Running a Peer Group Digitally vs. In Person: What Actually Changes

Before you book a room or send a calendar link, decide this: is the group’s job to build relationships between people who already work together, or to give isolated employees a source of contact they don’t otherwise have? The answer determines whether digital or in-person delivery is the right default — and most programmes skip this question and default to whichever is cheaper to schedule.

Gallup’s 2024 workplace data gives a reason to take the question seriously rather than treat it as logistics. Fully remote employees report loneliness at 25%, hybrid employees at 21%, and fully on-site employees at 16%. That’s not a small gap. It means the mode you choose to deliver a connection intervention interacts with the mode your employees already work in — a remote worker joining a video call for peer support is adding another screen to a day already made of screens, while an on-site worker joining the same call is getting something genuinely different from their day.

The two failure modes, named

The Zoom Room problem. You run a digital peer group with good intentions and a real agenda, and by session three it has the texture of every other meeting on the calendar: cameras half off, people multitasking, the facilitator doing most of the talking. Digital delivery inherits the trust and attention deficit of whatever platform culture your organisation already has. A support group is not exempt from Zoom fatigue just because its purpose is nicer.

The Postcode Problem. You run it in person and it quietly becomes a group of whoever is already in the building on Tuesdays at lunch — geographically colocated, seniority-skewed toward whoever has flexible time, and invisible to your remote or shift-working staff, who are statistically the ones with the highest loneliness rate in the first place. In-person delivery, done carelessly, systematically excludes the population most likely to need it.

Neither mode is safe by default. Each has to be designed against its own failure mode.

What changes operationally, mode by mode

Design element	In person	Digital
Group size that works	6–10 comfortably fills a room and sustains conversation	5–7 is the practical ceiling before video calls become one-to-many broadcasts
Facilitator load	Lower — physical presence does some of the work of holding attention	Higher — facilitator must actively manage turn-taking, silence reads as disengagement not reflection
Natural on-ramp	Shared food, shared arrival time, incidental chat before/after	None exists by default; must be built in (a five-minute open mic before the agenda starts)
Who it reaches	Whoever is colocated and has schedule flexibility	Whoever is online at that hour, but reaches distributed and shift-working staff
Cost per session	Room, sometimes catering, staff travel time	Platform licence, none of the physical overhead
Drop-off risk	Lower once someone shows up once, physical commitment carries weight	Higher — leaving a call has no social cost the way leaving a room does
What breaks the group	Format staleness (same room, same seats, same round)	Camera-off culture and multitasking

The table understates one thing: the two formats produce different kinds of relationships, not just different attendance rates. In-person contact carries incidental information — how someone actually looks, whether they laughed at something before the agenda started, whether they lingered after. Digital contact strips that out and replaces it with reliable, low-friction recurrence. Neither is a worse version of the other. They’re different instruments.

A sequenced approach to choosing and running the mode

1. Map where your target population actually is. If the people you’re trying to reach are distributed across sites, shifts, or remote arrangements, in-person delivery structurally cannot reach most of them regardless of how well you run it. Start with the map, not the preference.

2. Pick a primary mode, then build the other in as a supplement, not an equal option. Running “hybrid” as two co-equal tracks from day one usually produces two thin groups instead of one strong one. Choose a primary mode based on step 1, and treat the other as an occasional bridge — a quarterly in-person session for a mostly-digital group, or a dial-in option for people who can’t make the room on a given week.

3. Build the on-ramp deliberately, especially digitally. In-person groups get five minutes of unstructured arrival time for free. Digital groups do not — that has to be scheduled. Open the call five minutes early with no agenda and name it as arrival time, not as the meeting starting late.

4. Set a size ceiling that matches the mode, not the room or the interest list. A digital group of twelve is not a discussion; it’s an audience. If interest exceeds the ceiling, run two groups rather than one large one. This is a staffing cost, and it’s worth stating that plainly to whoever approves the budget.

5. Give the facilitator a different brief for each mode. In person, the facilitator’s job is largely to keep the conversation moving and notice who isn’t talking. Digitally, the facilitator’s job includes managing the platform itself — muting the person with the barking dog, calling on people by name because visual turn-taking cues are weaker, and treating silence as ambiguous rather than assuming it means agreement or disengagement.

6. Decide your cadence before you decide your mode. A weekly digital fifteen-minute check-in and a monthly ninety-minute in-person session are solving different problems — frequency of light contact versus depth of occasional contact. Don’t let the mode decision quietly also decide the cadence; make both choices on purpose.

7. Track attendance by role and location, not just headcount. If your in-person group is 80% site-based staff and your remote workforce is 40% of the organisation, that’s not a neutral fact — it’s evidence the format is excluding the group most at risk, per Gallup’s remote-loneliness figures. Fix it by adding a digital cohort, not by declaring the in-person group a success because it’s well attended.

What this means in practice: don’t ask “should this be digital or in person?” as a single organisation-wide decision. Ask it once per population you’re trying to reach. A single mode rarely serves both your on-site and remote staff well, and pretending hybrid delivery solves this without added facilitation capacity is how programmes end up with two thin groups instead of one working one.

Evidence status

Claim	Evidence status
Remote workers report higher loneliness than hybrid or on-site workers	Solid — large-scale Gallup workplace survey data
Lonely workers miss more work and cost more in absenteeism	Solid — Cigna’s 2020 workplace report, though the effect size is self-reported
Workplace isolation and loneliness are distinct constructs with different drivers	Reasonably supported — direct empirical treatment in healthcare-worker remote-work research
Structured, purposeful group activity outperforms unstructured contact	Moderate — qualitative synthesis of social prescribing, not workplace-specific
Face-to-face socialising has declined independent of remote work trends	Solid on the trend, silent on workplace-specific causes — American Time Use Survey reporting
Digital delivery reaches people in-person delivery misses	Plausible from workplace loneliness-by-mode data, but no direct trial compares reach between formats

What this does not solve

None of the sources here compare identical connection programmes run in both modes with the same population — the evidence is workplace loneliness prevalence by mode, not intervention effectiveness by mode. This guidance is inference from adjacent data, and it should be treated as a design starting point, not a proven protocol.

It also doesn’t solve for the population that shows up for neither format: employees whose loneliness is severe enough, or whose distrust of workplace-sponsored anything is high enough, that no scheduled group — in a room or on a screen — is the right first step. The American Enterprise Institute’s friendship data shows a substantial share of adults, particularly men, now report no close friends at all. A fifteen-minute weekly check-in will not reach someone at that end of the range, and a programme design document that implies otherwise is overselling itself to whoever signs off on the budget.