Hybrid Work Statistics: Adoption, Productivity and Policy

If you are writing or revising a hybrid work policy, you will be shown plenty of statistics, and they rarely seem to agree. In Buffer's 2023 survey, 98% of the remote workers who answered wanted to work remotely for the rest of their careers. Badge data from office buildings in 10 big U.S. cities showed them about half as busy in one September 2026 week as in early 2020. Gallup found that about half of U.S. employees whose jobs can be done remotely worked hybrid in May 2026. These figures can all be accurate, because they count different people, in different years, in different ways.

This guide brings together the figures that come from the original surveys and datasets, each with its date and the group it describes. It covers how many people work hybrid, what employees want compared with what employers offer, how often hybrid workers actually come in, what the strongest evidence says about productivity and retention, and what the research suggests about setting the policy and testing it on your own teams.

DisclosureJake Jorgovan co-founded Alpha Apex Group, which is mentioned in this article.

How many people work hybrid?

About half of U.S. employees whose jobs can be done remotely work hybrid. In Gallup's survey of May 2026, 52% of remote-capable employees worked in a hybrid arrangement (opens in a new tab), 26% worked exclusively remote and 22% worked on-site. Gallup says the hybrid share has held relatively steady since 2022. Technology is the exception it notes: in its 2025 analysis, remote-capable tech employees were as likely to be fully remote (47%) as hybrid (45%), and 9% were fully on site.

Sources that count everyone who works give lower numbers, because they include people whose jobs have to be done in person. The table sets three current sources side by side.

Three differences explain most of the gap between rows:

• The group counted. Gallup counts only people whose jobs can be done remotely. The SWAA (opens in a new tab), run by economists including Nicholas Bloom and Steven Davis, counts full-time wage and salary employees in every kind of job (people who earned at least $10,000 in a prior year and worked five or more full days in the survey week). BLS Table A-41 (opens in a new tab) counts everyone at work.
• The definition. Gallup counts someone as hybrid if they work remotely from 10% to less than 100% of the time, so its on-site group includes people who work remotely occasionally. The SWAA counts someone as hybrid if they worked at least one full day at home and one at the workplace in the previous week.
• The period. BLS and the SWAA classify each person by a single week, so a hybrid worker who happened to stay in the office all that week counts as not teleworking (BLS) or fully on site (SWAA). Gallup classifies people by the share of their time they work remotely.

Education changes the picture sharply. In the same BLS table, among people at work aged 25 and over, 37.8% of those with a bachelor's degree or higher teleworked during the September 2026 survey week, against 24.1% of everyone in that age group.

The longer trend shows how far things have moved. The Census Bureau reported (opens in a new tab) that the share of U.S. workers who primarily worked from home rose from 5.7% in 2019 to 17.9% in 2021. In a Pew Research Center survey of 5,775 workers published in March 2023 (opens in a new tab), 61% of U.S. workers did not have jobs that could be done from home. Among those who did, the share working from home all the time had fallen to 35% (from 55% in October 2020), while hybrid had risen to 41% (from 35% in January 2022). Since then, Gallup's data show hybrid as the settled norm for remote-capable employees.

What employees want compared with what employers offer

Employees want somewhat more flexibility than they get. When Gallup asked remote-capable employees in November 2025 where they would prefer to work, 61% chose hybrid, 33% exclusively remote and 6% on-site. Set against where they were working in May 2026, hybrid is close to what people want, fully remote is a little scarcer than demand, and on-site work is far more common than the 6% who prefer it.

Employee responses to the SWAA show a similar gap. Its October 2026 update reports that employers offer fewer fully remote jobs and more fully on-site jobs than employees want. The work-from-home days employees say their employers plan for them have hovered around 1.3 to 1.5 a week, and since mid-2022 they have resembled the amount people actually work from home.

The 98% figure from the opening comes from Buffer's 2023 State of Remote Work (opens in a new tab), a survey answered by 3,000 remote workers. It describes people who already work remotely, so it cannot tell you what your office-based staff want. Before using any preference figure, check who answered.

How often hybrid workers come into the office

On average, hybrid workers spend a little under half their week in the office. Gallup's September 2025 analysis (opens in a new tab) found hybrid workers spent 46% of their workweek in the office, the equivalent of 2.3 days, up from 42% in 2022. All of that increase came in 2023, with no movement in the year before the analysis.

Across all workers, not only hybrid ones, the SWAA (which surveys U.S. residents aged 20 to 64 who earned at least $10,000 in a prior year) found that about 26% of full paid working days in September 2026 were worked from home.

Office buildings show how unevenly that time is spread across the week. Kastle Systems, which counts each cardholder's first building entry of the day from its access-control data, reports a weekly occupancy barometer (opens in a new tab) for 10 major U.S. metro areas, measured against February 2020 levels. For the week including Tuesday, September 22, 2026, the 10-city average was 54.7%, and the peak day, Tuesday, reached 65.3%. Tuesday was the peak day in eight of the ten cities, and Kastle says that week's average may also reflect the Yom Kippur holiday on the Monday. Kastle notes its data are not a national statistical sample, and the figures change weekly.

The gap between the weekly average and the peak day matters for space planning. If most teams choose the same anchor days, the office has to be sized for the busiest day, while the average day leaves much of it empty.

Does hybrid work hurt productivity or retention?

