Presence is a terrible proxy, and most remote leadership still uses it

Only 54% of managers strongly agree they trust remote employees to be productive, and only 57% of employees strongly agree they feel trusted. That gap, from Gallup’s hybrid-work research, is the operating condition of distributed leadership, not a side effect of it. Managers still hunt for substitutes for walking the floor: green dots, camera-on meetings, response latency in chat. Those signals are cheap to observe and almost worthless as measures of work. They also punish the people who do the actual job well: the engineer who goes dark for four hours to finish a design, the analyst in a later timezone who ships at 07:00 local.

The difficulty is not “building culture on Zoom.” The difficulty is that the office hid three separate problems behind one building. Trust was inferred from seeing people. Accountability was inferred from being in the room. Inclusion was inferred from whoever spoke last in the corridor. Remote work does not create those problems. It removes the camouflage. After this, you should be able to replace presence-based management with three explicit systems: a trust model you can actually operate (ability, benevolence, integrity), a small set of outcome metrics that survive timezone lag, and team working agreements that treat culture as a written protocol rather than a vibe.

I would not bother with most of the “remote leadership frameworks” sold as canvases and values posters. They do not specify what a manager does on Tuesday when a deliverable is late and the person is nine hours away.

The mechanism people keep getting wrong

Distributed work fails at coordination, not at motivation. Gallup finds that 48% of hybrid workers say their team has never discussed a formal or informal plan for how to collaborate, yet teams with a hybrid collaboration plan are 66% more likely to be engaged and 29% less likely to be burned out. Managers themselves name communication (48%) and collaboration (44%) as their top hybrid challenges. That is a process failure.

Microsoft’s telemetry, ending 15 February 2025 and excluding EU and education tenants, describes the cost of filling that vacuum with meetings. Employees in the heaviest-ping quintile are interrupted every two minutes during core hours, 275 times across a 24-hour day. The average worker receives 117 emails and 153 Teams messages per weekday. Fifty-seven percent of meetings are ad hoc. Nearly a third of meetings span multiple time zones, up 35% since 2021. Meetings after 20:00 are up 16% year on year. In the 2023 Work Trend Index, the average Microsoft 365 user already spent 57% of time communicating and 43% creating; 68% said they lacked uninterrupted focus time. Inefficient meetings ranked as the top productivity disruptor. Teams meetings and calls per week had roughly tripled since February 2020.

So the default remote operating system is this: unclear ownership, then more meetings to compensate, then after-hours chat to compensate for the meetings, then managers reading online-status as commitment. That loop is what people experience as “low trust” and “poor culture.” It is not a personality problem.

Nicholas Bloom, Ruobing Han and James Liang’s randomised trial at Trip.com, published in Nature in June 2024, is the cleanest evidence we have that location is not the productivity variable people think it is. 1,612 graduate employees in Shanghai (395 managers, 1,217 non-managers) were assigned by birthday parity to five office days or three office days plus two at home. Over six months, attrition was 33% lower in the hybrid group. Performance reviews, promotion rates over two years, and lines of code for engineers showed no measurable difference. Managers had predicted a productivity hit; they reversed that view by the end of the trial. Resignations fell among non-managers, women, and people with long commutes, not among managers.

Fully remote work is a different animal from two days at home. Bloom has said as much: most negative findings in the literature are about unmanaged full remote, often in narrowly measured jobs, not about designed hybrid. Treat them as different designs.

Gallup’s location splits (Q2 2026, US full-time employees) put exclusively remote at 37% engaged and 52% thriving, hybrid at 34% and 54%, on-site remote-capable at 29% and 49%. Overall US engagement was 31%. Global figures in Gallup’s State of the Global Workplace series have moved: exclusive-remote engagement was reported at 31% in the 2025 report against 23% hybrid, then later global snapshots showed exclusive remote still ahead but falling (25% vs 24% hybrid in one 2026 write-up of the 2026 report). I would not collapse those years into one “remote people are more engaged” slogan. The consistent finding is narrower: among remote-capable roles, forcing the office is associated with lower engagement than giving people a designed remote or hybrid arrangement, and the manager still accounts for about 70% of the variance in team engagement. Location is not the lever. The manager is.

Eurofound’s EWCS 2024 first findings put full-time telework at 3% of EU-27 employees, regular hybrid telework at about 9%, occasional telework at 16%. Hybrid and occasional telework come with more work in free time and more contact outside hours. Europe’s labour market is not “everyone is remote.” It is a minority of teleworkable jobs, with boundary strain as the documented cost. If you lead a Europe-based team, your problem is often hybrid plus always-on, not pure distributed isolation.

Trust is ability, benevolence and integrity, not a feeling

Roger Mayer, James Davis and F. David Schoorman defined trust in 1995 as the willingness to be vulnerable to another party’s actions based on the expectation they will perform a particular action important to you, irrespective of your ability to monitor or control them. Trustworthiness is perceived ability, benevolence and integrity (ABI).

