This page is time-sensitive. Every figure below is dated, and the picture is as of October 2026, using the latest US government releases available on 15 October 2026 and reports from hiring-software companies that publish their own data. Labour statistics are revised, and some of the newest numbers here are marked preliminary by the agency that published them.
The short version: there are plenty of openings, employers are slow to fill them, and the applications that reach a human have usually been sorted by software first. That changes what a good search looks like more than it changes whether jobs exist. The figures are US unless stated; the UK, Canada, India and others publish their own, and the direction is similar in many places but not identical.
What does the job market look like as of October 2026?
Steady on the surface, slow underneath. Unemployment is low by historical standards, openings are near their 2019 level, but the rate at which employers hire and workers quit is well below it.
| Measure | Latest | A year earlier | Same month, 2019 |
|---|---|---|---|
| Unemployment rate | 4.2% (Sept 2026) | 4.4% (Sept 2025) | 3.5% (Sept 2019) |
| Job openings | 7.08 million (Aug 2026, preliminary) | 6.92 million (Aug 2025) | 7.18 million (Aug 2019) |
| Hires, as a share of employment | 3.3% (Aug 2026, preliminary) | 3.2% (Aug 2025) | 3.9% (Aug 2019) |
| Quits, as a share of employment | 1.9% (Aug 2026, preliminary) | 2.0% (Aug 2025) | 2.3% (Aug 2019) |
| Median time unemployed so far | 11.5 weeks (Sept 2026) | 10.1 weeks (Sept 2025) | Not compared here |
| Unemployed 27 weeks or more, share of all unemployed | 27.1% (Sept 2026) | 23.6% (Sept 2025) | Not compared here |
All rows come from the BLS public data API, checked on 15 October 2026. The unemployment figures are from the September 2026 household survey; the openings, hires and quits rates are from the August 2026 Job Openings and Labor Turnover Survey, the latest month available.
Why is it hard to get hired when openings are near 2019 levels?
Because an opening is an intention to hire, and the hiring itself has slowed. In August 2026 employers had roughly the same number of openings as in August 2019 but hired at 3.3% of employment instead of 3.9%.
Quits tell the other half of the story. When fewer people leave their jobs, fewer seats open up behind them, and employers feel less pressure to fill a role quickly. A quits rate of 1.9% against 2.3% in 2019 describes a market where people are holding on to the jobs they have, and where people without one wait longer for a seat to open.
The queue behind each posting is longer too. Ashby, counting applications per job at mostly US tech companies, found that the average business role drew three times as many applications in January 2024 as in January 2021, and the average technical role 2.6 times as many; it is the most recent comparison of its kind we found.
The waiting shows in the duration data. More than a quarter of unemployed people in September 2026 had been searching 27 weeks or more, a larger share than a year earlier.
New graduates feel it most. The Federal Reserve Bank of New York’s tracker for recent college graduates said that in the second quarter of 2026 “The unemployment rate stayed elevated at about 5.6 percent, and the underemployment rate edged up to 42 percent.” That compares with 4.2% for the whole labour force in September 2026, a later month, so read the gap as approximate.
On the employer side, Ashby, an applicant tracking system, followed a fixed group of hundreds of companies through 2024 and 2025. It found hiring rose overall in 2025, but companies with fewer than 100 employees “reduced quarterly hiring by as much as 25%” compared with early 2024, while the largest companies drove the growth. It added that “Talent teams are interviewing significantly more candidates per hire”. More interviews per hire means a longer process for everyone in it.
How is the tech job market in 2026?
Recovering from a deep trough, but nowhere near the 2022 peak. Postings for software developers are up on a year ago; employment in the information sector is down.
- Postings. Indeed’s index of US software development job postings, published through the St. Louis Fed’s FRED database with February 2020 set to 100, stood at 78.3 on 2 October 2026. A year earlier it was 64.3. Its high point was 233.9 at the end of February 2022. For comparison, Indeed’s index of all US postings stood at 103.8 on the same October day, so software is still lagging the wider market.
- Employment. BLS counted 2.739 million jobs in the information sector in September 2026, a preliminary figure, against 2.859 million in September 2025. BLS also counted 2.355 million jobs in computer systems design and related services in September 2026, a preliminary figure, against 2.388 million a year earlier.
- What gets posted. Ashby’s July 2026 analysis of more than 200,000 postings found that engineering and data roles “have referenced AI about half the time over the last year-plus”, and that smaller companies write AI into their postings more often than large ones.
The practical reading: more software roles are being posted than in 2025, but fewer than before the pandemic, while the wider market is above that line, and a larger share of the roles ask for AI experience. If you have shipped anything that uses a model, put it in the first lines of your resume.
What has AI changed about getting hired?
Mostly the first filter. Applications are cheaper to send, so there are more of them, and employers increasingly use software to sort them before a person reads any.
