Senior-level positions accounted for 69.3% of software development job postings in Q1 2026, against about 14% of postings across all US occupations [N22a">N22a]. Entry-level postings fell 7.5% year over year as of May 2026 [N22]. The developer market did not simply shrink. It inverted toward seniority. That inversion is the first of the developer hiring statistics on this page.
Key takeaways
This is a reference page for software developer hiring statistics in 2026. Every figure below carries its number, the organization that published it, and the period it covers, so a sentence can be lifted whole without losing its meaning. Each was verified against the publisher's own page, not a secondary write-up.
The scope is developer-specific where the published data allows it and broader tech hiring statistics for 2026 where it does not — every entry says which.
Where sources conflict, both appear. There is a live disagreement inside the US Bureau of Labor Statistics' own estate about the median developer salary, and two Ashby reports differ by a day on time to fill. Neither is averaged away here.
This page is the numbers. Our guide to the true cost of hiring a developer is the argument built on top of them.
The sharpest single statistic in developer hiring is a mix, not a volume. In software development, senior-level positions accounted for 69.3% of job postings in Q1 2026. Nationwide across all occupations, only about 14% of postings were for senior positions and almost half, 46%, were entry-level [N22a">N22a].
Indeed Hiring Lab published the finding on 23 July 2026. It classifies seniority by algorithmically reading the responsibilities, experience, education and certifications described in a posting rather than the job title, and reports roughly 95% accuracy against manual sampling [N22b">N22b].
The same analysis found senior-level postings up 14.7% year over year as of May 2026, while entry-level postings declined 7.5% year over year [N22]. For a team deciding which role to hire first, that mix matters more than the headline volume.
Volume tells a more complicated story, and it needs two numbers to be honest. US software development job postings on Indeed grew by almost 15% between the launch of Claude Code in late February 2025 and July 2026, a period in which job postings overall declined by 7% [V5, V5a]. That sounds like a boom. It is a rebound off a deep trough: those same postings remain about 27.5% below their pre-pandemic level, against a 1 February 2020 baseline of 100 [V5d">V5d]. Both come from the same Indeed Hiring Lab post of 8 July 2026, and quoting either alone misleads.
Indeed decomposed that increase two ways. Senior roles account for 71% of the growth in software development postings between May 2025 and May 2026 [V5b">V5b], and 37% is attributable to jobs that mention AI in their title [V5c">V5c]. These are two views of the same increase, not two separate slices, so they cannot be added.
One caution runs through all of it: postings are advertisements, not hires.
The three "time" metrics are not interchangeable. Time to first fill is the interval from a job being opened to the first hire being made; the median for technical roles is 75 days [V1]. The interview loop is the interval from a candidate's first interview to their last; the median for technical roles is 17.9 days [V1c">V1c]. Note that Ashby's own wider "time to hire" concept is broader still — it describes it as the sum of many steps across the hiring process — so this figure covers the interview phase only. Time to hire under a third definition, used in Ashby's recruiter productivity report, has a median of 40 days for technical roles [N2]. Same vendor, three constructs. None can be substituted for another.
Three different published numbers describe "how long hiring takes", and they measure three different intervals. Printing them together without definitions is how a 75-day process gets reported as an 18-day one.
Time to first fill, defined as the time from a job being opened to the first hire being made, has a median of 75 days for technical roles and 60 days for business roles across Ashby's customer base [V1]. That comes from Ashby's recruiting operations benchmarks report, drawn from 54 million applications and 93,000 jobs, January 2021 through March 2026 [V1b">V1b]. Broken out by seniority within the same dataset: junior roles 52 days, mid-level 63 days, senior 71 days, with senior roles taking 37% longer to fill than junior ones [V1a">V1a].
A second Ashby report, built on a larger and differently-scoped dataset of 109 million applications and 247,000 jobs [N3], puts median time to first fill for technical roles at 76 days [N1]. The one-day gap is a dataset difference, not a measurement dispute. This page uses 75 days from the operations benchmarks report [V1] and does not average the two.
Time to hire, defined in that same operations benchmarks report, is a different interval again: hired technical candidates take a median 17.9 days from first interview to last, and business candidates 14.4 days [V1c">V1c]. Ashby describes its wider "time to hire" concept as the sum of many steps across the hiring process, so treat this number as the interview phase only. It is not comparable to the 75-day time-to-first-fill figure and does not replace it.
