how an EdTech startup generated 18,000 monthly sessions with programmatic SEO

James builds an online learning platform for professionals upskilling in data science and machine learning. strong product, good NPS, and a content problem: he was competing with Coursera, DataCamp, and LinkedIn Learning for every keyword that mattered.

"I can't outrank them on 'learn Python online'," he said. "they have domain ratings in the 80s and teams of content writers. I need a different angle."

he was right. the different angle was programmatic SEO — building pages at scale from structured data rather than writing content one post at a time.

the opportunity

FlowIntent's keyword research identified a pattern in how professionals search for upskilling content: they search for specific skill combinations, not generic learning categories.

searches like:

• "Python for data analysts with Excel background"
• "machine learning for software engineers without maths degree"
• "SQL for marketers transitioning to data roles"
• "statistics for product managers"
• "deep learning for Python developers"

these queries had real search volume (200–1,200/month each) and almost no purpose-built content. the big platforms had generic pages. nobody had built pages that directly addressed the "X skill for Y background" query pattern.

James's platform had structured data on 40+ skills and 20+ professional backgrounds. the combinatorial maths produced 800 potential page templates. not all of them had search volume — but FlowIntent's keyword validation identified 340 that did.

the build

the programmatic architecture used a template with 6 core sections:

1. why [skill] matters for [background] professionals (background-specific framing)
2. what you already know that transfers (reduces intimidation, improves conversion)
3. what you'll need to learn from scratch (honest about the gap)
4. the learning path (structured as a 3-phase journey)
5. time to competency (specific, not vague)
6. what jobs this opens up (commercial signal, AEO-optimised)

each page was generated from James's structured data — skill descriptions, background profiles, learning path data — with the template providing the framework and the data providing the specificity.

340 pages went live in a single deploy. FlowIntent's technical SEO audit confirmed indexation strategy: XML sitemap submitted, crawl budget allocated, internal links from category pages to all programmatic pages.

results at 4 months

pages indexed: 340/340 organic sessions (month 0): 1,200/month → (month 4): 18,400/month top-10 rankings: 0 → 127 trial signups from organic: 0/month → 340/month paid conversions from organic: 0/month → 41/month

month 4 organic traffic exceeded James's entire previous year of organic traffic. the top-performing pages were the "X for Y background" templates, with the SQL for marketers page generating 1,200 sessions/month alone at position 3.

AI visibility: 18 programmatic pages appeared in Perplexity answers for relevant queries. the structured, specific format of the pages — clear answers to specific questions — made them well-suited for AI extraction.

what didn't work first

the first 50 pages deployed had thin content — the template was too short and the background-specific customisation wasn't specific enough. Google indexed them but didn't rank them.

the fix: expanded the template to include real learning path details (specific libraries, tools, projects), added FAQ schema to each page, and increased minimum content length from 400 to 900 words. the expanded template required more structured data from James — about 2 additional hours of data input — but produced pages Google treated as substantive rather than thin.

the lesson for programmatic SEO: the template has to do real work. pages that feel generated get treated as generated. pages that feel useful — even if produced from a template — rank and earn citations.

the takeaway

programmatic SEO isn't a shortcut — it's a strategy for teams who have structured data and a clear combinatorial opportunity. not every business has both. James had both.

the 18,000 monthly sessions came from 340 pages that took 3 weeks to build. that's roughly 53 sessions per page per month from a single deployment. the equivalent in traditional content production would have required years and a team.

the opportunity is real for any platform or database business with structured data on entities, locations, skills, or comparisons. if you can enumerate the combinations, you can build the pages. FlowIntent's keyword validation tells you which combinations have demand.

related reading: Programmatic SEO — the Complete Guide — when to use it, how to structure templates, and how to avoid thin content penalties. | The 9-Section SEO Content Brief — the brief format that informed the template structure for James's 340 pages.