The myth: Google’s March 2024 core update, the integration of the Helpful Content System into Google’s core ranking systems, and the rise of AI Overviews killed programmatic SEO as a defensible growth strategy.
The reality: They exposed the difference between thin, keyword-swap templates and database-rich programmatic SEO built around genuinely useful information.
The evidence is visible across today’s search landscape. In 2026, Wise still generates roughly 60 million monthly organic visits through its currency converters, Zapier around 16 million through more than 50,000 integration pages, and Canva more than 100 million through its template gallery. At the same time, G2, ZoomInfo, Datanyze, and a long tail of low-differentiation publishers have lost about 76%-99% of their organic traffic.
These outcomes are not contradictory. They reflect the same algorithmic shift separating programmatic pages that deliver unique value from those built primarily to capture search demand.
Key findings
- Google’s recent updates raised the quality threshold by rewarding pages that provide unique value and filtering out thin, repetitive content created primarily to capture search traffic.
- The strongest programmatic SEO strategies are built on proprietary, structured data. Sites such as Wise, Zapier, Zillow, Canva, and Tripadvisor continue to perform because each page delivers information that cannot be easily replicated.
- AI-generated answers are reshaping search behaviour. As more informational queries are answered directly in search and AI assistants, transactional tools, calculators, structured databases, and other task-oriented pages have become more resilient sources of organic visibility.
- AI search and traditional search do not evaluate content in the same way. Pages supported by statistics, credible sources, structured data, and clear entity signals are better positioned for AI citations alongside organic rankings.
- Publishing thousands of pages is no longer a competitive advantage on its own. Every programmatic page should justify its existence by solving a specific user need, offering unique information, and contributing measurable value.
The myth: “Programmatic SEO is dead”
The belief that programmatic SEO no longer worked was grounded in real evidence. Following the September 2023 Helpful Content Update and Google’s March 2024 core update, several high-profile websites reported significant traffic losses, reinforcing the perception that large-scale programmatic publishing had become a liability.
| Example | What happened |
| HouseFresh | Lost more than 90% of its Google traffic following Google’s Helpful Content Update and subsequent core updates. |
| Glenn Gabe’s HCU analysis | Of roughly 400 heavily impacted sites, only around 22% recovered more than 20% of their lost traffic by August 2024. |
| G2 | Estimated organic traffic declined by 80–85%. |
| ZoomInfo | Estimated organic traffic declined by about 76%. |
| Datanyze | Estimated organic traffic declined by about 96%. |
| Causal | Lost approximately 99.5% of its organic traffic after publishing 1,800 AI-generated articles. |
Taken together, these examples reinforced the belief that programmatic SEO itself had become the problem. That conclusion, however, was only partly correct. The issue was not programmatic SEO as a publishing method, but a specific type of programmatic SEO built around thin, low-value pages.
The reality: what the algorithm actually targeted
Google’s March 2024 spam policies define scaled content abuse as content “produced at scale to boost search ranking, whether automation, humans, or a combination are involved”. Google’s guidance focuses on whether content serves a genuine purpose and provides value to users. Google’s AI content guidance makes the same distinction explicit: appropriate use of AI or automation is not against the guidelines; using automation with the primary purpose of manipulating ranking is.
The published policy does not mention programmatic SEO by name. Instead, it asks whether each page provides genuine value to users. Looking across the sites that continued to perform, a clear pattern emerges:
Pages that continued to perform
- Wise: Live exchange rates, fee calculations, and transfer options backed by proprietary financial infrastructure.
- Zapier: Integration pages built from proprietary app compatibility data.
- Zillow: Property pages tied to individual real estate listings.
- Canva: Template gallery where the page itself is the product.
- TripAdvisor: Destination pages enriched with reviews, traveller ratings, and FAQs.
The sites that lost visibility also shared common characteristics:
Pages that lost visibility
- ZoomInfo: Heavy losses in /companies and /people pages built around thin, aggregated data.
- Datanyze: Similar declines across the same types of pages.
- RemoteOK: A sharp decline in ranked keywords after Google’s updates affected large numbers of templated job pages.
The pattern is consistent. Pages backed by unique data and genuine utility remained resilient, while thin, repeatable pages became far more vulnerable to Google’s quality systems.
The March 2026 update added another dimension. Lily Ray’s analysis at Amsive found a shift toward originator domains, such as the actual employer, hotel, or product provider, rather than aggregators and comparison sites. Travel, jobs, and finance were among the sectors where government, professional, and first-party sources gained visibility. This “go-to-source” pattern makes it increasingly difficult for programmatic SEO strategies that rely primarily on republishing third-party information.