Time to hire under a third definition, used in Ashby's recruiter productivity report, has a median of 40 days for technical roles in Q2 to Q3 2025, 75th percentile 57 days; business roles 30 days median, 46 at the 75th percentile [N2]. Same vendor, same word, different construct. Always say which one you mean.
Scheduling takes a real slice of the calendar. Across 2.8 million candidates who received at least one interview and more than 5.1 million interview events between Q1 2020 and Q2 2024, the median gap from availability request to interview start was 6 days, and about 65% completed a first interview within 7 days [N17, N17a].
At startups under 25 employees, time to hire drops by almost 30% when a recruiter is involved: 62 days without one, 42 days with [N18].
All of these describe Ashby's customer base, which skews venture-backed, tech-heavy and modern-ATS. They clear the bar because Ashby publishes its sample size and its definitions, not because they are a census.
What a 75-day median does to a delivery plan is a separate problem, and one we've written about in why teams can't hire engineers fast enough.
Applications went up and interviews went down. Across Ashby's customer base, applications per hire rose from roughly 100 in early 2021 to a peak of 319 in Q4 2024, settling at 291 in Q1 2026, while the application-to-interview rate fell from 7–8% in 2021 to between 3.6% for technical roles and 4.7% for business roles in Q1 2026 [N16]. More candidates, a smaller share of them ever speaking to anyone.
Pass-through by stage, same dataset: 35% of candidates clear the screen, 24% clear the onsite, 95% clear post-onsite steps, and 81% clear the offer stage. Referred candidates do better at every gate, at 52%, 36% and 85% respectively [N9].
On offers, the widely-quoted headline is dated. Ashby's offer acceptance report covers January 2021 through March 2024 and finds an average offer acceptance rate of 78% across that three-year period, with technical roles at 73% and business roles at 84%, drawn from 230,000 applications that reached the offer stage [N7, N7a]. Year by year it ran 77% in 2021, 76% in 2022 and 81% in 2023 [N7b">N7b]. That dataset ends in March 2024 and should carry the date whenever it is cited.
The newer March 2026 dataset breaks acceptance down by source instead, which is more useful anyway: technical referrals accept at 84%, technical sourced candidates at 76%, and business referrals at 89% [N8].
Across the whole funnel, the number most leaders actually feel is offer conversion: application to accepted offer. In Q1 2026 that was 7.3% for technical roles and 10.4% for business roles [N10].
Startups run different numbers. Offer acceptance hovers around 80%, and remote roles show a 9% higher acceptance rate than in-office ones, across more than 1,200 venture-backed startups, 32,000 hires and 11 million applications [N11, N11a].
The cost of hiring a developer is paid in your engineers' calendars before it is paid in salary. Technical roles now average 17.6 interviews per hire, up 52% from roughly 11 in 2021; business roles average 11.7, up 36% from about 8 [N12]. Those interviews cluster into 5 interview events per technical hire and 4 per business hire [N14], across Ashby's customer base from January 2021 through March 2026.
Converted to time, in Q1 2026 a technical hire consumed 23.3 interviewer-hours against 12.2 for a business hire. Engineering roles ran 24.7 hours and data roles 24.9, while customer support took 8.9 [N13]. Roughly three working days of engineering time per technical hire, most of it spent on people who never join.
Take-homes are rarer than the discourse suggests: only 13% of hires included a take-home component [N15].
Whether all that interviewing buys better decisions is a separate question. The current reference point is Sackett and colleagues' 2022 meta-analysis in the Journal of Applied Psychology. It re-examined how personnel-selection validity had been corrected for range restriction, and found that "most of the same selection procedures that ranked high in prior summaries remain high in rank, but with mean validity estimates reduced by .10–.20 points", and that "structured interviews emerged as the top-ranked selection procedure" [N19]. It supersedes the 1998 Schmidt and Hunter estimates that most hiring blog posts still quote. Per-method coefficients from the 2022 revision are not reproduced here.
