Here is the thesis I keep returning to: programmatic SEO is neither a growth hack nor a spam factory by nature. It is a tool that works exactly as well as the underlying data and user intent behind it, and fails in direct proportion to how thin and undifferentiated its pages are.
I have watched teams launch ten thousand templated pages and gain nothing but crawl budget problems. I have also watched a solo developer build a business on the back of a few hundred structured city pages. The difference was not the production method. It was whether each page gave a real person something they could not get anywhere else.
This post tries to give you the honest version of the programmatic SEO conversation, including the case for it, the case against it, and a framework for deciding which side you are actually on.
What Is Programmatic SEO?
Programmatic SEO is the practice of generating large numbers of web pages from a structured data source and a repeatable template, with the goal of ranking for many related search queries at once.
The template handles the layout and surrounding copy. The data source provides what makes each page unique: a currency exchange rate, a job listing, a city's cost-of-living breakdown, a software integration's trigger-action list. Done right, each page is genuinely distinct and genuinely useful, even if the template wrapping it is identical. Done wrong, you get a thousand pages that all say roughly the same thing with the city name swapped out.
That distinction is the entire argument.
The Case For Programmatic SEO
The strongest argument for programmatic SEO is that the best examples of it are not shortcuts at all. They are the most efficient way to deliver a large body of genuinely useful, structured information to people who are searching for exactly that information.
Real examples that worked, and why
Zapier's integration pages are the canonical pSEO success story, and it is worth understanding why they actually worked rather than just citing the traffic numbers. Each page in Zapier's /apps/ directory is generated from Zapier's own integration database. The page for "Slack + Google Sheets" contains the actual triggers, actions, and supported fields for that specific pair. It is not a thin template with a few sentences about "connecting your tools." It is functional documentation with a direct path to setting up the integration. Ahrefs' February 2023 case study put Zapier's blog at 1.6 million monthly organic visits, with integration pages contributing roughly 16% of total organic traffic.
Wise's currency-pair pages are another example that holds up under scrutiny. Each /currency-converter/usd-to-eur/ page shows a live exchange rate, a historical chart, a conversion calculator, and a fee comparison. The data is proprietary and real-time. You cannot get that specific combination anywhere else with that level of accuracy. As of a July 2024 Ahrefs snapshot, Wise was generating 6.4 million monthly organic visits, driven substantially by that currency-pair template system.
Canva's /templates/ pages follow the same logic. The page for "Instagram post templates" contains actual usable templates. You can interact with them. The value is not the copy on the page; it is the product embedded in the page. A March 2025 analysis by Daydream estimated Canva's template directory at 13.1 million monthly organic visits across about 21,000 pages.
Nomad List, built largely by a single developer, runs structured data pages for cities around the world, with cost-of-living scores, internet speed rankings, weather, and safety data. An Ahrefs October 2023 snapshot put it at around 41,000 monthly organic visits across roughly 26,000 pages. Not blockbuster numbers, but built by one person, with genuine utility, and presumably a reasonable return.
Sources: Ahrefs case study (Zapier, Feb 2023); Ahrefs SEO case studies index (Wise, July 2024); Daydream/Ahrefs (Canva, March 2025); Ahrefs (Nomad List, Oct 2023). All figures study-reported; tools and measurement dates differ.
What these examples share
Strip away the brand names and the scale, and the pattern is consistent. Each program works because:
The data is proprietary or aggregated in a way that requires access, infrastructure, or user contributions that competitors do not have. Wise has real-time rates. Zapier has first-party integration data. Canva has actual usable templates. Nomad List has its own crowd-sourced database. None of these pages could be cloned by a competitor without rebuilding the underlying data product.
Each page answers a specific, real search query. "USD to EUR exchange rate" is a query with genuine intent. "Slack Google Sheets integration" is a query with genuine intent. The template is not manufacturing queries; it is matching pages to queries that already exist in volume.
The template adds structure, not padding. The template handles navigation, layout, and consistent headings. It does not generate paragraphs of boilerplate copy to hit a word count. The content is the data, not the wrapper.
This is what Google's scaled content abuse policy is actually distinguishing: pages made "for the primary purpose of manipulating search rankings and not helping users" versus pages that happen to be generated at scale because the underlying data is large and structured.
