Programmatic SEO: How Indie Makers Build Thousands of Pages That Rank

What is programmatic SEO?

Programmatic SEO (pSEO) is creating landing pages automatically from a database and a page template, so one piece of design and code produces hundreds or thousands of pages, each aimed at a different long-tail search. A traditional SEO article targets one keyword with one hand-written page; programmatic SEO targets a pattern of keywords with one template.

Every pSEO project has three parts:

1. A head term and modifiers. The head term is the thing (“cost of living”, “remote jobs”, “timer”); modifiers are what changes (a city, a skill, a duration). Head × modifiers = your page list.
2. A dataset with a row per page and the facts that make each page different: prices, listings, ratings, specs, results of a calculation.
3. A template that turns a row into a page that genuinely answers the search — title, H1, data, comparison, FAQ and internal links.

When programmatic SEO is a good fit

• People search for the same thing with many variations (places, products, pairs, amounts).
• You have, or can build, data that differs meaningfully for each variation.
• Each page can answer the search better than a single general page could.

If you only have five variations, write five pages by hand. If you have 5,000 variations but the same sentence with a different city name on each, you don’t have a pSEO project, you have a spam problem (see pitfalls). For the wider SaaS picture, see SEO for SaaS.

Programmatic SEO examples: what 7 indie sites actually publish

Most pSEO articles repeat the same big-company examples (Zapier, Tripadvisor, Wise). We looked at indie products instead. On 2 November 2026 we fetched each site’s XML sitemap and counted URLs by template. Together they list 491,674 URLs.

How to read this: a sitemap is what a site asks Google to crawl, not what Google indexed or how much traffic it gets. Remote OK’s total also includes about 297,000 user profile pages.

What the examples have in common

• The data is the product. Remote OK’s pages are live job listings customers pay to post; TrustMRR’s are startups with revenue verified through Stripe; Nomad List’s are crowdsourced city data. The pages are a view of a database that exists for users, not for Google. In September 2024 Levels reported Nomads.com at about $61K a month and Remote OK at $29K a month (source: x.com).
• Each record gets several angles. Nomad List’s sitemap gives each of its 1,384 cities eight templated sub-pages (cost of living, coworking, people, nearby, similar, next, ranking and a guide) on top of the main city page, then compares cities in pairs.
• Tools count as content. Stagetimer’s timer pages carry about 150 words of text, but each is a working countdown — the search “1 hour 5 minute timer” wants a timer, not an essay.
• Small projects work too. Marc Lou reports 29,100 Google visitors across three micro-sites — BooksCalculator, GamifyList and WorkbookPDF — built once and left alone. WorkbookPDF is a single template with 25 pages; GamifyList’s data came from a virtual assistant he paid $150 to collect 600 gamification apps (source: marclou.beehiiv.com).

Other indie pSEO stories

• Ian Nuttall found pSEO by accident: people searched “£[salary] after tax” for countless amounts, so he fed the salary from the URL into his UK tax calculator. He went on to build many programmatic sites; the biggest sold for $500K, and his Practical Programmatic course has made more than $200K (sources: highsignal.io, practicalprogrammatic.com).
• Frey Chu built a thrift-store directory as an SEO experiment: listings scraped with Outscraper, then checked and enriched by hand against Google Maps, published as long pillar pages. It reached about 60,000 monthly visits and $2,000+ a month (source: wpgeodirectory.com).
• Danny Postma’s HeadshotPro generated 200+ city pages (“professional headshots [city]”) alongside blog posts in 2023, per a third-party breakdown (source: indiehackers.com). On 2 November 2026 its sitemap no longer listed city pages — a reminder that templated pages get pruned when they stop pulling their weight.

The 8 programmatic page patterns (and where the data comes from)

Almost every pSEO project is one of these patterns. Pick the one your data supports.

For a SaaS, the most valuable patterns are usually integrations, alternatives, use cases and free tools, because they bring people close to buying. Directories and data sites lean on locations, entities and comparisons — see our directory business playbook.

How to do programmatic SEO, step by step

1. Find a head term with enough long-tail variants

Brainstorm head terms around your product, then check the variations in a keyword tool. Marc Lou’s rule of thumb: keyword difficulty under about 20, 500+ monthly searches for the head term in one country, and at least 20 long-tail variants (source: marclou.beehiiv.com). Then Google a few variants: if small sites, forums or thin pages rank, there’s room; if every result is a giant with exactly your data, pick another pattern.

2. Build a dataset nobody else has, or that’s better organised

Unique data is the whole moat. Ways makers got theirs:

• Your own product’s records — the strongest kind (Remote OK jobs, TrustMRR revenue, Uneed listings).
• Public APIs and scraping, then manual enrichment. Frey Chu splits listing data into “need to know” facts (hours, phone) and “want to know” details (amenities, kid-friendliness), and spends his effort on the second group because that’s where competitors are thin.
• Cleaning messy data. For his $50K niche site, Ian Nuttall had a developer write Python scripts to turn messy location data into a usable database (per a write-up) (source: goodreads.com).
• Paying for collection. Marc Lou’s $150 virtual assistant for 600 apps; a few hundred dollars of outsourcing for richer venue data.
• Computation. Calculators and converters create a unique answer for every input.