Software often reads your application first
Ashby’s 2026 report on startup hiring, covering more than 1,200 venture-backed startups, found 60% of its startup customers using its AI features in the third quarter of 2025. One of those features lets a hiring team “define criteria that can help parse through evidence” of whether a candidate meets the essentials of the role. In a separate report, Ashby says one of its own staff “reviewed 1,500 applications in 6 hours using AI-Assisted App Review”. Those figures cover Ashby’s customers, not every employer, but the direction is clear.
Forms ask more questions
Ashby also found that “the number of questions per application has nearly doubled in the past four years”, across 4.8 million applications from 2021 to 2024, and that 46.8% of 2024 applications included long-form questions. The same report describes recruiters who “feel they’re stuck in a sea of AI-generated responses”. Generic AI-written answers are now the background noise a reader filters out; a specific sentence about something you built stands out more than it used to.
Fewer entry-level openings in AI-exposed jobs
Research on AI and early-career jobs by Erik Brynjolfsson, Bharat Chandar and Ruyu Chen at the Stanford Digital Economy Lab, built on ADP payroll data, was updated in August 2026. It finds “no evidence of widespread, economy-wide job displacement”, but that employment of workers aged 22 to 25 in AI-exposed occupations “now stands 19% below where it would be had it kept pace with that of their less-exposed peers”. The authors say the gap “operates primarily through reduced hiring of young workers rather than increased separations”, and they present it as an early, descriptive signal rather than proof of cause. If you are early in your career, the takeaway is that the front door for junior roles in those fields has narrowed, so other routes matter more.
Fraud checks
Ashby says “fraudulent candidates have begun creeping their way into the droves of inbound applications” and has built detection for them. Expect more identity checks and video steps, and keep your LinkedIn, GitHub and resume consistent with each other so you pass them without friction.
How long is a job search taking in 2026?
Plan for about three months, and be ready for longer. No official statistic measures a finished job search, but BLS tracks how long currently unemployed people have been looking.
That figure was a median of 11.5 weeks in September 2026, up from 10.1 weeks a year earlier. The average was 24.8 weeks, pulled up by the people who have been looking for a long time. Both measure searches still in progress, so they are a floor for the people counted, not a forecast for you. People switching while employed are not in these numbers at all.
What shortens it is mostly within your control: how well each application matches the role, how early you apply, and how many of your efforts reach a person rather than a queue. If you have already sent a large number of applications with little back, the 200-applications diagnostic walks through what to check first.
What works in a 2026 job search?
A narrower search with more direct contact. Each item below answers one of the changes above.
- Target fewer roles, and match them visibly. Pick one or two role types. Use the posting’s own words for skills you really have, in the opening lines of your resume, so both software and a skimming human find them.
- Apply in the first week. Older postings have older piles. Sort by date and skip anything you would be applicant number one thousand on.
- Answer screening questions as a filter, not a formality. Location, work authorisation and notice period answers can end an application before anyone reads it. Skip roles whose hard requirements you miss.
- Bring proof of work. A repository, a short case study, a metric you moved. AI can write a fluent cover letter for anyone; it cannot invent your shipped project.
- Use referrals where you have them. Ashby found referrals held steady at about 18% of hires from 2021 to mid-2025 and stayed near 17% regardless of how many inbound applications a job got. A warm introduction remains one of the few channels that does not grow more crowded as application volume rises.
- Contact the person who hires. At large companies, that is a hiring manager. At a startup of ten people, it is usually a founder, and a short, specific email reaches them before any screening software does.
- Follow up once or twice. Silence is often a busy inbox, not a no.
- Track what gets answers. After two weeks, move time toward whichever channel produced conversations.
None of this is new advice. What is new is how much worse the alternative performs: a high-volume, low-match search now meets more applicants per posting and more software screening than it did in 2021.
Where do startups fit when large-company pipelines are this crowded?
They are the clearest case for contacting people directly. Small companies lean on sourcing and their founders’ own networks, and some make their first hires before they write a job post.
Ashby’s July 2026 data supports the sourcing point from another angle: tech roles with AI in the title “are less likely to be filled from inbound or referral applications” and are filled more often through agencies and sourcing. Its startup report found remote startup roles received 42% more inbound applications than in-office ones, so a remote posting at a startup is crowded too. A note to a founder at a company that just raised money reaches a person with a budget and an empty org chart, often before the posting exists.
The manual version is simple and slow: watch funding and launch news, find the founder, write a short note that connects their news to your work, send it from your own email, follow up. Finding startups before they post roles covers the first step, while the founder cold-email guide gives you wording to adapt.
Backdoor runs that routine for you. Its morning sweep covers funding announcements, TechCrunch, Crunchbase, YC, Product Hunt, GitHub, Wellfound and company careers pages, and it keeps the companies that suit your background. For each pick it hunts down the founder’s email and writes up why your experience lines up with what they are building. You approve, edit or skip every draft, the approved ones leave from your Gmail, and as many as three follow-ups stay in the same thread, stopping when someone answers. Its focus runs from pre-seed through Series C, and it tells people who want to blanket large employers with applications that it is not for them, which fits the split this page recommends: applications for the big employers, direct notes for the small ones.
