How to find remote jobs that are actually remote

In August 2026, 10.6 percent of the people at work in the US worked all of their hours from home — 16.5 million out of the 155.5 million at work that week, in the Current Population Survey's telework table (BLS, table A-41). Another 11.0 percent teleworked part of the week and commuted for the rest. That second group is why a posting headed "Remote" so often turns out to mean something else, and telling the two apart comes down to three things you can check before you fill in a single form: the occupation, the employer's industry, and four specific lines in the posting.

How many jobs are actually fully remote?

The CPS asks people who worked during the survey reference week whether they teleworked or worked at home for pay, and for how many hours. In August 2026, 21.6 percent said yes — 33.6 million people. Split that group by hours and it lands close to even: 11.0 percentage points of the workforce teleworked some of their hours, 10.6 all of them.

So slightly fewer than half the people teleworking in any given week are doing it full time. The rest are in on Tuesday.

Telework in the US, August 2026

21.6%Of people at work teleworked at all

10.6%Teleworked all of their hours

16.5MPeople working entirely from home

What this number is not: it counts workers, not postings. BLS publishes no count of job openings advertised as remote, so nothing here tells you what share of the listings on a board are honest about the arrangement. It describes the far end of the process — what people are actually working, after every offer has been made and every relocation clause has been signed.

Which occupations go fully remote, and which never do

Percent of people at work who teleworked, by occupation group

CPS telework rate, August 2026. Counts anyone who teleworked for any part of the reference week.

Management, professional and related		35.9%
Sales and office		23.3%Where customer support sits
Service		5.8%
Natural resources, construction, maintenance		3%
Production, transportation, material moving		2.9%

Management, professional and related occupations teleworked at 35.9 percent in August 2026. Production, transportation and material moving occupations teleworked at 2.9 percent (BLS table A-42). That is a twelvefold gap, and no amount of searching closes it.

Which leads to the least popular sentence in this post: the occupation you apply into predicts your odds of fully remote work better than any filter on any job board. A Content Marketing Manager sits in the management and professional group, at the top of that chart. A Customer Support Specialist sits in sales and office occupations, at 23.3 percent. Neither figure is a promise about any one employer. Both are the base rate you are arguing against when you decide the right posting must be out there somewhere.

Then look inside the leading group. Of the 35.9 percent of management and professional workers who teleworked, 19.0 percentage points did it for part of their hours and 16.9 for all of them. Even in the most remote-friendly occupation group in the country, hybrid is the more common arrangement. A Content Marketing Manager who treats fully remote as the default for the role has the base rate backwards, and will read every hybrid posting as a bait-and-switch rather than as the ordinary shape of the job.

Remote, hybrid and "remote-eligible" are three different offers

None of these words are standardised, and no federal source defines any of them. What each one leaves unanswered, though, is consistent enough to turn into a checklist.

How to read this: the headline is written to attract applicants, and the application form is written by the people who run payroll. When the two disagree, the form is the one that decides where you are allowed to live.

What O*NET tells you about a job before the posting does

O*NET's Work Context section is the closest thing to a public answer on whether a job can be done from a kitchen table, and it costs nothing to check.

Take Marketing Managers (11-2021.00), the occupation O*NET files most content marketing work under. One hundred percent of respondents reported using electronic mail every day, and 90 percent reported telephone conversations every day. Face-to-face discussion is much thinner: 19 percent reported it once a week or more but not every day. Working with a group or team is rated extremely important by 83 percent. That describes a job made of scheduled conversation, and scheduled conversation survives a video call.

Now run the same read on Customer Service Representatives (43-4051.00), where a Customer Support Specialist is indexed. Telephone conversations: 100 percent every day. Electronic mail: 83 percent every day. So far the two occupations look alike. But face-to-face discussion runs at 51 percent every day, and 25 percent of respondents reported physical proximity of "very close (near touching)".

That last figure is the tell, and it is not describing a support inbox. It is describing a counter. One occupation code covers both the person answering chat tickets from a spare bedroom and the person standing at a hotel desk, which is why O*NET's own reported titles for 43-4051.00 run from Customer Service Specialist and Customer Support Representative through to Call Center Representative and Guest Service Agent. Search those variants one at a time. The code itself is too wide to search on.

A caveat on using O*NET this way. "Content Marketing Manager" does not appear in the reported-title list for Marketing Managers at all, a list that includes Brand Manager, Marketing Communications Manager, Product Marketing Manager and Marketing Director. Those lists lag what employers post by years. Use them to widen a search, never to decide what your own job is called.

Where remote work sits, by industry

Industry moves the odds as hard as occupation does, and it is the variable most people ignore while sorting a board by job title.

CPS telework rates by industry, June 2026, reported in BLS's The Economics Daily.