That definition is useful because it tells you what to stop doing. Monitoring is the opposite of trust on their terms. If you need the green dot, you have not built trust; you have built a surveillance substitute. Ability is “can they do this job.” Benevolence is “will they consider my interests, not only theirs.” Integrity is “do they adhere to principles I find acceptable.” Remote work starves all three of the informal evidence the office supplied. You have to manufacture the evidence.

Ability: make competence visible without theatre

Ability cannot be inferred from meeting fluency. Quiet people in a second language look less able on a 12-person call than they are. I would put ability evidence in artefacts: design docs, decision logs, pull requests, customer tickets closed, forecasts that later match reality. A weekly written status that answers “what moved, what is blocked, what I need” is a better ability signal than a standup where everyone performs busyness.

The failure mode is the demo culture that rewards whoever can narrate. If your only review surface is a call, you will systematically underrate people in later timezones and overrate people who share your accent.

Benevolence: the manager has to go first

Benevolence is the piece remote work kills fastest, because small favours and context disappear. Gallup’s prescription is unromantic: one meaningful conversation per week with each direct report, covering progress, barriers, wellbeing and collaboration. Only 22% of employees strongly agree they received meaningful feedback in the last week. That weekly conversation is how you demonstrate you will not use distance as an excuse to ignore load.

Fully self-determined hybrid schedules look kind and are not. Gallup finds people with fully self-determined hybrid calendars are 76% more likely to name burnout as their greatest challenge, 57% more likely to name reduced work-life balance, and 52% more likely to struggle to meet customer needs. Benevolence here is constraint: the team agrees when people must overlap, and the manager defends focus time against the 275-ping day.

Integrity: same standard, same career path, written down

Integrity collapses when on-site people get the interesting work. Six in ten exclusively remote workers say they would be extremely likely to look for a new job if remote flexibility were removed. Proximity bias is not a vibe. It is the promotion file that fills with hallway comments about people the committee actually sees.

Gallup says reinforcing communication, accountability and development can raise trust by nearly 30 percentage points. I read that as: publish the performance standard, apply it to people in the office and people who are not, and put development conversations on the calendar so they cannot be “forgotten” for whoever is off-camera.

If you need a test: would you still know this person was excellent if they never spoke in the all-hands? If the answer is no, your trust system is still theatrical.

Accountability metrics that survive a twelve-hour lag

Activity metrics (messages sent, hours online, meetings attended) are anti-metrics. They reward the infinite workday Microsoft documented. Outcome metrics have to be lag-tolerant, jointly observable, and hard to game by performing busyness.

John Doerr’s OKR rule is still the right filter: key results must describe outcomes, not activities. If the KR contains “consult, help, analyse or participate,” it is an activity. Completion of all key results must actually deliver the objective. For a distributed team I would cap a person at three objectives and three to five key results per cycle. More than that and you have reconstituted the status meeting as a spreadsheet.

What to measure, by work type

A software team can use cycle time, change fail rate, escaped defects, and customer-facing reliability, not stand-up attendance. Trip.com could count lines of code because they had an RCT and a two-year review process; I would not use lines of code as a management KPI. Bloom’s team used it as a research check that hybrid engineers were not shipping less. Using it as a target is how you get unreviewable diffs.

A customer team can use time-to-resolution, repeat-contact rate, and CSAT or CES on a rolling four-week window. A policy or research team can use artefacts shipped against a published calendar: briefs, models, decisions recorded. A sales team already has a pipeline; the remote-specific failure is crediting only the people on the HQ call.

Every metric needs an owner, a refresh cadence, and a “what happens if it is red” rule. Red without a conversation is surveillance. Conversation without a metric is therapy.

The weekly operating cadence I would actually run

This is a sequence, because order matters.

Written goals visible to the team, not locked in a manager’s doc. If two people cannot see the same priority list, you do not have alignment. You have folklore.
A team working agreement that states overlap hours, response-time SLAs by channel, and which decisions require a meeting. Gallup: team-designed hybrid norms get 91% fairness ratings versus 73% when leadership imposes the calendar.
One weekly 1:1 that is not a status dump (status is written). The call is for judgment, load, and career. Skip this and you will discover problems in the attrition survey.
A decision log. Date, decision, owner, dissent. Remote teams re-litigate constantly because the hallway never happened.
A review that looks at outcomes and at who got airtime. If the same two people spoke in every meeting, the process is selecting for extraversion and timezone, not quality.

If a metric cannot be inspected asynchronously, it will collapse into a meeting, and you are back in the 57% ad-hoc-call regime.

What “getting it wrong” costs

If you measure presence, high performers in other timezones leave. Bloom’s hybrid treatment cut quits by a third, concentrated in the groups most harmed by commute and rigidity. If you measure nothing, you get the other failure: the person who has been blocked for nine days because asking for help on Slack feels like an admission. If you measure too many things, people optimise the dashboard. Fatal versus fixable: a bad quarterly OKR is fixable in one cycle. A promotion cycle run on proximity is not. That file is already in compensation.

