Programmatic SEO examples: six live page sets, their URL patterns and the data on each page

Many lists of programmatic SEO examples give you a famous brand and a traffic estimate you cannot check, which makes them hard to learn from. This article gives you what you can check instead: the URL pattern, the data that fills each page, and the reason each page deserves to exist on its own. Every page below was opened on 13 October 2026, and each description covers only what that page showed.

Three questions do most of the work when you study someone else's set:

• What changes in the URL? The variable tells you the dataset: an app, a city, a currency pair, a product pair, a category.
• What changes on the page? In a weak set, only the name in the heading changes. In a strong one, most of what you read comes from the row behind the page.
• What would be lost if the pages were merged? If one general page could answer every variation as well, the set has no reason to exist.

If you want the method itself first, the guide to programmatic SEO explains how it works and where it fails. Here the focus stays on six sets you can open yourself.

Pattern: zapier.com/apps/[app]/integrations/[app], one page per pair of apps Zapier can connect, such as Google Sheets and Slack.

Data: Zapier's own catalogue of what each app can do. The Google Sheets and Slack page lists both apps' supported triggers and actions (a new spreadsheet row, a message posted to a channel), ready-made workflows for the pair, such as sending Slack messages whenever Google Sheets rows are updated, and an FAQ about connecting the two. Request the same address as Markdown, as an AI agent can, and it returns a plain list of both apps' triggers and actions with their inputs.

Why each page earns its place: someone searching for the pair wants to know whether the two apps connect and what they can automate between them. That answer differs for every pair, and Zapier holds it. The set also deals with the obvious duplicate: /apps/slack/integrations/google-sheets and /apps/google-sheets/integrations/slack both load, and both declare the second as the canonical URL, so search engines are asked to treat them as one page.

The lesson scales down. Say your invoicing app connects to 30 tools, but for 22 of them all you can say is that a connection exists. Build pages for the eight where you can show the setup, what syncs and the limits, and list the other 22 on one integrations hub until they have more to say.

Pattern: numbeo.com/cost-of-living/in/[City], one page per city.

Data: prices submitted by contributors, grouped into restaurants, markets, transport, utilities, childcare, clothing, rent and salaries, each with an average and a range. The Leeds page says the city had 644 entries in the past 12 months by 55 different contributors, gives the date of the last update, estimates monthly costs excluding rent for a single person and a family of four, and links to nearby cities with their distances.

Why each page earns its place: what a meal or a one-bedroom flat costs really does differ by city, and the page tells you how much evidence sits behind its numbers. The template also changes when that evidence is thin. The Truro page says some data are estimated due to a low number of contributors, asks residents to add prices, and points to nearby cities with a strong community of contributors, such as Plymouth and Exeter.

Indeed's salary pages add a second variable: uk.indeed.com/career/[role]/salaries/[City]. The page for software engineers in London gives an average based on 2.6k salaries taken from job postings on Indeed in the past 36 months, the date it was updated, the highest-paying places nearby, related roles from junior to staff, and current openings: the obvious next step for anyone checking a salary.

Pattern: wise.com/gb/currency-converter/[from]-to-[to]-rate, one page per currency pair, repeated on Wise's regional sites, which each page links to by country and language.

Data: the live mid-market rate in a converter, a chart running from 48 hours to five years, the high, low, average and change over the last 30 and 90 days, and conversion tables in both directions, from 1 to 10,000. Some sentences are written from the numbers: the page states the 30-day high and low and what that makes the 30-day average.

Why each page earns its place: the searcher wants today's rate, and often wants to know whether today is a good day to convert. The rate moves constantly, so a page per pair that updates itself answers better than any written article could. Each page also links to the top pairings for the same currency and to sending money, so it leads into the product rather than ending at a number.

Pattern: rtings.com/headphones/tools/compare/[product]-vs-[product]/[id]/[id], one address for any two headphones RTINGS has tested.

Data: lab results for both products from the same test bench, grouped under sound, design, isolation, microphone, active features and connectivity. For the Sony WH-1000XM5 against the Bose QuietComfort Ultra Headphones, the page adds a written verdict: the Bose headphones are more comfortable and better built, with a slight edge in noise isolation, while the Sony headphones support the LDAC codec and can be used passively when wired.

Why each page earns its place: measurements nobody else has, set side by side for the exact pair a buyer is weighing. The set also limits itself. Any two tested models can be compared, but the page for a pair with no written verdict, the Audio-Technica ATH-ANC70 and the AKG K391-NC, carries a noindex, follow robots tag and a one-line description. The Sony and Bose page has no such tag, and its reversed address declares the Sony-first page canonical. In the pairs we checked, visitors could compare anything, while search engines were pointed only at the pair with something written about it. The guide to indexing programmatic pages covers when to use each of those signals.

Pattern: canva.com/resumes/templates/ for the category, canva.com/resumes/templates/[style or role]/ beneath it, and one page per template under canva.com/templates/.

Data: Canva's own template library. The teacher resume page showed 50 of 355 templates, linked to sibling pages by style (professional, modern, minimalist) and by role or stage (college, tech, sales), and answered three questions for its own category, such as how long a teacher resume should be.

Why each page earns its place: someone searching for teacher resume templates wants to choose from real designs, and the page has hundreds. The detail worth copying is the FAQ. It is written for teachers rather than shared across every resume category, so even the block that usually ends up as boilerplate answers this page's search. Directories work the same way: each category page is only as good as the inventory and the category-specific help behind it. The guide to programmatic SEO templates shows how to plan blocks like these.

Set the six side by side and the same habits show up:

• The data is theirs, or kept live by them. An integration catalogue, contributors' price reports, job postings, lab tests, a template library; even Wise's rates, which anyone can look up, arrive with their own history and a way to act on them. That is why copying the URL pattern alone gives you the template without the reason it works.
• They show where the numbers come from. Numbeo and Indeed say how many reports sit behind a figure and when it was updated; Wise shows the range over 30 and 90 days; RTINGS names its test categories.
• They treat thin rows differently. Numbeo flags estimated figures and sends readers to better-covered cities nearby; RTINGS kept the pair it had written nothing about out of the index.
• They collapse duplicates. Zapier and RTINGS point the reversed pair to one canonical address.
• They link sideways and onwards. Nearby cities, other pairs and sibling categories, plus a next step: build the workflow, send the money, apply for the job, open the template.

One honest limit: none of these pages shows how much search traffic its set earns, and every one of these sites has an established brand behind it. Judge an example by what each page contains, and expect a newer site using the same pattern to need data at least as good.

If you would rather start from the searches your own site already appears for, AutoPSEO's content engine reads your Search Console data, finds searches with no page about them, sets aside the ones your pages already cover, and writes articles in your site's format and languages. The topics documentation shows where those topics come from.