What Is a Doorway Page?

The term is also translated as bridge page or gateway page in other languages. The classic explanation describes a very specific technique: a text-heavy page with little design, almost no images or JavaScript, built so that a crawler could read it without difficulty. It relied on a "hallway page" linking to it, and as soon as a visitor arrived from the search engine, an immediate redirect sent them to the real website, without ever letting them see the page that had just been indexed. The user entered a one-way street with no way back.

That description fits a very specific period: sites built in Flash or with excessive JavaScript, technologies a crawler back then could barely process. The doorway page worked as a readable stand-in for content that otherwise would not have been indexed at all. Over time, the technique was mostly repurposed for something else: piling up ranking factors through keyword stuffing, instead of solving a real accessibility problem.

Google no longer frames the problem that way. In its current spam policies documentation, it classifies this practice under doorway abuse and defines it through two traits that do not depend on any specific technology: sites or pages created to rank for specific, similar search queries, which lead users to intermediate pages that are not as useful as the final destination. There is no mention of instant redirects or code a crawler cannot read. It is about the result the searcher actually gets.

The practice still falls under black hat SEO, and it is often mentioned in the same breath as cloaking, even though the two are different techniques: cloaking shows different content to the crawler and to the person, while a doorway page usually shows the same content to both, only that content delivers less than it promises.

Google's documentation lists several doorway abuse patterns, and none of them depend on the technology the page is built with. The first is running multiple websites with slight variations in the URL and home page, aimed at covering as many variants of the same query as possible. In practice, that means multiplying near-identical domains or URL slugs for the sole purpose of appearing several times for the same search, not to offer genuinely different content.

The second pattern is running multiple domains or pages targeted at specific regions or cities that all funnel toward a single page. This is the case built most often today, and the one almost nobody recognizes under this name: dozens of pages like "service in Boston", "service in Austin" or "service in Denver", each optimized for a local search intent, that actually lead to nothing different depending on the city. This pattern gets a closer look in the next section, because it is the one with the greatest practical impact.

The third is pages generated solely to funnel visitors into the genuinely useful or browsable part of a site, without contributing anything on their own. The fourth is creating substantially similar pages that sit closer to a list of search results than to a browsable, meaningful hierarchy, the kind usually reflected in a well-built sitemap. In both cases the problem is not one isolated technical element, but a group of pages that exists for the search engine rather than for the person visiting it.

All four patterns share the same core: several pages competing for similar queries, with none of them justifying its own existence on its own terms. That is different from duplicate content that appears by accident, a common technical error with no intent to manipulate results, even though in practice both problems can coexist on the same site and are worth reviewing together.

Of the four patterns, the one built most often, and recognized least often as a problem, is the regional or city-page case, because it carries none of the signals that used to be associated with a doorway page: no unreadable text, no instant redirect, no obvious keyword stuffing. The page exists, has a design, and can even be well written. The problem is not how it looks, but what it delivers.

The difference between a legitimate city landing page and a doorway page does not depend on the intent of whoever built it, but on a checkable fact: whether that page offers something the destination page does not. A landing page for local SEO with genuine value includes information that only makes sense for that city or that area: a real address, opening hours, a specific service area, references or customer cases from that location, availability or pricing that changes by region. A doorway page, on the other hand, repeats the same generic text with the place name swapped out, and ends up funneling the visitor toward the general page, whether or not a redirect is involved.

This distinction has a practical consequence for any business with a presence across several cities or countries: not every regional variant is suspicious. When the variant is a matter of language rather than content, the correct technical signal is hreflang, which tells Google which version to show depending on the user's language or region without that implying redundant pages competing with each other. The problem is not having several pages for several areas, it is having several pages that say the same thing with a different city name in front.

A business with offices or genuine coverage in ten cities has legitimate reasons for ten distinct pages. A business with a single location that builds ten identical city pages to show up in ten local searches falls, under Google's current definition, into the second doorway abuse pattern, whether or not there was any intent to deceive anyone. What matters is not what whoever built the pages intended, but what the person visiting the page actually gets.

• Give every city or region page something specific to that place that the general page does not have: an address, opening hours, a service area, local references, or any detail that genuinely changes by location, not just the name.
• Link those pages from the main navigation and from the sitemap, so any visitor can reach them without depending on a specific search.
• Avoid redirecting the user automatically the moment they click. If a page no longer adds anything distinct after review, merge it with a 301 redirect to the page that works, instead of keeping it alive through a redirect chain to another site.
• Use canonical or noindex when a variant is nearly identical to another and both should not compete for indexing, instead of letting dozens of near-identical pages get indexed all at once.
• Check with your own data whether those pages get their own traffic and conversions, or only generate bounces back to the parent page; that shows whether they genuinely serve a different search intent.
• Consolidate whatever turns out, once measured, to be pure thin content with no contribution of its own.

• Confusing any landing page optimized for a specific query with a doorway page. A page with its own purpose and content that supports it does not fall into this category, no matter how tightly it targets a keyword.
• Assuming it is enough to avoid the old techniques to be safe. Google's crawlers have processed JavaScript normally during crawling for years, so the current test is not whether a page is technically readable, but whether it offers the user anything of its own.
• Copying the same text across city pages and swapping only the place name. That is duplicate content and a form of geographic keyword stuffing at the same time, and both are common signals of thin content.
• Not checking whether those pages produce a soft 404: a page with no real content left that still returns a 200 status instead of a proper 404 error or a meaningful redirect.
• Forgetting to check robots.txt and the sitemap to confirm which version of each page is actually meant to be indexed, especially after merging or removing city pages.