Remote's Edge Vanished. The Boss Didn't.

The remote satisfaction edge is mostly an unadjusted number. Christos Makridis, writing in The Conversation on October 7, says that once you count pay, communication, and the manager, fully remote workers are not happier than people in the building. The only group that still looks a bit more satisfied is the one that works remotely sometimes. He calls that difference modest.

That is a productivity finding, not a culture essay. If your week is organized around “remote is the perk that makes the job tolerable,” the perk may be standing in for a manager who replies, a workload that fits the day, and someone you can ask without waiting three months. We already wrote the calendar version of a shorter week in the four-day review that is not a Friday off. We wrote the hours an assistant gives back and then takes again in the botsitting piece. We wrote a list of remote tactics in the 2026 strategies roundup. This one is the variable those pieces keep bumping into. Location is not the variable that survives the controls.

The satisfaction edge was the unadjusted number

Makridis, an associate research professor of information systems at Arizona State University, co-authored the study with Jason Schloetzer. They used the Payscale database: responses from nearly 165,000 U.S. employees between 2020 and 2023. The surveys asked whether people felt appreciated, whether pay looked fair and transparent, whether communication was useful, whether they had chances to develop, and whether they had a good relationship with their manager.

Before those controls, semi-remote and fully remote workers can look more satisfied. Makridis says that appearance needs the other factors in the model. Better pay and more supportive practices travel with some remote jobs. If you compare a well-paid remote role to a poorly managed on-site role and then credit the sofa, you have measured the package, not the commute.

After the job-specific variables, the edge for fully or mostly remote workers disappears. “Sometimes remote” is the leftover. Modest, in his wording, means you should not rebuild a policy around it. A company that announces a one-day-at-home rule because of this paper is reading past the sentence. The sentence says the location effect shrinks when the workplace is held constant. It does not say Tuesday at home is a treatment with a known effect size you can paste into a slide.

He also used the Gallup Workplace Panel to follow the same workers over time. That pass reinforced the cross-section: you have to look at the workplace as a whole. A one-time survey that asks “remote or not” and “happy or not” will keep producing the headline the unadjusted cut produces. The panel is there to stop that.

Makridis discloses that he is a senior researcher at Gallup and receives foundation funding to study remote work. That is not a reason to toss the Payscale cut. It is a reason to notice that one of the stress figures below comes from a Gallup report, and the author is not a stranger to that shop. The Payscale sample and the Gallup stress table are different instruments. Do not average them into one “remote score.”

A good manager shows up more on site

The manager item is a 5-point agreement with “I have a great relationship with my direct manager.” Four or five counts as favorable. It is the employee’s rating, not a checklist of one-on-ones.

A favorable rating lined up with higher job satisfaction in every arrangement. The association was stronger for people working on site than for people who always or sometimes worked remotely. Makridis offers one reading, and he marks it as possible rather than proven: working in the same place gives you more chances to use a supportive manager, the mentoring and the development conversations. Distance can also soften a bad relationship. If the manager is the problem, not seeing them every day is a kind of relief. That second half matters. A return-to-office memo that assumes proximity only helps is ignoring the case where proximity is the tax.

This is the practical split. If your manager already replies, already sets a priority, and already notices finished work, the office is an extra channel, not the channel. If your manager does none of those, a badge swipe will not install the habit. The paper’s on-site association is about employees who already have the relationship. It is not a recipe for creating one by seating chart.

Younger trainees are the group Makridis says may want the building for a different reason. Someone still learning the job can use easier access to colleagues. Someone who wants an uninterrupted block at home can resent the same commute. A single attendance rule treats those as one preference. The Payscale result is what you get when you stop doing that and still find location weakly related to satisfaction.

Stress does not follow the satisfaction chart

Job satisfaction, daily stress, and burnout are different questions in this writeup. Liking the work and being exhausted can sit in the same week. So can liking the home setup and feeling cut off. Makridis separates them on purpose.

On stress, Gallup’s 2026 State of the Global Workplace report, as he cites it, puts substantial stress on the previous day at 41 percent for fully remote employees, 46 percent for hybrid employees, and 46 percent for on-site employees whose jobs could be done remotely. The remote number is lower. It is not low. Four in ten fully remote workers still reported substantial stress yesterday. A policy deck that says “remote fixes stress” is using a five-point gap and dropping the base rate.

The on-site figure is specifically people whose jobs could be remote. That is the right comparison if you are arguing about a laptop job. It is the wrong comparison if you paste it onto a ward, a shop floor, or a route. Those jobs are not in that 46 percent.

He uses physicians as a separate illustration that satisfaction and stress coexist. Among physicians in organizations that took part in an American Medical Association survey in 2024, about three-quarters reported being satisfied with the job, while almost half reported a great deal of stress from the job, including paperwork. The published summary does not say how many doctors were in both groups. You cannot turn “three-quarters satisfied” and “almost half stressed” into a Venn diagram. You can say a profession can post both numbers at once, which is enough to stop treating a satisfaction score as a wellness score.

A different Gallup cut, nearly 7,500 full-time U.S. employees, ties burnout to conditions rather than to the postcode of the desk. The factors closest to burnout in that research, as Makridis lists them, were unfair treatment, crushing workloads, unclear expectations, poor communication and support from managers, and unreasonable time pressure. Home can remove a bad commute. Home can also slow a question that used to take a hallway. Neither one clears an overloaded queue or a manager who will not say what “done” means.

If you are tracking productivity, track the queue and the reply time. A badge report will not show the three-month wait below. A satisfaction pulse that only asks about the office will not show the workload item Gallup put next to burnout.

Asking for help got slower

The October 7 piece is about satisfaction. A related reporting line, from a Yahoo Finance article dated October 5, is about whether anyone answers.

