A location page works when it matches a search people make in that place and says something that could not be pasted onto another city's page. It fails when it is one template with the city name swapped, which Google's spam policy calls doorway abuse.
On one service client's site, 65% of the impressions on 13 city pages came at position 10 or better over the 90 days to 30 August 2026. Smaller cities did best. The Jaipur page held 98% of its impressions on page one and Delhi held 26%, though Delhi's page was the longest of the set.
At a glance
- On page one, 13 city pages
- 65% of impressions, 90 days to 30 August 2026
- Strongest
- Jaipur 98%, Indore 97%, Ahmedabad 94%
- Weakest
- Mumbai 27%, Delhi 26%
- Our limit on shared text
- 40% between any two pages
Competition decides more than length
We measured the share of each page's impressions that came on Google's first page. The pattern followed how crowded each city's results were.
| City page | On page one |
|---|---|
| Jaipur | 98% |
| Indore | 97% |
| Ahmedabad | 94% |
| Lucknow | 90% |
| Chennai | 86% |
| Chandigarh | 81% |
| Bengaluru | 79% |
| Surat | 72% |
| Hyderabad | 64% |
| Pune | 58% |
| Kolkata | 42% |
| Mumbai | 27% |
| Delhi | 26% |
Delhi's page was the longest, at 1,650 words with eight sections and six neighbourhoods, and the least repetitive of the set. It still came last. Jaipur reached the top with 1,333 words and four neighbourhoods.
In smaller cities fewer businesses compete for the same searches, so a good page reaches page one sooner. Start there, and treat the largest metros as a long job.
Split a weak page's searches before judging it
A page's average position hides two kinds of search. Delhi's page appeared for 119 searches. The 16 that mentioned Delhi averaged position 11.7, a few places from page one. The other 103 were national searches the page was never meant to win, and they averaged 36.1. Together they gave a page average of 24.4, which made the page look far weaker than it was.
Before you rewrite a city page, separate the searches that name the city from the ones that do not, and judge the page on the first group.
The duplication test we run
Location pages built from one template share most of their words, and Google's spam policy lists pages for specific cities or regions that funnel visitors to one page as doorway abuse. So we measure how much text any two pages share before they go live.
We take every run of eight words in the main content of each page, with city and neighbourhood names removed, and count the runs two pages have in common. The 13 India pages, written from a template before we measured, shared a median of 58.8%. Four later pages for cities abroad, each written from scratch, shared between 13.0% and 16.8% with their closest match. Our limit is 40%.
Adding new paragraphs to a duplicated page helps less than it seems. When we lengthened the opening answer on five of the template pages, the median overlap only fell from 58.8% to 55.3%. Writing each page separately is what fixes it.
Source: Google Search Central: Spam policies, doorway abuse (updated 28 August 2026).
What makes a location page worth having
- The words people in that city search. Check Google's suggestions city by city. One of the client's Delhi pages was named after a phrase that Search Console never showed anyone using. It earned its impressions instead on a phrase asking for the best provider in Delhi.
- Facts that belong to that place. For the pages abroad, the time difference changed the whole answer. In summer, Toronto is 9.5 hours behind India, London 4.5 hours and Dubai 1.5 hours, while Singapore is 2.5 hours ahead.
- An honest line about where the work is done. Never imply an office you do not have: no local address you cannot staff, and no local number that rings somewhere else. Google's Business Profile guidelines rule out virtual offices, and a false address can get a profile suspended.
Source: Google Business Profile Help: Guidelines for representing your business on Google.
Searches with a map are a different job
When Google shows a map for a city search, the businesses in the map take most of the attention, and a city page alone will not get you into it. That depends on your Google Business Profile, its category and its reviews. City pages do best on searches without a map, and in the AI answers that Google builds from pages that rank.
What we changed on our own site
We launched this site with 20 city pages. Reading them back, too much of each page could have sat on any other, so in October 2026 we replaced them with five researched guides to the largest markets and redirected the rest. Five pages that each say something new serve a reader better than twenty that repeat each other.
Questions people ask
Are location pages bad for SEO?
No. Pages that give each place something useful of its own are fine. Google's spam policy is aimed at pages made only to rank for similar searches, which then funnel visitors to one page.
How many location pages should I make?
As many as you can write separately, with facts that belong to each place. Start with smaller cities, where fewer businesses compete, and add the large metros once those pages work.
How different should location pages be?
In our checks, no two pages share more than 40% of their text, measured on runs of eight words with place names removed. Different facts matter more than different wording.
Do I need an office in each city?
No, but never suggest you have one. Say where the work is done. Google's Business Profile guidelines do not allow virtual offices, and a profile with a false address can be suspended.