Job Taxonomies

The Problem

Since Google's AI-driven spam and helpful content updates, programmatic SEO strategies that job boards relied on for years can now trigger low-quality content classifiers. Thin category pages, auto-generated combinations of location × job title × seniority that produce thousands of near-empty pages — these tactics went from effective to toxic.

At the same time, most job boards inherit taxonomies from official classification systems (ISCO, O*NET, SOC) that work for government statistics but fail at white-collar and technical roles. "ICT Professionals" doesn't match how anyone searches for software engineering jobs.

The result: you're either ranking for the wrong terms, creating pages Google won't index, or both.

How I Work

I build custom taxonomies from the ground up, using your actual data and real search demand.

Data-Driven Term Discovery. I pull candidates from six sources: your own Search Console keywords, competitor keywords via Semrush and Ahrefs, competitor site structures via scraping, official taxonomies as a baseline, normalized job titles from your own content (cleaned, deduplicated, clustered by count), and on-site search queries from your users. Each source catches what the others miss.

Taxonomy Architecture. Terms are organized into a hierarchy that balances search volume, indexing viability, and content depth. Every category page must have enough job listings to avoid thin content flags, enough search volume to justify its existence, and enough semantic differentiation from sibling pages to avoid cannibalization.

Search Page Lifecycle. Category pages aren't static — they fluctuate as jobs are posted and expire. I define the rules: when does a page with 2 results get noindexed? When does an empty page redirect? When does it come back? These transitions need to be automated and tested.

Internal Linking Strategy. Links between category pages are built on semantic similarity, not just hierarchical relationships. "Data Engineer" links to "Data Scientist" and "Backend Developer," not just to a parent "IT Jobs" page.

Implementation. I work directly with your engineering or PM team through the full cycle: analysis, scoping, requirements, build, test, rollout, and monitoring. Typical project duration is 4–6 months for one language.

Technologies & Tools

Search Console, BigQuery, Ahrefs, Semrush, custom scraping for competitor taxonomies, NLP clustering for job title normalization, ISCO/O*NET/SOC reference frameworks.

Results

• EU Job Board (Denmark): Custom taxonomy + migration — 3x organic traffic growth over 2.5 years, leading to acquisition.
• EU Niche Job Board (UK): New taxonomy replaced a broken post-migration structure — full traffic recovery from a 50% loss.
• EU Aggregator (UK): Taxonomy built for a new job board from zero — 250x traffic growth in 8 months.

A taxonomy project pays for itself in the first quarter of organic growth. Let's scope yours.