Remote and hybrid work, honestly

Few workplace questions have generated more confident assertion on thinner evidence. A few things are reasonably well supported, and a great deal of what is claimed is not.

What the better evidence shows

Productivity effects are mixed and generally small. A well-known randomised trial at a Chinese travel company found a substantial performance improvement among call-centre staff working from home, alongside lower attrition — but call handling is measurable, individual and unusually suited to the arrangement.

Later randomised work on hybrid arrangements among graduate employees at the same firm found no significant difference in performance or promotion, with clear reductions in attrition.

The pattern across studies is that effects depend heavily on the type of work, and that the largest and most consistent finding is about retention rather than output.

Commuting time is a genuine gain. The most reliably positive component, and it is arithmetic rather than psychology.

Time-use research has consistently found commuting among the lowest-rated daily activities, and removing it returns hours that mostly go to work, sleep and household activity.

Weak ties suffer. Analyses of communication networks after large-scale shifts to remote working have found networks becoming more static and more siloed, with fewer connections across groups.

That matters because cross-group connections are disproportionately the route by which information and opportunity travel.

What is asserted without much support

That innovation requires proximity. Frequently claimed, and the evidence is thin and contested.

Some studies find advantages for co-located teams on certain creative tasks; others find no difference or advantages for distributed groups on idea generation, where working separately avoids the convergence that group discussion produces.

The honest position is that different stages differ: generating options may work better apart, and resolving disagreement usually works better together.

That culture cannot be built remotely. Culture is largely the accumulation of what is rewarded, how decisions are made and how people are treated, none of which requires a building.

What does become harder is the informal transmission of norms to new joiners, which is a real and specific problem rather than a general one.

That everyone prefers it. Preferences vary widely by role, seniority, housing, caring responsibility and personality, and aggregate survey figures conceal that variation.

Who it works less well for

Worth stating, because the costs fall unevenly.

New joiners and junior staff. Learning by observation, incidental correction and informal mentoring are all substantially harder at distance, and the loss is invisible in the short term.

People without good space at home. Remote work is considerably more pleasant with a room and a door than with a kitchen table in shared accommodation, which makes the arrangement's benefits partly a function of housing.

People who rely on work for social contact. A real and under-discussed effect, particularly for those living alone.

The proximity bias problem

The most serious risk in hybrid arrangements specifically.

Where some people are present and others are not, there is reasonable concern that visibility translates into opportunity — and evidence from earlier remote-working research found promotion penalties for remote staff even where performance was equal.

The mitigations are structural rather than attitudinal: measure outputs rather than presence, run meetings so that remote participants are not disadvantaged, and check the distribution of promotions and assignments against attendance patterns.

If the check is never run, the bias is invisible and persists.

What actually makes hybrid work

Coordinated days rather than individual choice. The failure mode of hybrid is a half-empty office where everyone attends on different days, incurring the commute without the interaction.

A purpose for the office day. Sitting on video calls in a building you travelled to is the worst available combination, and it is extremely common.

Days in person work when they are used for the things that need presence: complex disagreement, planning, onboarding, and unstructured contact.

Written defaults. Distributed work degrades quickly where decisions live in conversations. Writing things down benefits everyone and is essential where not everyone was there.

Explicit availability norms. Without them, remote work expands to fill the day, which is the mechanism behind most reported increases in hours.

The reasonable summary

Retention improves, commuting time is recovered, output changes modestly in either direction depending on the work, and network structure degrades in ways that take years to show up.

None of that supports a universal answer, which is why the organisations handling it best are the ones treating it as a design problem per team rather than a policy per company.