Working Life · Performance and collaboration
Working from home or the office: Where do we actually work better?
Monday's assignment is a demanding report. Tuesday calls for a conversation in which several half-formed ideas need to become something new. Should both happen in the same place? Arguments about working from home and office attendance often assume one must win everywhere. Yet deep individual work, onboarding, brainstorming and team coordination call for different conditions. We focus on knowledge work that can realistically be done with digital tools. Jobs in care, retail, trades and many other sectors face fundamentally different constraints. Just as importantly, working fully remotely is not the same intervention as spending one or two days at home each week.
For suitable office jobs, well-designed hybrid arrangements can improve job satisfaction and reduce employee turnover without necessarily harming performance. Research on communication and scientific collaboration also makes a case for deliberately preserving in-person contact, especially when building new connections and developing ideas. The evidence does not crown either fully remote work or full-time office attendance as universally most productive.
Position A: The strongest case for working from home
Less commuting, greater control over time
Not travelling to the workplace every day saves journeys and potentially money. In ten European organizations studied by Eurofound, employees identified the time saved on commuting as a leading reason to opt for hybrid work; some said they used that time for rest or personal responsibilities. That is a meaningful benefit, but it is not proof that every minute saved translates into higher productivity. The payoff depends on commuting distances, living arrangements and individual circumstances.
Stronger retention in a randomized trial
A randomized trial from August 2021 to January 2022 involving 1,612 university-graduate employees at Chinese travel-technology company Trip.com compared a schedule offering two home-working days and three office days with five office days. Over six months, 4.8% of the hybrid group left, compared with 7.2% of the office-only control group. That is a 2.4-percentage-point reduction, or roughly one-third in relative terms. Job satisfaction also improved. The trial tested one hybrid policy among specific occupations—not unlimited home working across the economy.
Space to focus when tasks and circumstances allow
A quiet home environment can suit writing, analysis and programming, provided equipment, connectivity and caring responsibilities allow uninterrupted work. Eurofound's ten case studies report greater autonomy and concentration as perceived benefits. These are qualitative findings from selected organizations, not a measured average productivity boost for every worker. Homes can also be noisy, cramped and full of interruptions.
Performance does not automatically require attendance
In the Trip.com trial, equivalence tests supported no meaningful difference in performance reviews or measured lines of code within the predefined testing bounds. The researchers found no statistically supported difference in promotions over up to two years; that does not prove complete equality. This challenges the sweeping claim that offering two days at home necessarily harms performance or careers. It establishes neither a productivity gain for every job nor the same outcomes for fully remote work.
Position B: Why the office still matters
New connections do not reliably happen on screen
Using communication records for 61,182 Microsoft employees in the first six months of 2020, researchers reported that firm-wide remote working made collaboration networks more siloed and static, with fewer links across groups. Communication also shifted from synchronous towards asynchronous channels. Such changes may make discovering new colleagues and exchanging unfamiliar knowledge harder. But this study follows patterns during the pandemic; it does not prove that every face-to-face meeting produces better work.
Being together can help ideas take shape
A bibliometric study examining about 20 million research papers and four million patent applications found that geographically distributed teams were less likely to generate highly disruptive advances than co-located teams. The authors point in particular to early-stage conceptual work. This is an observational study across many decades, not a randomized trial of present-day employees. Geographic distance between scientists is also not equivalent to two days of working from home.
Learning the ropes is a separate challenge
New colleagues need to absorb unwritten rules, discover whom to ask and build relationships. Being present can make casual questions, mentoring and immediate feedback easier. Significantly, the Trip.com trial excluded interns and employees still on probation, partly because on-site learning mattered for them. Its encouraging findings therefore should not be casually extended to new hires. Conversely, simply placing a new employee in an office is not a substitute for structured support.
Home is not automatically a healthy workstation
A kitchen table is not necessarily a safe substitute for ergonomic equipment, clear working hours or daily social contact. The joint WHO/ILO technical brief highlights physical and psychological risks and offers guidance on managing telework. Eurofound's case studies also stress appropriate equipment and worker involvement. This does not show that offices are always healthier; it means both settings require deliberate design and safeguards.
