
Zapier is the first name on almost every list of programmatic SEO examples. Build a page for every app, then a page for every pair of apps, and watch the long-tail traffic roll in.
So I tested it. On October 9, 2026, I pulled US search results for six integration searches, the kind those case studies love to quote. I expected at least a few first-place finishes. Zapier’s integration pages ranked first for none of them, and for one search Zapier wasn’t in the top 20 at all.
That doesn’t mean Zapier programmatic SEO failed. It means the version people keep copying is a few years old. Below is what the system looks like now, what’s broken, and what a SaaS team should take from it.
Quick answer: Zapier programmatic SEO is a database-driven page system. Zapier generates a page for each app it supports, a page for each pair of apps, and a page for each pre-built workflow (a “Zap template”). Every page is filled with data only Zapier has: the triggers and actions each app supports. In 2026 those pages still rank, but app makers’ own pages and Google’s AI Overviews now sit above them on many integration searches.
Three page types do the work.
The pages work because of the data. Every pair page lists the real triggers and actions both apps support through Zapier, and nobody else can publish that list. Zapier’s own blog post on programmatic SEO (last updated February 2025) says as much: the triggers and actions are the unique data behind the pages.
The crawl control is tidy too. Zapier’s robots.txt blocks any URL with a query string (Disallow: /*?) and blocks landing-page variants under /apps/. It lists 11 sitemaps, split across templates, the blog, automations and other sections. Most SaaS sites won’t need crawl control at this scale, but the idea carries over: keep bots off URL variations you don’t want indexed. That’s a core part of technical SEO on any large site.
Third-party traffic numbers are estimates, and six searches are a sample, not a census. Read the direction, not the decimals.
Zapier’s February 2025 post says Zapier would likely rank at or near the top for a search like “google sheets integrations.” The pair pages are the heart of Zapier programmatic SEO, so that’s where I started. Here’s what the results looked like this month.
| Search (US) | Est. monthly searches | Zapier’s best organic rank | What ranks above Zapier |
|---|---|---|---|
| slack trello integration | 480 | Not in the top 20 | Reddit, YouTube and Appzi hold the top three organic spots |
| gmail slack integration | 260 | 4th (pair page) | Google Workspace Marketplace, two Slack pages |
| google sheets slack integration | 110 | 2nd (pair page) | A Slack help article |
| google sheets integrations | 90 | 3rd (app page) | Make, Google Cloud docs |
| typeform hubspot integration | 70 | 6th (pair page) | Typeform, HubSpot, HubSpot Community, a guide from Default |
| google sheets notion integration | 40 | 6th (a blog post; no pair page in the top 20) | Two Notion pages, Reddit, Google Workspace Marketplace, n8n |
All six results pages also showed an AI Overview.
App makers caught up. Slack, Notion, Typeform and HubSpot now run their own integration and marketplace pages, and Google often ranks the app maker’s own page above Zapier’s for “connect my app” searches. Competitors copied the URL pattern too. Make and n8n use the same app-to-app integration structure, and Make outranks Zapier on “google sheets integrations.”
One result bothered me more than the rest. Zapier’s own blog post uses a Google Sheets and Notion search as its example of pair pages ranking near the top. For nearly the same search today, no Zapier pair page shows up in the top 20. I also ran the exact phrase from that post, “google sheets and notion integration.” Zapier’s best result there is a community forum thread, the second organic result.
If the pair pages slipped, where does the traffic come from now? Ubersuggest’s list of Zapier’s top 40 US pages by estimated organic traffic contains zero integration pages. It’s the homepage plus blog posts like “best wireframe tools,” “best AI chatbots” and “best note-taking apps.”
The same tool estimates Zapier’s monthly search traffic fell from about 22.2 million in October 2024 to about 3.7 million in September 2026. Its count of ranking keywords dropped from about 1.5 million to about 310,000.
I’d handle those numbers with care. The monthly series is jumpy (about 7.9 million in March 2026, about 2.9 million in July), and part of a sixfold drop could be how Ubersuggest counts. But the direction matches the results pages: the pair pages in Zapier programmatic SEO rank lower than the case studies claim, and the listicles carry the load.