Contractors and vendors need the same outcome objects and must not be excluded from the decision log. Excluding them is how you get two cultures on one backlog. Employees with caregiving constraints need the overlap window to be real and short, not “be flexible” as a synonym for “always available.” Cross-timezone teams should default to async documents and reserve live time for conflict and invention, which is Edmondson’s split: remote is adequate for routine check-ins; ideation and structuring from nothing are substantially better face to face.

Psychological safety is lower on a call, so you have to over-structure it

Amy Edmondson is blunt: distributed work inhibits candour. Ease of speaking up is mechanically lower on remote formats. The remedy is not “be nice.” It is process: go around the group, use names, make silence more expensive than a short comment, use chat on purpose. Leaders must be process managers, not hosts of an organic conversation that never happens.

Psychological safety is not comfort and it is not winning the argument. Edmondson’s useful correction: disagreement from a leader is not the removal of safety; it is often evidence that disagreement is allowed. Safety is permission for respectful candour, not protection from feedback. It is also team-local, not a corporate programme. You cannot decree it from the CEO. You can run your own team so that questions and bad news arrive while they are still cheap.

Google’s Project Aristotle (2012) put psychological safety first among the factors separating high-performing teams. That finding is still used as a slogan. The operational content in a digital team is: who can halt a release, who can say the estimate is a lie, who can say the customer evidence does not support the roadmap, and whether that message can be sent in writing without a political tax.

I would make three behaviours non-optional in remote meetings: a named round for dissent before a decision, a written recap in the last five minutes (Microsoft found 55% of people said next steps were unclear and 56% found meetings hard to summarise), and a rule that cameras are a team choice for the type of meeting, not a loyalty test. Camera-on for a 90-minute status call is how you fake presence. Camera-on for a one-to-one about load can be useful. Do not confuse them.

Inclusive culture is a protocol, not a pizza Friday

Culture in a distributed team is the set of rules that determine who gets information, who gets stretch work, and whose clock the team treats as default. If those rules are unwritten, they default to HQ.

Hofstede’s dimensions still predict the collisions. High power-distance colleagues will not contradict a manager on a recorded call. High uncertainty-avoidance colleagues will want the working agreement in writing before they will move. Collectivist colleagues may treat public critique in Slack as an attack on face. Earley and Mosakowski’s work on hybrid team cultures is the practical warning: highly heterogeneous teams do not magically average into a third culture; they need an explicit team culture or they fragment.

Write the protocol.

Overlap hours: a two- or three-hour window, not “core hours 09:00–17:00 HQ,” which is colonialism with a calendar. Channel SLAs: chat for things that can wait four hours, email or tickets for things that can wait a day, phone for outages. Language: working language stated, and documents preferred over speech so non-native speakers can be precise. Recognition: public for outcomes, private for correction. Religious and national holidays: a published team calendar, not “use your PTO.”

Eurofound’s 2024 e-survey found men had slightly more access to hybrid work than women, and for the first time more women than men working entirely at the workplace. Desire to work exclusively from home rose from 13% in 2020 to 24% in 2024. If your “inclusive culture” is an office social calendar, you are selecting against the people hybrid work retained in Bloom’s trial.

Gen Z loneliness shows up in Gallup’s hybrid notes as a risk when people come to an office and still do not interact. Forced office days that are just laptop farms are expensive isolation. If you spend the money to colocate, use it for the work Edmondson says is worse on a call: divergence, planning, teaching.

Fully remote companies should budget periodic in-person time as a project cost, not as a loyalty ritual. Hybrid companies should stop pretending that “three days in” is a culture strategy if those three days are not designed. Team-designed schedules, Gallup’s 91% fairness number, are the better default.

The leadership job is smaller and stricter than the posters say

Managers account for 70% of the variance in team engagement. That is either a burden or a relief. You cannot wait for a corporate “digital culture” programme. You can run ABI evidence, a handful of outcome KRs, a written working agreement, a weekly 1:1, and structured speaking rights on calls.

What I would not do: mandate cameras as a trust policy; use after-hours responsiveness as a high-performer signal when Microsoft already measures 50-plus messages outside core hours and a 16% rise in meetings after 20:00; run hybrid as “come in whenever you like” when that pattern is the one Gallup ties to burnout; treat OKRs as a cascade of activities.

What I would do in the next 48 hours, if I were taking a distributed team: publish the three outcomes that define this quarter; kill one recurring meeting whose artefact can be a document; put a 25-minute 1:1 on the calendar for every direct report this week; ask the team to draft overlap hours and channel SLAs and then sign them. Fairness ratings jump when the team writes the rule.

The expensive mistake is still the same one the 54% of distrusting managers are making: looking for proof of work in the wrong place. The proof is the artefact, the metric, and the person telling you bad news while you can still act. If you only see them when they are lit up on a grid of faces, you are not leading a digital team. You are supervising a broadcast.

High-Impact Remote Leadership & Digital Team Alignment