Fact check: What can the studies actually support?
In the randomized Trip.com trial of 1,612 employees, six-month attrition dropped from 7.2% to 4.8%, while satisfaction rose. The relative change is (7.2 − 4.8) / 7.2 = 33.3%. Random assignment supports a causal interpretation of the offer of two home-working days in this company—not a claim about every job worldwide.
In the hybrid trial, equivalence tests supported similar performance ratings within the testing bounds, while no difference in promotions was detected. Microsoft's study concerned widespread remote work early in the pandemic and reported more siloed communication networks. The findings need not contradict each other: one firm's performance reviews and cross-team communication are different outcomes measured in different settings.
The large bibliometric study linked distributed scientific and patent teams to fewer disruptive advances. Its design cannot demonstrate that ordering every employee into an office on a set number of days would improve any firm's creativity. It is a reason to protect opportunities for joint idea development—not an exact formula for mandatory attendance.
Eurofound's 2025 research found markedly different hybrid arrangements in ten organizations. Its 2025 EU e-survey, published in 2026, also reports tensions when employees' actual arrangements clash with their preferences. The online survey is not a probability sample of all European workers, and case studies are not controlled interventions. None of these sources determines a universally optimal ratio or the costs of every individual's commute.
Our assessment
A more useful question than 'home or office?' is 'what does today's work need?'. Quiet individual tasks may benefit from flexible days at home; onboarding, difficult conversations, informal ties and collaborative concept work deserve intentional opportunities to meet. The strongest case for hybrid working is therefore not simply splitting the difference, but matching place to purpose. Teams can reserve shared days for discussion and learning while allowing focused work away from the office. Performance should be assessed through meaningful outcomes rather than hours spent in view. Equipment, availability, fairness across job types and genuine employee choice also matter. A customer-facing small business has different needs from a software team, and a newcomer may need different mentoring from a seasoned colleague. The randomized evidence against an automatic productivity penalty is substantial within its setting, but the long-run effects on culture, innovation and careers remain uncertain across many work arrangements. Either slogan—'the office is obsolete' or 'remote work never works'—outruns the evidence.
Sources and references
Editorial source and language recheck: 7 November 2026. Article approved by the operator on 7 November 2026; separate publication checks remain open. Scope is mainly digitally compatible knowledge work, not a statement about Austrian employment law. Trip.com tested a randomized offer of limited hybrid work for selected workers. The Microsoft analysis reflects widespread remote working during the pandemic, with a published correction to a figure legend. The innovation study is bibliometric, not causal. Eurofound's 2025 ten-case study and non-probability 2025 e-survey are not population-wide effectiveness trials. WHO/ILO provide occupational guidance, not an experimental productivity effect. Numerical checks and disclosed interests appear in the research dossier. No third-party video transcripts were used. Conflict of interest: co-author James Liang is Trip.com’s co-founder, chairman and a shareholder. This does not automatically invalidate the trial, but is relevant to interpreting it.
- Bloom, Han & Liang (2024), Nature: Hybrid working from home improves retention without damaging performancePeer-reviewed, randomized controlled trial; original article; DOI 10.1038/s41586-024-07500-2
- Yang et al. (2021/2022), Nature Human Behaviour: The effects of remote work on collaboration among information workersPeer-reviewed observational/natural-experiment analysis; correction published 5 Oct 2021
- Lin, Frey & Wu (2023), Nature: Remote collaboration fuses fewer breakthrough ideasPeer-reviewed long-run bibliometric analysis; association, not randomized intervention
- Eurofound (17 June 2025): Shaping the future of work: Inside Europe's hybrid work strategiesEuropean Union agency; synthesis of ten qualitative organizational case studies
- WHO and ILO (2 Feb 2022): Healthy and safe telework, technical briefJoint expert technical guidance; risks and prevention, not a randomized productivity trial
- Eurofound (19 Jan 2026): Uneven picture of a changing Europe — 2025 e-survey resultsNon-probability online survey; descriptive preferences and reported experience