What 421 Product Manager Job Postings Actually Ask For

Every open role on 62 company job boards was read; 421 of them carry a product manager or product management title. Each was checked against a fixed list of 44 terms.

What is reported is how many postings use a term, not how often. Nine mentions of “roadmap” in one posting counts once, exactly like a single mention in a bullet nobody reads. The ± is a 95% interval — where two rows sit closer together than that, treat them as tied.

The top four are not tools, methods or metrics. roadmap, cross-functional, prioritization, stakeholder management — in that order, 85.5% down to 57%. Every one of them describes how a product manager works with other people, and none of them is something you install.

The tool stack is nearly invisible. The most-named tool is SQL at 16.4%, and it is a query language rather than a product tool. Jira appears in 1.7% of postings, Figma 2.4%, Amplitude 1.7%, Confluence 0.7%, Looker 0.7%, Mixpanel 0.2%, Tableau 0.2%. Productboard appears in 0.

So is the Agile vocabulary. Agile is named in 8.6% of these postings, sprint in 3.1%, Scrum in 2.1%, Kanban in 0.5%. “Lean startup” or “lean product” appears in 0. A resume built around ceremonies and certifications is answering a question these postings mostly did not ask.

The generic word beats the specific one. Experimentation is named in 24.2% against 2.1% for A/B testing — roughly 12 times as many. The same pattern holds for metrics: engagement 19% and retention 15% against OKRs 3.3%, north star metric 4.5% and NPS 1.4%. Carry both forms.

An MBA is named in 5.5% of postings. Fewer than mention AI or machine learning (28.3%), and far fewer than mention prioritization (59.6%).

Want the version of this for one specific role rather than the aggregate? Run the posting and your resume through the free checker and it will name the terms that posting uses and yours does not.

Source. The Greenhouse job board API, published by companies so their own careers pages can render their listings. Nothing here was scraped and nothing is private. Point the collector at the same endpoint and you will get these numbers back; both it and the analyzer are in the repository.

Selection. Matched on the title alone, requiring “product manager” or “product management”. Internships are excluded. Nothing else is: because the title already pins the profession, a word like “developer” or “engineer” in one of these titles names the product domain rather than the job, and “Staff Product Manager, Developer Experience” is a product manager. A wider exclusion list was tried first and dropped 14 genuine PM postings for that reason.

Duplicates. One req advertised in four cities comes back as four records. Company, normalized title and content length together decide whether two records are the same posting — the rule is shared with the other studies here so their sample sizes stay comparable.

Counting. Document frequency: the share of postings mentioning a term at least once. “Discovery” is counted only in its qualified forms — product, customer, user or continuous discovery, or discovery process, work or phase. The bare word appears in roughly four times as many postings, but nearly all of that is ordinary English rather than the practice, so counting it would overstate the finding.

Three terms were dropped rather than reported. Linear reaches 0.7% loosely and 0.2% once “linear regression” and “linear algebra” are excluded — too rare to report and dirty at that size. Aha! appears in one posting, and its trailing exclamation mark makes a word-boundary match unreliable. “Platform” appears in 85% of postings but is ordinary English in a technology job description rather than something a candidate can act on, so reporting it would pad the table without informing anyone.

Limits worth stating. These are 62 mostly US technology companies hiring through one ATS vendor. They are not a random sample of product management jobs, and the mix skews to well-funded product companies. Terms outside the 44-term list are not counted, so absence here is not evidence of absence in the market — no certification terms were measured at all. It is a snapshot of open roles on 2026-09-09, not a trend.