Why do the unit economics now sort differently?
Two structural shifts have changed the math.
- First, search behaviour has changed. SparkToro’s 2026 research found that 68% of Google searches in the U.S. now end without a click. As AI-generated answers become more prominent, many informational queries are resolved directly on the search results page. This has reduced the opportunity for long-tail informational pages to attract organic traffic, making traditional programmatic SEO strategies less effective for those query types.
- Second, the queries where pSEO continues to work are the ones AI cannot fully replace: transactional queries, database and calculator pages where live data is required, and template galleries where the page itself is the product. Traffic volume and traffic value are increasingly diverging for publishers with structured, data-rich pages. As AI-generated answers satisfy more informational searches, the remaining clicks are more likely to come from users with a specific task or higher purchase intent. The ghost traffic metric is the right frame for measuring what GA4 will miss in this environment.
The Princeton Generative Engine Optimization (GEO) paper (KDD 2024) found that AI citation systems reward different signals than traditional organic ranking. Pages enriched with statistics, quotations, and cited sources consistently achieved higher visibility in AI-generated answers, while keyword stuffing was the only tested optimization that reduced visibility. These findings suggest that AI search rewards content quality and supporting evidence more than keyword repetition.
What defensible pSEO looks like in 2026?
The examples above reveal a consistent pattern. Defensible programmatic SEO in 2026 comes down to three principles:
- Start with the database, not the template: Build only as many pages as you have genuinely unique, structured data to populate. Wise has currency pairs and live rates. Zapier has compatibility matrices. Zillow has MLS feeds. Without irreplaceable data per row, no template structure rescues the asset from being judged thin. The right question before building a page set is no longer “what keywords can I target?” It is “what data do I own that nobody else can publish about my entity inventory?”
- Match query type to page type: Transactional queries with live data, calculators, comparison tables, and template galleries continue to perform. Informational queries where AI Overviews dominate are no longer the right battleground for templated pages. The minimum viable pSEO build now starts at ~100 pages (submit to Search Console, wait three weeks, scale only the template variations that demonstrate traction), not 10,000. Validate before scaling; the cost of indexing thousands of pages and discovering they trip the quality classifier is months of recovery, not days.
- Build for AI citation eligibility alongside organic ranking: AI engines cite some programmatic pages (databases, calculators, structured comparisons) and skip others. The signals that lift citation rates are statistics, structured data, Q&A formatting, and E-E-A-T markers. Schema coverage and entity authority do double duty: they help organic ranking and AI citation simultaneously. A page with statistics, schema, and structured Q&A is a different asset class than a keyword-swap variant, and the prioritization queue needs to treat them differently. Across the examples in this article, the pattern is consistent: database-rich pages remain resilient, while template-rich pages continue to lose ground.
The pattern that survives in 2026 is not “programmatic SEO” or “no programmatic SEO.” It is programmatic SEO where each URL passes the test Google has applied since 2023: does this page genuinely help the user, or does it exist primarily to capture search traffic? The casualties shared the second pattern; the survivors shared the first. The algorithmic updates of the last two years did not judge how pages were created. They judged whether each page deserved to exist. That is the standard to apply before refreshing, consolidating, or removing any programmatic page.
If you’re unsure which of your programmatic pages are helping or hurting your visibility, ReSO can help. We track how your content performs across ChatGPT, Perplexity, and Google AI, highlighting the pages that strengthen your AI visibility and the gaps holding you back.
Book a call with ReSO to see where your programmatic SEO stands.
FAQ
Did Google explicitly ban programmatic SEO?
No. Google’s scaled content abuse policy targets low-value content produced at scale to manipulate rankings, regardless of whether it is created by humans or AI. Programmatic SEO itself is not prohibited.
Why did G2 and ZoomInfo lose traffic when Wise and Zapier did not?
Wise and Zapier build pages from proprietary, high-value data. G2 and ZoomInfo relied more heavily on large numbers of thin, aggregated pages with limited differentiation, making them more vulnerable to Google’s quality systems.
Will AI Overviews kill programmatic SEO?
Not entirely. AI Overviews affect many informational searches, but programmatic pages that help users complete tasks, such as calculators, live-data tools, and structured databases, remain valuable.
Should I delete my existing programmatic pages?
No. Review them first. Keep pages that offer unique data and genuine value, and consolidate or remove pages that differ only by keywords without adding meaningful information.