The Case Against Programmatic SEO
The case against is real, and I want to give it its full weight rather than treat it as a straw man.
Thin pages at scale get penalized at scale
The failure mode of programmatic SEO is not subtle. When the data is thin, the pages are thin. When the pages are thin and there are ten thousand of them, Google's systems notice. The March 2024 core update, which folded the Helpful Content System into core ranking, was explicitly designed to demote this pattern at scale.
Search Engine Journal's analysis of 49,345 monitored sites found 837 were fully deindexed in the update. Those sites collectively lost over 20 million monthly organic visits. An Originality.ai study cited in the same analysis found that all affected sites showed signs of AI-generated content, with half having 90 to 100 percent of their posts AI-generated.
The specific failure pattern at G2 and ZoomInfo, documented by AirOps in a February 2025 analysis, is instructive. G2 built database-generated pages at scale for software categories and comparison queries. ZoomInfo generated millions of pages for employee contact lookups. Both saw reported "boom and bust" cycles as Google rewarded initial crawling and then later demoted pages that were not providing meaningful value beyond what was already indexed elsewhere. The AirOps source is a secondary analysis without a direct primary data link on the traffic figures, so treat the specific numbers with appropriate caution, but the pattern is consistent with what Google's own policy documents describe.
John Mueller's warning has two halves
Google's John Mueller made a statement in July 2023 that circulates widely in the SEO community. The version most people share is: "Programmatic SEO is often a fancy banner for spam." What most citations omit is the rest of the sentence: "To be fair, programmatic SEO is not always spam but hey: Forever the optimist." Search Engine Roundtable documented the full quote.
The context matters too. Mueller was responding to Lily Ray documenting sites that scraped data from other sources, ran it through templates, and published at volume. His skepticism was about that specific practice, not about programmatic page generation as a category. But his instinct that most practitioners conflate the two is fair.
The maintenance burden is real
Even well-executed pSEO programs carry costs that rarely appear in the case studies. Data goes stale. Templates need updates when the underlying product changes. Pages that ranked well can quietly decay when the data they pulled from is no longer accurate or useful. A Wise currency page with a rate that has not updated in six months becomes a trust problem, not an SEO asset.
At scale, auditing programmatic pages for quality, staleness, and accuracy is a significant operational commitment. Most teams underestimate this going in. The launch is fast. The maintenance is ongoing.
AI search is raising the bar for specificity
Generative engines like ChatGPT, Perplexity, and Google's AI Overviews do not reward volume. They reward specificity and citation-worthiness. The KDD 2024 GEO paper found that adding sourced statistics lifted generative-engine visibility by roughly 31% and adding quotations from authoritative sources lifted it by roughly 41%, on a simulated engine. Pages that are thin templates, even if they rank on traditional search, are structurally poor candidates for AI citation because they contain no quotable, attributed claims.
If your pSEO program generates pages that say "The best hotels in Austin, TX include a variety of options for business and leisure travelers," you are not a candidate for a Perplexity citation. If your page says "Austin has 47 Marriott-brand properties within 10 miles of the downtown convention center, with a median rate of $189 in Q1 2025 (source: Marriott public data)," you might be.
How to Tell If Programmatic SEO Is Right for You
The decision is not about scale. It is about data. Answer these questions honestly before building:
Do you have data that is genuinely unique? If your data comes from publicly scraped sources that any competitor could replicate, your pages are not differentiated. Unique means: first-party (you collected it), licensed exclusively, aggregated in a way that requires non-trivial infrastructure, or enriched with user contributions no one else has.
Does a real search query correspond to each page template? Use a keyword tool to verify search volume before building. If you are generating "Denver plumber comparison" pages but the query has no search demand, you are not doing SEO; you are creating crawl budget waste.
Is the template adding value, or filling space? If you removed the template wrapper and left only the data, would the page still be useful? If the answer is no, your template is load-bearing in a way that signals thin content.
Can you maintain data freshness at scale? If the data powering your pages has a shelf life, you need a refresh mechanism before you build, not after.
Does the program survive a Google penalty? This is the stress test. If a manual action or algorithm update deindexed 30% of your programmatic pages tomorrow, would the business survive? If the program is a core traffic strategy with no complementary editorial investment, the concentration risk is significant.
