Reach Hire your agent

We analyzed 63,651 job postings. Here are five things job seekers should know

Reach's matching engine ingests job postings from 22 sources: aggregator boards, company career pages on Greenhouse, Lever, Ashby and Recruitee, public APIs like Adzuna and USAJOBS, and niche boards. Between June 11 and July 27, 2026 that added up to 63,651 postings. Because we score every posting against real candidates, we see the parts of the market that job seekers usually discover one painful application at a time. Here are five of them, with numbers.

1. More than half of "remote" jobs are not remote for you

58% of remote-labeled postings restrict hiring to a specific country or region (7,832 of 13,544)

Of the 13,544 postings in our inventory labeled remote, 7,832 name a country or region requirement somewhere in the posting: "Remote (US only)", "Anywhere in Europe", "must be eligible to work in Canada". A majority of the remote market is closed to you before you write a word of your application.

The sneakiest version is the posting that says nothing at all. German-market employers, for example, are required by hiring law to add a gender marker like "(m/w/d)" to job titles. Our inventory holds 1,728 such postings, and 1,536 of them, 89 percent, name no country anywhere in the location field. They read as location-free to a job seeker in Ohio, but the employer expects candidates in Germany, Austria or Switzerland.

What to do: before investing in an application, search the posting for your country's name, "eligible", "authorized" and "time zone". If a remote posting names no location scope at all, treat that as a question to answer, not a green light.

2. When the employer writes the posting, salary appears less than half the time

31% to 47% of postings on employer-run ATS boards state a salary

Across postings scraped directly from company career pages, salary transparency is a coin flip at best: 47 percent on Greenhouse boards, 41 percent on Ashby, 31 percent on Lever. Job leads arriving through LinkedIn alert emails in our pipeline stated a salary exactly zero times out of 277.

Aggregators look far better on paper. Adzuna listings in our inventory carry salary figures 97 percent of the time, but that reflects a mix of US state disclosure mandates and the aggregator's own market estimates, not a sudden outbreak of employer honesty.

What to do: treat aggregator salary figures as an estimate of the market, not an offer. If pay is a hard constraint, filter on postings that state it, or expect to spend interview rounds discovering a mismatch. Reach users set a pay floor per persona and postings below it are gated out automatically.

3. The entry level is nearly invisible

10 to 1 senior-cue titles outnumber junior-cue titles (10,027 vs 1,025)

Titles carrying a seniority cue skew overwhelmingly upward: 10,027 postings say Senior, Staff, Principal or Lead, while just 1,025 say Junior, Entry Level, Graduate or Intern. That is under 2 percent of the whole inventory explicitly welcoming people starting out.

The flip side: most titles carry no level cue at all, and many of those roles are genuinely open to less experienced candidates. On our own platform, users who set a strict experience-level filter silently lose a large share of their matches. When we audited personas with a level filter set, roughly 1 in 3 were gating away 40 percent or more of the postings that otherwise fit them.

What to do: if you are early-career, do not filter to postings that say "junior". Apply to cue-less titles where you meet the core requirements. The word "senior" in a title is a signal; the absence of any level word is not.

4. No single job board shows you the market

~2% of postings appear on more than one of our 22 sources under the same title and company

Matching postings by exact title and company across sources, only about 2 percent of rows in our inventory are cross-listed. That measure is conservative, since the same role can appear with slightly different titles, but the shape is unmistakable: each board carries mostly inventory the others do not have. The QA-specialist board and the four-day-week board and the big aggregator barely overlap.

What to do: a job search that lives on one site samples a thin slice of the market. Either budget time to sweep several boards on a schedule, or use tooling that does the sweeping for you. This fragmentation is the reason Reach scrapes 22 sources instead of reselling one feed.

5. Part-time work is barely labeled at all

4.8% of postings carry any part-time signal (3,065 of 63,651)

Only 3,065 postings in the inventory mention part-time, per-diem or PRN anywhere in the description. Everything else is silently full-time. If you need reduced hours, the market does not just have fewer options, it has fewer findable options, because "part-time" is rarely a filterable field and almost never in the title.

What to do: search descriptions, not titles, and treat silence as full-time. Reach personas with an hours preference only match postings with an explicit part-time signal for exactly this reason.

Methodology

Counts are from the Reach matching inventory: 63,651 unique postings collected June 11 through July 27, 2026 from 22 sources including Adzuna, USAJOBS, Greenhouse, Lever, Ashby and Recruitee company boards, Himalayas, We Work Remotely, RemoteOK, Working Nomads, Jobspresso, 4 Day Week, QA Testing Jobs, Hacker News Who's Hiring threads and Craigslist. Deduplication is per source; the cross-listing figure matches exact title plus company across sources and understates true overlap. Salary presence means a structured minimum or maximum was attached to the posting. Seniority and part-time figures use word-boundary pattern matching on titles and descriptions.

You are welcome to cite these figures with a link to this page. If you want a cut of the data we did not publish, ask us at [email protected].

Browse open jobs

Data Scientist (2,258)Data Engineer (1,840)Data Analyst (1,005)Data Entry Clerk (74)

More guides

HomeGuides › We analyzed 63,651 job postings. Here are five things job seekers should know