The Most Common Job Title in Tech Covers 2% of the Jobs

Type "software engineer" into a job search. Whatever comes back, you have just addressed about two percent of the market.

We normalised the title on all 9,247 open roles across the 817 tech companies currently hiring, and the shape of what came out explains a frustration a lot of people have with searching for work and cannot quite name.

5,526 titles for 9,247 jobs

That is the headline. Nine thousand jobs, five and a half thousand distinct titles.

The most common one:

"Senior Software Engineer", the single most common job title in the industry, accounts for 193 roles. That is 2.1% of the board.

All ten titles above, combined, are under 10% of open roles.

The long tail is not a tail, it is the body

Here is the number that reframes it. Of those 5,526 distinct titles, 4,829 are used by exactly one company.

Which means 4,829 open roles, 52% of everything available, carry a job title that literally no other employer in our set uses.

Not variations on a theme. Unique strings. "Member of Technical Staff". "Forward Deployed Engineer". "Software Engineer, Data Infrastructure and Acquisition". Somebody's internal ladder vocabulary, published outward as if it were a category.

The market does not have a title taxonomy. It has 817 private ones.

Why search feels broken

This is the mechanical reason keyword job search underperforms, and it is worth being precise about because the usual advice ("use more keywords!") does not survive contact with these numbers.

If you search "Senior Software Engineer" you see 193 roles. You do not see the Member of Technical Staff at a lab that would have been your best offer, the Forward Deployed Engineer role that is 70% the same work, the Senior Backend Engineer at a company that just words it differently, or the "Software Engineer II" that is your level under a different ladder.

You could add synonyms. Add ten and you are still at 12% coverage, and you have now also pulled in ten titles' worth of roles that are not your job.

The half of the market with a unique title is unreachable by keyword in principle. You cannot search for a string you have never seen, and there are 4,829 of them.

Titles are not levels either

The seniority words are doing less work than they look like they are.

"Staff Software Engineer" appears at 107 companies, and at some of them staff sits above senior, at others it is a parallel IC track, and at a few it is what they call a strong senior to avoid a promotion conversation. "Member of Technical Staff", which shows up at eight companies in our set, spans everything from new grad at one to principal-equivalent at another. It is a title that deliberately carries no level information, which is the point of it.

So the title tells you neither the work nor the level with any reliability. It mostly tells you which company wrote it.

What actually works

Search the company, not the title. This is the whole reason we built Remoet the way we did, so read the following knowing we are not neutral. Companies are a stable, finite, comprehensible set. There are 1,121 of them in our data and maybe forty you would seriously want. Titles are an unbounded vocabulary invented independently by each of those companies. Picking the axis with 1,121 values over the one with 5,526 is not a preference, it is arithmetic.

Star the companies whose stack and product you want, and read everything they open. Forty companies is a list you can actually hold. When one of them posts a Member of Technical Staff role, you read it, because it is them, and you never needed to guess the noun.

Search the stack instead of the title. If you do want to filter across the whole market rather than a shortlist, technology names are far more standardised than job titles. "Go" means Go everywhere. "Staff Engineer" means six things.

Read the requirements, not the header. Obvious, and most people still filter on the title in the search results before ever opening the post. The header is the least standardised part of the document.

Let something else read them. This is the honest case for pointing an agent at job data rather than a search box: an agent can read 9,247 descriptions and reason about whether the work matches, which is exactly the task a keyword filter cannot do and a human does not have the hours for. If you want that, connect one and describe the job you want rather than guessing its name.

Caveats

Our normalisation lowercases and trims but deliberately does not merge different titles, so "Senior Software Engineer" and "Sr. Software Engineer" may still count separately where the punctuation differs. That inflates the 5,526 somewhat. It does not change the shape: even generous merging leaves the top title in low single digit percentages.

Counting a title as "used by one company" is sensitive to that same issue in reverse, so treat 4,829 as approximately half the board rather than exactly 52%.

We track 1,121 companies weighted toward product companies with real engineering orgs. A sample heavy on enterprises with formal job architectures would look more standardised than this. A sample heavy on startups would look less.

And the tail grows. Every company that invents a ladder adds titles to the world and none are ever retired.

A small prediction we are willing to be wrong about

Job titles have been getting less standardised, not more, and we think that continues, because a distinctive title is cheap and does recruiting work. "Forward Deployed Engineer" appearing at 57 companies is a title that barely existed a few years ago and spread because it sounded like something.

If that is right, keyword search gets worse from here, and the only searches that survive are the ones anchored to things with fixed vocabularies: companies and technologies.

If it is wrong and the industry converges on a shared ladder, come find this post in a year and tell us. Preferably on Discord, where we will be the ones who have to admit it.