Job Titles Have Stopped Being a Shared Vocabulary

Bars scaled to the largest value, 55,293

Count of distinct engineering job titles by how often that exact title string appears in the corpus. The first bar is the finding. 55,293 titles exist for one posting each, while only 44 titles in the entire engineering corpus are used more than fifty times.

Exact title string matches, case insensitive. To see even one engineering job in three you would need to search five thousand different title strings. Searching the ten most obvious ones shows you under three percent of what is open.

Source: Seeker job corpus, 288,411 active job postings, snapshot 2026-08-25. seekerscore.com/insights/job-title-fragmentation-2026

The scale of it

The distribution is not a long tail with a healthy head. It is almost entirely tail. 55,293 titles appear once. 6,926 appear twice. Only 44 titles in the whole engineering corpus are used more than fifty times. The top of the list looks reassuringly normal until you check the counts: Senior Software Engineer 538, Software Engineer 368, DevOps Engineer 349, Data Engineer 260, Senior Data Engineer 238, Senior DevOps Engineer 173, Staff Software Engineer 171, QA Engineer 160. Those eight titles together are under 2.5% of engineering postings. Everything else is somewhere in the other 97.5%, spread across sixty five thousand strings.

We tried to explain it away and could not

The obvious objection is that this is formatting noise rather than vocabulary. Employers append locations, requisition numbers, bracketed tags and legal suffixes, and each variant counts as a new title. So we stripped them. Removing everything inside parentheses and brackets, plus requisition and job numbers, takes the unique share from 84.9% to 82.8%. Formatting explains about two points of it. We then went further than is really defensible and stripped one trailing clause from every title, which destroys genuine information such as the team name in Senior Software Development Engineer, Prime Video Offers Expansion. Even under that deliberately aggressive normalization, 77.0% of titles are still used exactly once. The fragmentation is in the words employers choose, not in the punctuation.

It is not one job board doing it

Every large source in the corpus shows the same pattern, so this is not an artifact of one aggregator with messy data. Singleton rates run from 78.1% at the lowest to 92.4% at the highest, and no source falls below 78%. Sources that publish a single employer's own postings sit inside the same range as sources that aggregate thousands of employers. Whatever is happening, everyone is doing it.

What the unique titles actually look like

These are real titles from the corpus, each used exactly once: Network Deployment Technician, Global Network Deployment. UI Developer (.NET/React) (AVP). Senior Software Development Engineer, Prime Video Offers Expansion. Technology Product Manager AI and Data Strategy. Quality Assurance Associate (New Product Introduction). Tableau Development, Assistant Vice President. None of these are absurd. Every one of them is a reasonable description of a specific job written by someone who was not thinking about search. That is the point. Nobody is doing anything wrong, and the collective result is a vocabulary with no shared words in it.

Why this makes job search feel broken

Exact title search assumes titles are categories. They are closer to sentences. When you search Software Engineer you are matching a string that 367 postings happen to use, while thousands of postings describing the same work sit behind a string you did not think to type. The failure is silent, which is the worst property a search can have. You do not get an error. You get a short list and the reasonable impression that not much is open. Two candidates searching for identical work with slightly different vocabulary will see almost entirely different markets, and neither will know.

What to search instead

Search the things that stay constant when the title changes. Skills are far more stable than titles because a posting that needs Kubernetes says Kubernetes regardless of whether the role is called Platform Engineer, Infrastructure Engineer, SRE or Cloud Operations Specialist. Responsibilities are stable for the same reason. Role family is stable by construction. The practical habit: pick three or four titles that describe the same work, search all of them, and treat any single title search as a sample rather than an answer.

Methodology

Population is active postings in the Seeker corpus classified into the engineering role family, measured on 25 August 2026. Titles are compared as exact strings after lowercasing and trimming whitespace. The normalization ladder in the second section applies, in order, removal of parenthesised and bracketed content, removal of requisition and job identifiers of three or more digits, and finally removal of one trailing clause after a comma, dash, slash or pipe. Per source figures cover every source contributing at least 1,500 engineering postings.

What this does not show

Role family is assigned by our classifier rather than by the employer, and the engineering family is broad, so a minority of adjacent technical and operations roles are included. This measures the diversity of title strings, not the diversity of underlying jobs, and some genuinely different jobs share a title while some identical jobs carry different ones. The corpus is a sample of the labour market shaped by which employers and boards we ingest, not a census. Finally, this is a snapshot rather than a trend. We cannot yet say whether title fragmentation is increasing, because the corpus does not hold enough clean history to measure change.

Search by fit instead of by wording

This is the problem Seeker was built around. Rather than matching the title string you typed, it reads what a posting actually asks for and compares it against what you have actually done. Upload a resume and you will see roles you would never have found by search, because you would never have guessed what their employers decided to call them.

Frequently asked questions

Related Data

Guides

• Job Titles Explained →
• What Jobs Am I Qualified For? →
• Which Industries Hire Your Skill? →