The categories also split hard inside themselves, which is where a title-first search goes wrong. Wholesale and retail trade came in at 12.1 percent in June 2026, but that single figure hides wholesale trade at 21.6 percent and retail trade at 10.3 percent. Professional and business services averaged 41.8 percent, and inside it professional and technical services reached 55.8 percent while management, administrative and waste services sat at 16.0 percent. So a Content Marketing Manager weighing two offers is choosing between roughly one-in-two odds and roughly one-in-six, on a difference that never appears in the job title.

What this table is not: it is the employer's industry, not your job function. A Customer Support Specialist at an insurance carrier is counted inside finance and insurance at 59.0 percent. The same title at a restaurant group is counted inside accommodation and food services at 3.9 percent. Same work, same resume, two very different odds — which is a reason to choose the industry on purpose rather than let a job board's title match choose it for you.

Why entry-level remote jobs are so hard to find

Nobody filters a job board by age, but the telework rate splits on it anyway. In March 2026, 6.7 percent of employed people aged 16 to 24 teleworked, against 25.1 percent of those aged 25 to 54 (BLS, The Economics Daily) — nearly four times the rate.

CPS telework rates by group, March 2026, reported in BLS's The Economics Daily.

What this is not: it is not evidence that employers refuse to hire young people into remote roles. Age is standing in for occupation here. The 16-to-24 group is concentrated in food service and retail, work that cannot leave the building at all, and the 5.8 percent service-occupation rate in the chart above is the same fact seen from another angle.

The effect on a search is real all the same. Someone two years out of school, filtering a board to Remote for Customer Support Specialist postings, is competing inside a slice of the market that people their age mostly do not hold yet — and the way out is to change the occupation or the employer's industry, not the filter. The 4.4-point gap between women and men in the same month is the same story: it tracks which occupations each group works in, not a policy anyone wrote down.

Four checks on one posting

Everything above is a base rate, useful for deciding where a week of searching should go. On a single posting in front of you right now, four checks settle the question in about two minutes, which is worth doing before the forty minutes the application will take.

1. Open the application form and read the location field. A dropdown of named states is the real policy. If the field demands a city and the headline said Remote, someone has already answered your question.
2. Search the posting body for "hybrid", "on-site", "in office" and "days per week". The true arrangement usually sits in one line near the bottom, well below the paragraph about culture.
3. Compare it against the employer's other listings. If every other role on the careers page reads Hybrid — Chicago and this one reads Remote, that difference is worth a question rather than an assumption.
4. Ask two things in the recruiter screen. How many days a week is this role expected on site, and which states can it be hired in. Both are neutral, and neither has ever cost anyone an interview.

Is remote work actually shrinking?

The overall rate rose, then flattened. The full-time share is the part that slipped. BLS recorded a telework rate of 17.9 percent in October 2022 and 22.9 percent in the first quarter of 2024 (Beyond the Numbers, volume 14), then 22.6 percent in March 2026 and 21.6 percent in August 2026; over the year to March 2026 it stayed inside a band of 21.5 to 23.0 percent. Against that flat line, the share of teleworkers doing all of their hours at home fell from 54.0 percent in the first quarter of 2023 to 47.9 percent a year later, and stood at 49.1 percent in August 2026 — 16.5 million of the 33.6 million who teleworked at all, on table A-41's own figures. That is not fewer remote jobs. It is fewer fully remote arrangements inside a roughly steady number of remote-ish ones, which is precisely the pattern that makes a search feel dishonest when the data underneath it is not.

The series also has a hole in it: there is no October 2025 reading, because the appropriations lapse interrupted collection. Anyone drawing a smooth line through that autumn is drawing it themselves.

One more limit worth stating plainly. The CPS telework questions ask whether someone worked at home for pay in one reference week, so a fully remote worker who spent that particular week at a company offsite is counted as hybrid, and a warehouse supervisor who caught up on scheduling from the sofa on Sunday is counted as a teleworker. The numbers are a good guide to which occupations and industries permit remote work at scale. They are a poor guide to any individual job.

If you are applying for this role

Three specifics, whether you are chasing a Content Marketing Manager posting or a Customer Support Specialist one.

Search the title variants, not the category. For support, that means Customer Service Specialist, Customer Support Representative and Client Services Representative, from O*NET's list for 43-4051.00. For marketing, add Marketing Communications Manager and Product Marketing Manager. Each variant surfaces postings the others miss, and the fully remote ones are not evenly distributed across them.

Know the range before the form asks for it. BLS puts the median pay for customer service representatives at $44,770 a year, or $21.53 an hour, as of May 2025, across 2,666,000 jobs, with employment projected to fall 5 percent from 2025 to 2035. The Handbook says only that "working from home is also possible in some companies", and OES does not break those wages out by work arrangement, so there is no federal figure on whether remote pays more or less than on-site for the same title.

Ask the state question in week one, not at offer stage. It is the single fact most likely to end a process late, and the cheapest one to establish early.

Once the filtering is done, the applications are the slow part, and the location field is one of several dozen you will retype into every form. The platform fills the application from your resume, so the hours you have go into checking which postings are real rather than into typing your address again.