Location page similarity checker

01What this tool is actually measuring

Google penalises pages created primarily to rank rather than to help someone. On a multi location dispensary site, the most common form of that is a set of location pages built from one template, where the only thing changing between them is the city name and the address. To a human that looks efficient. To a reviewer it looks like content manufactured at scale.

The measurement here is deliberate about one thing: it strips out your navigation, header, and footer before comparing anything. That matters, because every page on a website shares that chrome, and leaving it in would return a high similarity score for even the best written site. What gets compared is the body content, the part that is supposed to differ from store to store, using overlapping five word sequences so rearranged sentences still register as the same material.

A scaled content penalty does not hit one page. It hits the pattern, which means every location goes at once.

02Why chains fall into this, almost every time

The business model pushes you straight into it. A chain with fifty stores wants a page for every city, every page needs to rank for near me searches in its own area, and the fastest way to ship fifty pages is to build one and pour in variables. That efficiency is precisely the thing that gets flagged.

It gets worse in cannabis specifically, for a reason that is easy to miss. Because paid advertising is closed to you on the major platforms, location pages carry more weight than they would in any other retail category. They are not a supporting asset, they are the acquisition channel for each store. So when the pattern gets penalised, you do not lose a slice of traffic. You lose the primary way customers find every branch, on the same day.

The embedded menu compounds it further. On most dispensary sites the menu loads inside an iframe or through JavaScript, which means the genuinely unique content on each page, the products, never reaches Google at all. Strip that out and what remains is the templated wrapper. That is why dispensary location pages often score far higher on this test than the team expects.

03How to read your score

Under 45 percent means each store is carrying genuinely distinct content. There will always be some overlap, because every location page covers similar ground, and that is expected rather than a problem.

Between 45 and 70 percent means real local content exists, but there is enough repeated boilerplate that a reviewer could reasonably call the set templated. This is the band worth acting on early, because fixing it while the site is healthy is far cheaper than recovering from an action later.

Above 70 percent is the pattern that precedes a manual action. At that level the pages are functionally the same document with different addresses, and the only reason it may not have been actioned yet is that nobody has looked. Treat a result in this band as urgent rather than as a metric to monitor.

Look at the highest matching pair rather than the average. One pair at 94 percent tells you the template is still intact, even if the average looks acceptable.

04Fixing it at scale without going insane

The instinct is to hand the pages to a writer and ask for unique content. That fails predictably. A writer who has never visited the stores cannot invent genuine local detail, so they pad, and the pages drift straight back toward filler that is different in wording but identical in substance.

What works is a per store data sheet, filled in once by someone who actually knows each location. Real parking situation. Nearby landmarks people navigate by. Named staff and how long they have been there. On site amenities. Local promotions and any products this branch carries differently. The writer then assembles a page from facts rather than imagination, which is both faster and impossible to fake.

Keep one structure across every page so the site stays coherent, and change only the substance inside it. The test for any paragraph is simple: if it would still make sense with a different city name swapped in, it is boilerplate and needs rewriting. Apply that test ruthlessly and the score drops on its own.

05Keeping it fixed as the chain grows

Most chains fix this once and then quietly rebuild the problem, because the next three stores open and someone copies an existing page to get them live quickly. The standard has to survive the pressure of a launch, which means it needs to be written down and owned by someone.

Make the data sheet part of the store opening checklist, sitting alongside the licence and the signage. Add a hub and spoke internal linking model, where a city or state page links down to each store and every store links back up and across to its nearest neighbours, so Google reads the chain as one organised local network rather than a pile of orphaned pages. Then re-run this check every quarter and after every new opening. It takes a minute and it catches drift before it becomes a recovery project.

It samples up to eight location pages and compares them. It does not check every page on the site, review your Business Profiles, or tell you whether a manual action already exists. If you suspect one, open Search Console and check Security and Manual Actions first, because that changes the entire response.

Scans a page the way a search engine sees it. Finds embedded menus Google cannot read, age gates blocking the crawler, and…

Detects which menu platform you run and whether its products actually reach the HTML Google reads. If your menu loads in an…

Paste any marketing copy and see the medical claims, potency language, and youth appeal that get dispensary marketing pulled. Runs entirely in…