3. Design one template that fully answers the search

• Put the long-tail keyword in the title and H1, naturally, once.
• Lead with the answer (the number, the list, the tool), then the detail.
• Add things a user can do on the page: sort, filter, compare, calculate, book. Marc Lou names sorting, pricing and booking options as the content worth including.
• Use conditional sections so pages with more data show more, and pages with almost none aren’t generated.
• Add internal links to related pages (nearby, similar, alternatives) — Nomad List’s “near” and “similar” pages are this idea at scale.

4. Generate pages, and write the parts a template can’t

Technically this is the easy part: a static-site generator or framework route (for example a [city] route in Next.js with static generation), WordPress with a directory plugin (Frey Chu uses GeoDirectory), or a no-code CMS fed from a spreadsheet. Keep pages light, server-rendered HTML.

The hard part is text that differs meaningfully per page: an intro that uses the row’s data, a summary of what’s notable, answers to the questions people ask about that specific item. Marc Lou used ChatGPT for pieces such as meta descriptions; the rest of each page should come from your data.

5. Publish in batches and get pages indexed

• Launch 20–50 of your best pages first, not 10,000. Watch Search Console for impressions and indexing for two to four weeks before generating the rest.
• Split large sitemaps by page type so you can see indexing per template in Search Console.
• Link every page from somewhere: category hubs, “related” blocks, breadcrumbs. Pages reachable only from a sitemap are discovered slowly.
• Earn some links to the domain. Marc Lou aims for around ten from launches and directories before expecting rankings; see where to launch and how to get backlinks.
• Request indexing for key hubs. Uneed’s founder used indexing tools that notify Google and Bing for a pSEO side project (source: solos.ventures). Google’s official Indexing API is restricted to job-posting and livestream pages (source: developers.google.com).

6. Measure, prune and improve

• After three to six months, sort pages by impressions, clicks and conversions.
• Improve the template for the whole set when a pattern underperforms, rather than page by page.
• Merge, noindex or delete pages with no impressions and no unique value.

Pitfalls: thin content, Google's scaled content policy and other ways pSEO fails

Programmatic SEO has a reputation problem because most of it is low quality. Google’s spam policies name two things that hit careless pSEO directly:

• Scaled content abuse: producing many pages mainly to manipulate rankings, with little value for users — whether written by people, AI or scripts. Examples Google lists include mass-generated AI pages, scraped or lightly reworded feeds, and stitched-together content that adds nothing (source: developers.google.com).
• Doorway abuse: pages built to rank for similar queries that only funnel visitors elsewhere — including many city or region pages that all lead to the same place (source: developers.google.com).

Google added the scaled content policy in its March 2024 update, which also targeted unhelpful content in general (source: developers.google.com).

What it looks like when it goes wrong

• Manual action. In March 2024 Google gave a “pure spam” manual action to the blog of Far & Away, an e-commerce site owned by Jacky Chou, whose AI-generated product guides he had shown off to followers; the blog was deindexed (source: futurism.com).
• Core update swings. Ian Nuttall’s relaunched programmatic site lost about half its search traffic around a core update and needed roughly six months of fixes before a later update restored it (per a write-up) (source: goodreads.com). He lists extreme patience through algorithm swings as a core principle of SEO businesses.
• Pages that never get indexed. Thousands of near-identical pages often sit in “Crawled – currently not indexed”. That’s Google telling you the template isn’t different enough.
• Empty rows. “Best rage rooms in [town]” with zero rage rooms is a page that should not exist. Only generate pages where the data is there.

Quality checklist before you publish

How long programmatic SEO takes, and what results to expect

Indexing can be quick once requested; ranking usually takes three to six months, according to Marc Lou, whose small pSEO sites settled at about 300 Google visitors a day (source: marclou.beehiiv.com). He also notes some readers reported no success with their own pSEO sites — results vary widely.

Programmatic traffic is usually low-intent per visit, so plan how it pays: ads and featured listings for directories (Frey Chu uses ads once a niche reaches roughly 10,000–50,000 monthly visitors), sign-ups for a SaaS, or affiliate links. Ian Nuttall’s rebuilt site earns mostly from display ads.

Programmatic SEO checklist

1. Head term chosen; 20+ variants with real search demand; page one beatable.
2. Dataset with a row per page and at least a few fields that differ meaningfully per row.
3. Rows with too little data excluded from generation.
4. Template answers the search above the fold and lets users do something (filter, compare, calculate).
5. Unique title, H1 and meta description per page; canonical tags; no duplicate URLs from filters.
6. Hubs and related-page links so every page is reachable in a few clicks.
7. Sitemaps split by template; Search Console and Bing Webmaster Tools set up.
8. First batch of 20–50 pages live; indexing and impressions checked before scaling.
9. A handful of real links to the domain from launches and directories.
10. A review every quarter: prune, merge or improve the weakest pages.

Want to see who’s doing this well? Browse makers like Ian Nuttall, Frey Chu and Pieter Levels, compare what products earn on the revenue pages, and check which frameworks and CMSs makers use on the stack page.