Developer salary statistics start with a conflict. Two BLS-sourced salary figures for software developers are both in circulation and both real. O*NET reports a median annual wage of $135,980, or $65.38 an hour, for Software Developers, and attributes it to Bureau of Labor Statistics 2025 wage data [V3]. The BLS Occupational Outlook Handbook page for the same occupation, last modified 28 August 2025, still carries the earlier vintage: a median annual wage of $133,080 in May 2024, with the lowest 10% earning under $79,850 and the highest 10% earning more than $211,450 [V3b">V3b]. This page treats $135,980 as the current figure because it reflects newer wage data. The two are not measuring different things. One page has been refreshed and the other has not.
Which occupation group a number covers matters as much as its vintage. Software Developers alone (SOC 15-1252) employed 1,693,800 people in 2024, with 115,200 projected job openings for 2024–2034, using O*NET's own label and growth rated "much faster than average" [V3a">V3a]; BLS Employment Projections put that group at 1,961,400 by 2034, a rise of 267,700 or 15.8% [N4]. The broader group "Software Developers, Quality Assurance Analysts, and Testers" is a different set: 2024 median pay $131,450 a year, 1,895,500 jobs, and a 15% job outlook for 2024–34 with employment change of 287,900 [V3c">V3c]. Mixing the two is the most common citation error here.
Salary, agency fee and employer cost are three separate quantities. The $135,980 median is employee gross salary [V3]. Contingency developer recruiting fees run 18% to 25% of the hire's first-year base salary, billed when someone starts, according to recruiting firm kore1 in June 2026 [V2]; that is a vendor's price, not an internal cost-per-hire. Employer compensation cost is different again: for private industry in March 2026, BLS Employer Costs for Employee Compensation put total compensation at $46.60 per hour worked, of which wages and salaries were $32.60, or 69.9%, and benefits $14.01, or 30.1% [V4]. For civilian workers the total was $49.32 per hour worked, split $33.72 wages and $15.60 benefits [V4a">V4a].
Dividing those two published BLS figures, $46.60 ÷ $32.60, gives about 1.43. That arithmetic is ours, not a published benchmark, and it describes all private-industry work rather than developers specifically. Read as a rule of thumb, budgeting a developer at salary plus roughly 43% again in employer compensation cost is defensible; citing "1.43×" as though someone published it is not.
Anyone weighing that total against a subscription developer cost should compare like with like: gross salary plus employer compensation cost, never salary alone.
There is no official turnover rate for software developers, and the nearest official proxy is an industry figure rather than an occupational one.
BLS JOLTS put the quits rate in the Information sector at 1.1% in June 2026, down 0.1 percentage points month over month, against 2.2% for total private and 4.2% for leisure and hospitality [N20]. Information is an industry classification: it includes publishing, telecommunications and media, and it excludes the many developers employed in finance, retail and manufacturing. It is not developer turnover. It is the closest thing published.
Whole-economy context for the same month: 7.4 million job openings (4.4% rate), 5.3 million hires (3.4%), 5.4 million separations (3.4%), 3.2 million quits (2.0%) [N21].
The Southeast Asian numbers that circulate as "2026 attrition projections" are neither projections nor developer-specific. Aon's 2025 Salary Increase and Turnover Study surveyed more than 700 businesses across Indonesia, Malaysia, the Philippines, Singapore, Thailand and Vietnam, fielded July to September 2025 [V6]. Its attrition figures are 2025 actuals across all industries: Philippines 20.0%, Singapore 19.3%, Malaysia 18.2%, Thailand 17.2%, Indonesia 15.0%, Vietnam 15.0%, regional overall 17.5% [V6a">V6a]. What Aon projects for 2026 is pay, not attrition: an average 5.3% salary increase across the region, from 7.1% in Vietnam down to 4.3% in Singapore [V6b">V6b].
Every credible remote figure available here is worker-reported, which changes what it can support. In Stack Overflow's 2025 Developer Survey, 32.4% of responding developers said they work fully remote, 19.9% hybrid leaning in-person, 17.9% fully in-person, 17.2% hybrid leaning flexible, and 12.6% described the arrangement as their own choice, on 33,686 responses to that question [N26]. US-based respondents reported the highest remote share at 45%, on 5,848 responses [N26a">N26a].
The survey drew over 49,000 responses from 177 countries [N27]. Of 48,178 respondents, 69.8% were employed [N27a">N27a]. It is opt-in. These percentages describe how surveyed developers say they work, not what share of developer jobs are advertised as remote, and should never be restated that way.