If you can answer yes to all of these, programmatic SEO is probably worth building. If you are rationalizing past a "no" on data uniqueness or search demand, the downside risk is real.
Doing It Responsibly If You Do
If you have cleared the decision framework, a few operating principles reduce the risk of a boom-bust cycle:
Index selectively. Not all programmatic pages deserve to be indexed. Use noindex or canonical consolidation for pages where the data is thin or overlapping. Google's systems are better than they used to be at detecting low-quality pages within a programmatic set, and a smaller indexed footprint of high-quality pages is preferable to a large footprint of mixed quality.
Invest in the editorial layer. The best pSEO programs are not purely programmatic. Zapier has a large editorial blog that earns links and supports the integration pages with authoritative context. Canva has design tutorials and a creator community. The template pages benefit from the editorial authority built around them. Treating pSEO as a standalone growth motion, disconnected from an editorial content strategy, produces a more brittle result.
Monitor page-level performance, not just aggregate traffic. A pSEO program that is generating 100,000 pages and 500,000 monthly visits can mask the fact that 90,000 of those pages earn zero clicks. Build dashboards that surface per-page click distribution early. Prune or consolidate pages that are not earning organic engagement.
Do a content gap analysis before scaling. Know what your competitors have already indexed and where genuine demand exists before you build the template. A well-scoped programmatic program that targets unmet demand is far more durable than one that replicates what already exists at scale.
Treat pSEO as part of a content estate, not a replacement for it. The programmatic pages earn traffic for high-volume, specific queries. Your topic clusters and pillar pages build the topical authority that makes Google trust your domain in the category. Both need each other. See also how E-E-A-T signals apply at the domain level, not just the page level.
For a grounding on what gets content demoted versus rewarded in 2026, the why AI content fails to rank breakdown covers the same undifferentiated-at-scale failure pattern from a different angle.
FAQ
Is programmatic SEO against Google's guidelines?
No, not by definition. Google's scaled content abuse policy targets pages generated "for the primary purpose of manipulating search rankings and not helping users." If your programmatic pages provide genuine utility and unique data, they comply with that standard. The policy is about intent and output quality, not production method.
How many pages is too many for a programmatic SEO program?
There is no hard limit. The question is not page count but page quality. A program with 500 genuinely unique, data-rich pages is better than a program with 50,000 thin pages. If you cannot answer what unique, user-facing value each page provides beyond its template, that is the signal to reduce scope.
What is the difference between programmatic SEO and scaled content abuse?
The Google policy language is the cleanest guide here: scaled content abuse is pages made "for the primary purpose of manipulating search rankings and not helping users." Programmatic SEO that puts real, useful, unique data in front of users who are searching for exactly that data is not abuse. Programmatic SEO that repackages publicly scraped data into keyword-targeted templates with no additional user value is exactly what the policy describes.
Does programmatic SEO still work after the March 2024 core update?
It works if it worked for the right reasons before. Programs built on proprietary, useful data with genuine search demand have not been systematically penalized. Programs built on thin templates at scale, particularly those using AI to generate filler copy rather than embedding real structured data, were the primary targets of the March 2024 update and subsequent core updates. The mechanism has not changed; the enforcement has become more consistent.
Can a small team run a programmatic SEO program?
Yes. Nomad List was built largely by one person. The constraint is not headcount; it is data. If you have access to a structured dataset that is genuinely useful and corresponds to real search demand, you can build a programmatic program at whatever scale your engineering resources support. The editorial investment, ongoing data maintenance, and performance monitoring do require sustained attention, but they do not require a large team.
The honest conclusion here is that programmatic SEO is a legitimate strategy when it is built on the right foundation, and a meaningful risk when it is not. The line between the two is not about scale or automation. It is about whether the underlying data gives users something they cannot get from the ten other pages that come up when they search.
In my experience building SparkBlog, the teams that succeed with content-led growth treat their data assets as the actual product and the content strategy, programmatic or otherwise, as the distribution mechanism. That framing changes what you build. Instead of asking "how many pages can we generate?" you start asking "what do we know that no one else has indexed yet?" That is the question worth answering before you touch a template.