Peter Cappelli at Wharton and Y. Jasmine Wu at the University of Texas at Austin looked at what they call relational capital: knowing whom to ask, feeling able to ask, and having a colleague willing to answer. The paper, “How Remote Work Is Reshaping the Office,” appeared online in California Management Review in July. The interviews are older than the news cycle. Yahoo’s writeup is inside the week. I am using the news piece, not pretending I re-read the journal article this morning.

The method in that writeup: 38 group interviews in late 2023, roughly 760 employees at one multinational financial services firm, in the U.S., the U.K., and India. That is one company. It is not the Payscale file of 165,000. Do not merge the samples because both mention remote work.

Inside that firm, people who called a colleague they already knew got help quickly. Otherwise help was more likely to be delayed. One relatively new hire with no work friends tried for three months to get an answer. The authors describe a loop. Slow replies make the next ask less likely. The relationship thins because the ask stopped, not because anyone scheduled a breakup.

That loop is a productivity leak you can see without a survey vendor. Count how long a new person waits for a first useful reply. If the wait is measured in weeks, the office-versus-home argument is a distraction. The missing object is a named person who is expected to answer.

Cappelli has made the newcomer point before, according to the Yahoo piece. Entry-level workers are the ones without the old address book. A remote policy that assumes everyone already knows whom to ping is a policy written by people who do. The three-month anecdote is one person in one interview set. It is not a median. It is a ceiling you do not want your own onboarding to approach.

Makridis’s leisure-time finding sits next to this, and it is his own earlier work, not the Cappelli interviews. Among younger workers in jobs that can be done remotely, the share of leisure time spent alone rose from about 25 percent in 2017 to about 40 percent in 2024. Control of the calendar did not automatically become time with other people. Friends and colleagues have their own calendars. A free afternoon is not a meeting.

He points to a separate paper by Natalia Emanuel, Emma Harrington, and Amanda Pallais. In a nonrandomized comparison of almost 600,000 people, they looked at mental health before and after the pandemic in jobs that could and could not readily be done remotely, leaving out 2020 and 2021. Remote workers who lived alone came out worst. Again: different study, different design, do not staple it to the 41 percent Gallup stress figure and call it one effect. The shared warning is narrower. Social access is a work variable, and living alone makes the remote version sharper.

A remote day only helps if the overlap is real

Makridis’s advice to employers is specific enough to use and narrow enough not to fake. If you are pulling people back, first name what they get from being in the same place, then put the office time on that. Teams that need frequent collaboration need overlapping hours, not a shared badge average. Newer employees need colleagues who can teach the job to actually be there. A Wednesday mandate where the people you need are on different Wednesdays is attendance theater.

He also says preferences and job requirements differ, so a manager has to pick an arrangement for the team in front of them rather than import one from a peer company. The part that does travel is feedback. Useful feedback, on site or remote, means talking about priorities, noticing the work, and helping people see the next step. Gallup’s Jim Harter, as quoted in the piece, treats that as a weekly habit. “These conversations don’t have to be long; 15 to 30 minutes, done consistently, is enough.”

Fifteen minutes is a calendar object. It is not a vibe. If your one-on-one is a status recital, it will not supply the manager relationship the Payscale item is measuring. The item is “I have a great relationship,” not “I attended a meeting.” A short weekly conversation that names one priority and one finished thing is closer to the mechanism than a three-day office rule with no overlap.

Two traps sit on either side. One is declaring the office irrelevant because the adjusted satisfaction gap vanished. The newcomer who waited three months, in the Cappelli interviews, did not have an irrelevant problem. The other trap is declaring the office sufficient. The burnout list (unfair treatment, load, unclear expectations, a manager who does not communicate) does not dissolve because the chairs are in one room.

What to change before you argue about badges

You can do this without a new tool and without a town hall.

Write down, for each person who started in the last six months, who they are supposed to ask and how fast that person is expected to answer. If you cannot name the pair, the Cappelli loop has room to start. The 760-person study does not give you a service-level number. It gives you the failure mode: unfamiliar ask, delayed help, fewer asks after that. Pick a reply window your team can actually keep. A same-day reply on a blocking question is a choice. A three-month reply is what happened when nobody chose.

Put the 15 to 30 minutes on the calendar as a repeating hold, and use it for priority and recognition, not for a tour of every ticket. Harter’s line is about consistency. Skipping three weeks and then holding a 90-minute retrospective is a different practice. The Payscale association was stronger on site, which is a reason to spend those minutes in whatever channel you actually share, not a reason to skip them because people are remote.

Separate the stress question from the satisfaction question if you survey at all. A pulse that only asks whether people like the flexibility will miss the 41 percent who had a substantially stressful day and still might click “satisfied.” The AMA physicians are the caution: high satisfaction and widespread stress in one profession, with no published overlap count. Ask the workload and the clarity items if you want the burnout side. Ask the manager item if you want the satisfaction side. Do not ask only “how many days.”

If you are the worker rather than the manager, the same facts cut the other way. A remote setup that leaves you waiting on strangers is not a focus strategy. It is a queue. Find the familiar colleague the Cappelli interviews said actually answers, and ask that person before you open a fifth chat with someone who has never seen your name. If you live alone and the Emanuel, Harrington, and Pallais pattern worries you, the 40 percent leisure-alone figure is a prompt to schedule the non-work contact on purpose. The research does not hand you a social app. It says the free hour does not fill itself.

None of this requires you to win the return-to-office argument. The October papers are useful because they shrink that argument. Location, after the controls Makridis lists, is a weak satisfaction lever. The manager relationship, the reply, and the load are the levers that stayed in the model. Spend the week on those. The badge report can wait until the new hire has someone who answers.