Before You Evaluate a Connection Programme, Audit Its Logistics

A programme manager reports that attendance at the new peer-support lunch group dropped from 24 people to 9 over six weeks and concludes the concept did not land. Nobody checks that week three was the week the room moved from the ground-floor café to a conference room three floors up with no signage. That is the whole evaluation problem in miniature: the outcome measure looks like a verdict on the intervention when it is actually a verdict on a door.

If you are building the case for a workplace connection programme — to keep it funded, to expand it, to report it honestly — you need to separate two different questions before you touch a single survey item: did people show up, and did the programme do something for the people who showed up. Logistics answers the first question almost entirely. Content answers the second. Most evaluations collapse the two and then draw conclusions the data cannot support.

Why this matters more than it sounds like it should

Gallup’s 2024 State of the Global Workplace found that fully remote employees report loneliness at 25%, against 21% for hybrid workers and 16% for those fully on-site — a gap driven substantially by proximity and incidental contact, not programming. A separate cross-sectional study of healthcare workers found that workplace isolation and loneliness are distinct constructs with different correlates, and that perceived social support moderates how remote arrangements translate into wellbeing. Put those together and the implication is blunt: where and how people encounter a programme physically shapes the outcome as much as what the programme contains.

Cigna’s 2020 workplace loneliness report adds the stakes. It found lonely workers miss work twice as often due to illness and five times as often due to stress — so misattributing a logistics failure as a programme failure does not just waste a budget line, it can lead an organisation to cancel something that was working for the people it reached.

The four variables that move attendance more than content does

Log these for every session, every cohort, every location. Not as an afterthought — as the first four columns in your tracking sheet, before any outcome measure.

1. Time of day and day of week. A session moved from lunchtime to 8am will lose parents, carers, and anyone on a commuter train schedule, regardless of how good the facilitator is.
2. Physical distance and legibility. A room that requires a badge, a lift, or a corridor nobody normally walks down adds friction that reads in the data as disengagement.
3. Cost to the participant. Even a nominal charge, or the requirement to buy your own lunch when the session used to include one, changes who attends.
4. Cadence. Weekly, fortnightly, and monthly programmes attract and retain different populations. A shift in cadence mid-programme is a design change, not noise.

None of these are novel insights about human behaviour. They are boring, and that is exactly why they get skipped in evaluation write-ups — nobody wants to explain to a funder that the intervention “worked” but attendance also depended on whether there was a lift.

The calendar invite problem

Name it so your team catches it next time: the calendar invite problem is what happens when a metric moves because a logistics detail changed, and the team reads it as a signal about engagement or programme quality. It shows up as a sudden drop after a venue change, a spike after free food is introduced, or a plateau that tracks exactly with a recurring meeting that competes for the same slot. The fix is not more enthusiasm from facilitators. It is a logistics log kept alongside the outcome data, so that when attendance moves, you can check the log before you write the narrative.

What this means in practice: Do not report attendance or connection-outcome trends without a parallel column recording time, location, cost, and cadence for each session. If a metric moves and none of those four changed, you have a signal worth investigating. If one of them changed, you have your answer, and it is not about the programme content.

Evidence status of the claims underneath this

Claim	Evidence status
Remote/hybrid work is associated with higher loneliness than on-site work	Reasonably well supported — consistent across Gallup’s global workplace data and workplace isolation research
Workplace isolation and loneliness are the same thing	Not supported — treated as distinct constructs with different drivers in the literature
Structured, purposeful group activity produces more benefit than casual contact alone	Supported by qualitative synthesis of social prescribing, though the underlying evidence base for social prescribing generally is described as heterogeneous with limited trial data
Loneliness measurement is standardised enough to compare programmes directly	Not supported — a 2023 review of the loneliness research base explicitly flags inconsistent measurement as a barrier to comparison

That last row matters for anyone trying to benchmark their programme against a published figure. Most of the workplace loneliness statistics in circulation come from different instruments, different populations, and different time windows. Use them to establish that the problem exists, not to set a numeric target your programme is expected to hit.

What this does not solve

A logistics audit tells you whether your attendance data is trustworthy. It does not tell you whether the programme is reaching the right people. Every fix described here — better timing, easier access, lower cost, steadier cadence — improves conditions for people who are already inclined to attend something. It does nothing for the employee who would never sign up for a peer-support group regardless of when it is held, or the one who has decided that admitting to loneliness at work carries a cost. That is a different problem, and no amount of logistics correction will show up in the data as solving it.