Older ranking counts from 2021 still show up in many Zapier SEO strategy write-ups. I couldn’t verify them. And even if they were right then, they describe a different search engine. If you’re staring at a traffic chart that fell off a cliff, the cause is rarely one thing, and I walk through how to find out why traffic dropped in a separate post.
This is the cost side of Zapier programmatic SEO that the case studies skip. Zapier’s own post warns that more pages mean more upkeep, and its pages show why. On one pair page and one template page, I found:
None of this will sink a domain with Zapier’s authority. On a site with a domain authority of 20, the same mess across 2,000 pages is how you end up staring at “Crawled, currently not indexed” in Search Console. If your site already has pages but no traffic, start there.
Older case studies miss this, and I think it’s the smartest thing Zapier programmatic SEO is doing right now.
The Gmail + Slack pair page has a section called “For AI agents & developers.” It explains how to call the same Gmail and Slack actions through Zapier’s MCP server or its TypeScript SDK.
The robots.txt file explicitly allows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot, Google-Extended and DuckAssistBot, and a Content-Signal line permits AI training, search and AI input. Then there’s zapier.com/llms.txt, which points to markdown copies of the templates and automations pages, an AI catalog file, an MCP server card and an AGENTS.md file.
None of this is for Google. Google’s documentation on AI features says you don’t need special files, AI text files or extra markup to appear in AI Overviews or AI Mode. Zapier is building for agents that act, and for an agent the question is which tool gets called, not which page ranks. The integration catalog that once fed Google now also feeds AI agents through MCP.
If you’re planning programmatic pages in 2026, put your effort into the data. Pages come and go. If you want help building for both Google and AI agents, that’s what my AI SEO work covers.
To be fair, a lot still holds up. The data on each page is real and specific, and nobody can copy triggers and actions from somewhere else. Crawl control keeps bots on the pages that count. Internal linking is dense: the template page I checked links to about 59 related templates and back to both app pages, and it carries an A to Z app index.
And Zapier still has a page on page one for five of the six searches I checked, four of them integration pages. It just isn’t first.
These are the takeaways I’d give a SaaS founder who points at Zapier programmatic SEO and says “let’s do that.”
Planning programmatic pages for a SaaS product and want a second opinion before you build thousands of them? Tell me what you’re working with through the contact page, and I’ll tell you what I’d test first.
You don’t need Zapier’s budget to run this audit. My free Technical SEO Auditor crawls your site and flags the technical issues behind most indexing problems. The GEO Checker shows which AI crawlers your robots.txt blocks and whether your content needs JavaScript to load.
And if you’re a SaaS team weighing programmatic pages and want a second opinion before you build 5,000 of them, that’s the work I do in SaaS SEO.
It’s Zapier’s system of generating search pages from its own app database: one page per app, one per app pair, and one per workflow template. Each page lists the triggers and actions available through Zapier, which is data competitors can’t copy.
Partly. The pages still reach page one for many integration searches, but in my October 2026 check they ranked first for none of the six searches I tested. App makers’ own pages, Make, Reddit and AI Overviews now take the top spots.
Yes, but study the 2026 version. Zapier programmatic SEO now shows both sides: why unique data works, and what happens when app makers, competitors and AI Overviews crowd the same searches.
Zapier doesn’t publish a page count. Its site says it connects 9,000+ apps, and its own programmatic SEO post says only “thousands upon thousands” of pages. Treat any exact figure in a case study as an estimate.
Yes, if the pages don’t carry unique data. Google’s spam policies treat mass-produced pages with little value as scaled content abuse. Start with a few dozen pages, check they get indexed and earn clicks, then expand.
Not for Google. Google says no special files are needed to appear in its AI features. An llms.txt file may help AI agents and developer tools find your docs, which is why Zapier has one. Treat it as optional and cheap, not as a ranking factor.
SEO cost in India runs from ₹8,000 to ₹2,00,000 a month for most businesses. Enterprise retainers go well past that....
Read Full Article →Most "traffic dropped after migration" post-mortems stop at redirects. Here's the fuller checklist I actually run, and why the drop...
Read Full Article →This site uses cookies for essential functionality and basic analytics. See the Privacy Policy for details.