The strongest evidence on this question comes from a randomized trial, because it compares similar employees who were assigned different arrangements. In 2021 and 2022, Trip.com, a large Chinese travel technology company, ran a six-month experiment with 1,612 graduate employees in engineering, marketing and finance. Those born on odd-numbered days were given the option to work from home on Wednesdays and Fridays; the rest stayed in the office five days a week. Most of those offered the option used it about one day a week, typically Friday, because large meetings and product launches tended to fall midweek. The results were published in Nature in June 2024 (opens in a new tab) by Nicholas Bloom, Ruobing Han and James Liang.

• Quits fell by a third. Attrition during the experiment was 4.8% in the hybrid group against 7.2% in the office group. The drop was larger for women, non-managers and employees with long commutes.
• No harm to performance. Performance reviews over the two years from the start of the trial showed no difference, and neither did lines of code written by the computer engineers. The researchers found no evidence of a difference in promotions.

Image

Zoom

Across four review periods, the Trip.com trial found no statistically significant performance-grade difference between the hybrid and office groups. This is evidence from one company; sample sizes fell as employees left.nature.com (opens in a new tab)

• Managers changed their minds. Before the trial, managers expected hybrid work to lower their own productivity by 2.6% on average. Afterwards, their average estimate was a 1.0% gain.

Image

Zoom

In Trip.com’s six-month randomized trial, attrition was 7.2% in the office group and 4.8% with a hybrid option: a 2.4-point, or one-third relative, reduction. One company’s result does not establish the effect everywhere.

Trip.com estimated that each quit cost about $20,000 in recruitment and training, and it announced in February 2022 that hybrid work would extend to all employees. The authors note the limits: it is one company, and they are not sure the results would hold for arrangements with one office day a week or less, because of training, innovation and culture. The trial also left out interns and employees on probation, because on-site learning and mentoring matter most for them. If you want to set your own quit rate against a national benchmark, this guide to employee turnover statistics explains the BLS figures.

Gallup's surveys add a different measure, engagement, and show an association rather than an effect. Its May 2026 data show 34% of hybrid employees and 37% of exclusively remote employees engaged, against 29% of on-site employees whose jobs could be done remotely.

Trust is the weak point. Just over half of managers of remote workers (54%) strongly agreed in Gallup's September 2025 analysis that they trust their teams to be productive when working remotely, and 57% of employees said they feel trusted. Gallup links four practices to greater trust: timely communication about the team, a sense of community, accountability for performance, and the same feedback and development opportunities wherever people work.

Who should set the hybrid schedule?

Gallup's surveys link who sets the schedule with how fair employees consider the policy. The same September 2025 analysis found hybrid schedules split roughly evenly: 34% of hybrid employees set their own, 35% had theirs set by their manager or team, and 31% had it set by the employer or leadership.

• 91% of employees whose team decides the schedule see the policy as fair, the same rate as those who decide for themselves.
• 73% see it as fair when the employer decides.
• Compared with team-set schedules, employees who set their own were 76% more likely to name burnout or fatigue as their greatest challenge, and 52% more likely to name meeting customer needs.

Gallup's conclusion is that hybrid works best when teams agree the rules, so colleagues can count on each other being in the office on the same days. Few employees have that: as of May 2025, only 11% were on teams that set their hybrid policy together.

Image

Zoom

Original plot of Gallup’s May 2025 U.S. hybrid-worker survey, reported September 2, 2025. The chart shows perceived fairness within each schedule-setting group; the observational results do not establish causation.gallup.com (opens in a new tab)

Costs and compliance

On costs, Global Workplace Analytics (opens in a new tab), a research-based consulting firm, estimates that "a typical employer can save about $11,000/year for every person who works remotely half of the time." It is a modeled estimate on a page first published in 2020. Your own figure depends on whether you actually give up space, which the peak-day pattern above can make harder.

Where people work also affects tax. New York is one example: its tax department says that for a nonresident whose primary office is in New York, days spent telecommuting count as days worked in the state (opens in a new tab) unless the employer has established a bona fide employer office at the telecommuting location. Rules differ by state and country, so a policy that lets people work from another state needs payroll and tax advice before it starts.

How to set and test a hybrid policy

The figures above describe averages. Your policy has to work for your teams, and the most useful evidence you can get is your own.

I have a simple rule for decisions like this. As I wrote in 2020, "When the path isn't clear, it's because you lack the facts to make the right decision." Action is how you get them. The Trip.com trial is a good example: managers who expected hybrid work to hurt productivity changed their view after six months of running it.

1. 1

   Decide what the office days are for

   Name the work that needs people together: onboarding, mentoring, planning, client work. Trip.com left interns and staff on probation out of its trial for this reason. If you cannot say what an office day is for, a required number of days will be hard to defend.

2. 2

   Check where people work before you allow it

   Confirm the payroll, tax and employment-law effects of any work from another state or country before it starts.

3. 3

   Set a floor and let teams choose the days

   Set the minimum the business needs, then let each team agree its own anchor days within it. In Gallup's survey, team-set schedules were seen as fair as often as self-set ones, and employees with self-set schedules were more likely to name burnout or fatigue as their greatest challenge.

4. 4

   Choose your measures and record the baseline

   Before the test starts, record quits, performance ratings, promotions, attendance by day and engagement, by team. Decide which results would justify keeping or changing the policy, and whether you will compare with a similar team or with the period before the test.

5. 5

   Run it as a dated test

   Run the policy for a fixed period, six months in the Trip.com trial. Enroll everyone eligible. At Trip.com the volunteer rate was low at first, volunteers were less likely to be women, and management suspected people feared that volunteering would signal low ambition.

6. 6

   Review the results and decide

   At the end of the period, keep, change or reverse the policy on the evidence, and tell employees what you found.

What would you learn if you tested your current policy on one team for six months?