The clearest measured AI effect on hiring is in job titles, not job volume. Distinct AI-touched job titles in the US grew from 264 in 2022, 2.6% of all titles with at least five postings, to 822 by Q1 2026, or 8.3% of all titles [N24]. Sixty-three percent of those US AI-touched titles are now non-tech roles, and in Q1 2026 Germany led Europe with 288 such titles (4.2%), ahead of the UK at 160 (2.7%) and France at 138 (3.3%) [N24a">N24a]. That counts distinct titles, a measure of how far AI language has spread through the taxonomy of work. It does not count postings.
What comes next is forecast, not measurement. In a panel fielded in July 2026 by Indeed Hiring Lab and Pulsenomics — roughly 120 responses on the postings and unemployment questions, but only 70 to 100 on the AI questions — 52% said AI would be at least a mild drag on employment, 35% expected net job gains, and 13% no effect. The same panel put the Job Postings Index down 1.4% by June 2027 [N25], and named Software Development among the fields it expects to lose jobs, with IT Infrastructure and Data & Analytics among those it expects to gain [N25a">N25a]. That is what a panel of forecasters believes, not a measurement of hiring.
The developer job market is smaller than it was in 2020 and tilted hard toward experience. Postings sit about 27.5% below pre-pandemic levels even after a near-15% recovery [V5, V5d], and 69.3% of what is posted is senior [N22a">N22a]. Hiring into it is slow and interview-heavy: a 75-day median from opening a technical req to the first hire [V1], 17.6 interviews and 23.3 interviewer-hours per technical hire [N12, N13], and a 7.3% application-to-accepted-offer conversion rate [N10]. Cost is a $135,980 median salary [V3] plus, by the ECEC arithmetic above, roughly 43% again in employer compensation cost [V4].
Which of the software development staffing models fits depends on how much of that 75-day median [V1] a team can absorb, and IT staff augmentation in 2026 is one of the routes teams take to shorten it.
For a leader who needs engineering capacity this quarter, those numbers describe a two-to-three-month gap between deciding and delivering, most of it spent on work no one wanted to do. That gap is the reason You-Source structures engineering as capacity you can turn on rather than a hire you have to complete first.
Median time to first fill for technical roles is 75 days, measured from a job being opened to the first hire being made, across Ashby's customer base [V1]. Broken out by seniority in that same dataset, junior roles take 52 days, mid-level 63 days and senior 71 days [V1a">V1a]. A different Ashby metric, the interview loop from first interview to last runs a median 17.9 days for technical roles, which covers the interview phase only [V1c">V1c].
O*NET reports a median annual wage of $135,980, or $65.38 an hour, for Software Developers, attributed to Bureau of Labor Statistics 2025 wage data [V3]. The BLS Occupational Outlook Handbook page for the same occupation still carries the earlier vintage, a median of $133,080 in May 2024 [V3b">V3b]. The broader group "Software Developers, Quality Assurance Analysts, and Testers" is a different set, with 2024 median pay of $131,450 a year [V3c">V3c].
Technical roles average 17.6 interviews per hire, up 52% from roughly 11 in 2021, across Ashby's customer base [N12]. Those interviews cluster into 5 interview events per technical hire and 4 per business hire [N14]. In Q1 2026 a technical hire consumed 23.3 interviewer-hours against 12.2 for a business hire [N13]. Only 13% of hires included a take-home component [N15].
Entry-level software development postings declined 7.5% year over year as of May 2026, while senior-level postings rose 14.7% over the same period [N22]. Senior-level positions accounted for 69.3% of software development job postings in Q1 2026, against about 14% of postings across all US occupations [N22a">N22a]. Postings are advertisements, not hires, so these figures describe what employers advertise rather than who they employ.
Salary, agency fee and employer cost are three separate quantities. The $135,980 median is employee gross salary [V3]. BLS Employer Costs for Employee Compensation put private-industry total compensation at $46.60 per hour worked in March 2026, of which wages and salaries were $32.60 and benefits $14.01 [V4]. Contingency developer recruiting fees run 18% to 25% of the hire's first-year base salary and are a vendor's price, not an internal cost-per-hire [V2].
