph and swap the city name for another of your stores. If it still makes complete sense, it is boilerplate. It does not matter how well it is written, how many synonyms it uses, or how recently it was rewritten.
“Visit our welcoming Salem store for a great selection and friendly staff” survives the swap, so it is boilerplate. “Parking fills by six on Fridays, so use the Cross Street lot behind the building” does not survive it, so it is unique.
My highest reading on a live chain was 94 percent, for two stores in different states.
The team had rewritten those pages. Twice.
Both rewrites started from the same document, so the words changed and the substance did not. Nobody had visited the stores, called the managers, or looked at anything the pages did not already say.
What Actually Makes A Location Page Unique
The answer is facts nobody can invent from a desk. That is the whole principle. Once you accept it, the work becomes plain rather than creative.
1. Get the facts before anyone writes. One sheet per store, covering the six things above. Each one has a person or a record that knows the answer, and none of them can be guessed.
2. Write from the sheet, not from the old page. Do not open the old page at all. This is the step people skip, and skipping it is why the last rewrite failed. The old page is the thing you are trying to escape.
3. Run the swap test on every paragraph. Anything that survives a city swap gets cut, or replaced with something from the sheet. It is mechanical, it takes minutes, and it catches almost everything.
4. Fix delivery too, or none of it counts. If your menu is invisible and your age gate blocks the crawler, your new paragraphs will be the only thing Google can read. There will not be enough of them.
For a twelve store chain, filling the sheets takes about a week of phone calls. After that the writing is fast, because the hard part of a location page was never the writing. What goes on the sheet, and the twelve minute intake that fills it in, is the part of the system worth copying.
What One Paragraph Looks Like, Before And After
Here is the difference in practice. The store below is made up for the example, but the shape is the one I see on real chains.
Before, written from the old page: “Our Salem location offers a wide selection of flower, edibles and concentrates in a welcoming space. Our knowledgeable budtenders are happy to help you find the right product.”
Run the swap test and it passes for every store you own. It says nothing a Salem customer could use.
After, written from the sheet: “We are on the corner of Front and Mill, across from the hardware store. The side lot fills up after work on Fridays, so the street spaces on Mill are your best bet. Most of our regulars drive in from the towns to the south.”
Swap the city and every sentence breaks. That paragraph could only have been written about one store, and it answers questions a real customer asks before they leave the house.
Where Shared Content Should Live
Some content really is the same for every store. Your brand story, your loyalty scheme, your returns policy and your legal notices do not change by town.
The mistake is repeating all of it on every store page. Each copy adds shared words to the comparison and pushes your overlap up.
Put the shared material on one page and link to it from each store. The store pages then carry only what is true of that store, plus a short link to the rest.
Keep legal text short on each page too. If your state requires a notice on every page, use the exact required wording and nothing more. A three paragraph disclaimer copied forty times is a big share of what gets compared.
Check it in about a minute
The location page checker samples your store pages, strips the shared wrapper, and shows the overlap for every pair. You see your worst pair, not a flattering average.
Why Chains Feel This And Single Stores Do Not
A dispensary with one location page has no similarity problem at all. This is a scale issue, and it gets worse in a set order.
At two stores, nobody notices. At six, a pattern exists but nothing has gone wrong. At twenty, the template has been reused often enough that it is plainly a template. At fifty, you are carrying the exact pattern those spam policies were written to find.
The failure at scale is not gradual either. A manual action applies to the pattern rather than to a page, so it can reach every store on the same day. A chain does not lose ten percent of its visibility. It loses the map.
If that has already happened, what to do in the first 24 hours matters more than anything on this page.
Do not spin the pages, and do not generate them.
Spinning gives you pages that differ in words and match in substance. That is the same problem with more steps and a worse paper trail.
Generating pages is no better. A template with a few variables, or an AI tool fed no new facts, makes the exact thing Google’s scaled content policy describes.
If you would be uneasy showing your process to a Google reviewer, that unease is the finding.
What To Do This Week
If your worst pair is above 70, stop publishing new location pages until the pattern is fixed. Adding more copies of a risky pattern while you fix it is the most common way these projects fail.
If you are in the 45 to 70 band, you have time, and the sheet is the place to start. It is not urgent yet. It will be, and it is far cheaper to do while nobody is under pressure.
For chains with dozens of pages in that band, the rebuild is what our multi-location engagement is scoped around.
If you are under 45, check again after your next site build. This problem is almost always brought in by a redesign, when somebody sensibly decides to standardize the location pages.
Any dispensary SEO agency worth hiring should check page similarity before a redesign launches, not after rankings fall.
If you take one thing away
Read your worst pair, not your average. Fix it by gathering facts, not by rewriting sentences.