Is programmatic SEO dead? What the 2024 to 2026 updates actually changed

Programmatic SEO is not dead. The version of it that was lazy is dead, and it has been dying in public since March 2024. The honest question is not whether you are allowed to generate pages at scale, but whether each page you generate would survive being read on its own.

The short verdict

Google has never had a rule against making pages programmatically. It has a rule against making pages that exist only to rank. The scaled content abuse policy defines the problem as “many pages are generated for the primary purpose of manipulating search rankings and not helping users”, and it says that judgement applies whether the pages are made by automation, by humans, or by some combination of the two. The method was never the issue. The intent behind the page is.

So if your location pages are forty near-identical templates with a town swapped in, the updates since 2024 have been steadily bad news. If each page answers a real question for a real place, programmatic generation is just a faster way to build pages you would have wanted anyway.

What the policy actually says

Google’s scaled content abuse policy, announced with the March 2024 core update and in force since, treats thin templated pages as spam no matter who or what produced them. The policy describes pages that “are generated for the primary purpose of manipulating search rankings and not helping users”, and it calls out one pattern by name: creating many pages where the content “makes little or no sense to a reader but contains search keywords”. That second line is the one most location-page estates fail. The words make sense; they just make the same sense on every page.

This is the same line the doorway pages guide draws, approached from the other side. A doorway page is one of many similar pages built to funnel people to the same destination; the scaled content policy in plain English is the broader rule that catches the same behaviour when it is produced at volume. Rather than re-derive either here, treat this page as the verdict and those two as the detail.

What the core updates since 2024 changed

The March 2024 core update folded much of the old “helpful content” judgement into Google’s core ranking systems and shipped the scaled content abuse policy alongside it. Every core update since, through 2025, has kept tightening how that judgement is applied: less tolerance for pages that are technically present but substantively empty, and more weight on pages that are demonstrably useful for the query.

Reporting from the 2024 and 2025 updates consistently pointed the same way: estates of hundreds of near-identical location pages tended to lose visibility, while sites whose pages were genuinely different held or grew. That is an industry observation rather than a figure Google has published, but it matches what the policy targets and what the duplicate content on location pages measurement predicts. Near-duplicates get folded together; one survives, the rest quietly stop ranking.

One 2026 update is worth naming precisely, because it is widely misread. The February 2026 Discover core update was Google’s first Discover-only core update. It changed how articles are surfaced in Google Discover, favouring more locally relevant content and less clickbait, and it did not touch how location pages are ranked in Search. If your location pages moved in early 2026, that update is not the cause. The Search-side pressure on thin pages comes from the core updates, not from a Discover feed change.

So is it dead, or not?

The lazy version is dead. Spinning one template into hundreds of pages, swapping the town and the odd adjective, is exactly the pattern the policy now names. The data-backed version is doing fine, because it was never really “programmatic SEO” in the gaming sense; it was just building useful pages efficiently. The dividing line is whether you can defend each page in front of a reader, not whether a script helped you make it.

That is also the difference you can measure before you publish, rather than guess at after a core update. Measuring location pages in Search Console tells you which ones earn impressions and which never do; the similarity measurement tells you which ones read as copies of each other.

Where Townsmith fits

Townsmith builds real, editable location pages, scores each one from 0 to 100 with its Quality Score, and holds the thin ones back behind an advisory gate, so you can catch the duplicate-template pages these updates target before they go live. It is not a guarantee of ranking and nothing can promise that. It is a way to see the page the way the policy would, while you can still change it. The features overview covers what it checks and where the line sits.

Common questions

Is programmatic SEO against Google’s rules?

No. Generating pages programmatically is not itself a violation. The scaled content abuse policy targets pages that “are generated for the primary purpose of manipulating search rankings and not helping users”, and Google states this applies whether the pages are made by automation or by humans. A programmatic page that genuinely helps a reader is fine; a hand-written page that exists only to rank is not.

Did the 2024 and 2025 core updates penalise location pages?

They penalised thin, near-identical location pages, not location pages as a category. A page built for a place it serves, with content that is true of that place and useful to someone searching there, is exactly what these updates reward. The estates that lost visibility were the ones built from a single template with little real variation between them.

Was the February 2026 Discover update a location-page penalty?

No. The February 2026 Discover core update was Google’s first Discover-only core update. It changed how articles are surfaced in Google Discover, not how pages rank in Search, so it does not demote thin location pages in normal search results. If your location pages shifted in early 2026, look to the core updates and to the pages themselves, not to that Discover change.

How do I tell if my pages are on the wrong side of the line?

Read two of your pages side by side. If the only differences are the place name and a swapped adjective, they read as copies of each other and are at risk. Measuring similarity between sibling pages, and checking in Search Console which pages actually earn impressions, turns that feeling into something you can act on before a core update does it for you.