28 AUGUST 2026 · 4 MIN READ

685,153 job postings have 408,567 different titles

We counted the titles across every live posting in our index. 685,153 jobs. 408,567 distinct titles. That is roughly one new title for every 1.7 jobs — and it explains why searching for work feels broken in a way nobody quite articulates.

Most job titles are used exactly once

356,119 titles appear a single time. Not once per company — once in the entire corpus. Eighty-seven per cent of the distinct titles in the market are a one-off: invented by one employer, for one opening, and never used again by anyone.

There is no shared vocabulary. There is a long tail with almost nothing else in it.

Why that breaks the search box

Keyword search assumes you and the employer picked the same words. At this level of fragmentation you mostly did not. Somebody hiring for what you would call a growth marketer has written "Demand Generation Manager", "Performance Marketing Lead", "Growth & Lifecycle Manager" or "Member of Marketing Staff" — and a keyword search for your words returns none of them.

The jobs you never see are not hidden. They are simply filed under a phrase you did not think to type.

What that means for how you search

The usual advice — search more synonyms, set up more alerts — is asking you to do a machine's job. Guessing four hundred thousand naming conventions is not a skill anyone should need.

It is why Otto matches on meaning rather than words: your CV and the posting become vectors, and similarity is computed between them. Search "help people recover from injury" here and you get Recovery Coach, Outpatient Therapist and Recovery Mentor — none of which share a keyword with what you typed.

Method: all active postings in the Otto Roles index on 28 August 2026, grouped by exact title string. No normalisation — "Senior Engineer" and "Sr. Engineer" count as two, which is precisely the point.