If you would rather let founders come to you, YC Work at a Startup and Wellfound let you build one profile that startup founders can search and contact directly. Neither costs job seekers anything.
A first month of job searching in October 2026
Week 1: narrow and fix. Choose two titles: product analyst and growth analyst. Rewrite the resume’s first third around one result with a number. Answer the standard screening questions once in a notes file so every answer is consistent.
Week 2: apply early, ask for referrals. Ten applications, each to a role under a week old whose must-haves you mostly meet. Message five former colleagues at companies that are hiring and ask whether they would refer you.
Week 3: add direct contact. Pick twelve startups that raised a seed or Series A round in the past two months. Send each founder a four-sentence note tied to their product, with a link to a short analysis of their public metrics.
Week 4: follow up and count. Follow up on founder notes that got no answer. Count replies by channel: applications, referrals, founder notes. Shift next month’s hours toward whichever produced interviews.
The numbers are illustrative. Change the titles and the proof to your field, but keep the structure: one week to fix the basics, then a weekly mix of early applications, referrals and direct notes, measured by replies rather than by how many forms you sent.
If a layoff is what started your search, laid off and can’t find a job covers the first two weeks and a weekly routine for the months after.
Checking these numbers yourself next month
The BLS series above are monthly, the New York Fed says its graduate unemployment data are “updated quarterly, generally in February, May, August, and November”, and the Indeed indexes on FRED are daily series. If you are reading this after October 2026, check the latest releases before relying on any figure above, and look for the direction of change rather than a single month’s number.
How we chose
We compare tools on what their own public pages say, checked on the dates shown, and label every claim. How we compare tools.
FAQ
how to get a job in 2026
Apply to fewer roles you clearly match, ideally within the first week of posting, and lead each application with evidence of work you have done. Ask for referrals where you have a connection, and contact hiring managers or founders directly at smaller companies. US data shows employers hiring at a slower rate than before the pandemic, so a narrow, well-aimed search beats a wide one.
how to get a job in 2026 as ai transforms hiring
Assume software reads your application before a person does, so state the must-haves from the posting in plain words and answer screening questions carefully. Ashby, an applicant tracking system, reports that 60% of its startup customers used its AI features in Q3 2025 and that application questions per job nearly doubled from 2021 to 2024. The channels AI screening touches least, referrals and direct notes to the person hiring, are worth more time than they used to be.
job market 2026 tech
Tech hiring is recovering slowly from a low base. Indeed's index of US software development postings stood at 78.3 on 2 October 2026, up from 64.3 a year earlier but still below its February 2020 baseline of 100, while BLS counted fewer jobs in the information sector in September 2026 than a year before. Roles that mention AI are a growing share of what is posted.
how long to get a job in 2026
There is no official figure for the length of a completed job search, but BLS reports how long unemployed people have been looking. In September 2026 the median was 11.5 weeks, and 27.1% of the unemployed had spent 27 weeks or longer searching, against 23.6% a year earlier. Plan for about three months and build a routine that holds up longer.
why is it so hard to get a job in 2026 when there are still millions of job openings
Because openings are not the same as hires. In August 2026 US employers listed about 7.1 million openings, close to August 2019, but hired at a rate of 3.3% of employment against 3.9% in August 2019, and fewer people quit to free up roles. Each opening also draws more applicants: Ashby's data showed applications per job roughly tripling between 2021 and early 2024, so the queue behind every posting is longer. Backdoor is one way around the queue at startups: it finds founders at companies that just raised and emails them for you once you approve the note.
Sources checked
Show 20 sources
- Unemployment rate (LNS14000000) | BLS, checked
- JOLTS job openings (JTS000000000000000JOL) | BLS, checked
- JOLTS hires rate (JTS000000000000000HIR) | BLS, checked
- JOLTS quits rate (JTS000000000000000QUR) | BLS, checked
- Median weeks unemployed (LNS13008276) | BLS, checked
- Share unemployed 27 weeks and over (LNS13025703) | BLS, checked
- All employees, information sector (CES5000000001) | BLS, checked
- All employees, computer systems design and related services (CES6054150001) | BLS, checked
- Average weeks unemployed (LNS13008275) | BLS, checked
- Software Development Job Postings on Indeed in the United States | FRED, checked
- Job Postings on Indeed in the United States | FRED, checked
- The Labor Market for Recent College Graduates | Federal Reserve Bank of New York, checked
- Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence | Stanford Digital Economy Lab, checked
- 2023 Trends Report | Applications Per Job | Ashby, checked
- Did hiring pick up in 2025? | Ashby, checked
- The State of Startup Hiring | Ashby, checked
- Are jobs asking more application questions? | Ashby, checked
- How many hires really come from inbound? | Ashby, checked
- The Rise of AI in Job Postings | Ashby, checked
- Backdoor: AI job agent that emails startup founders for you, checked