Are You Funding a Service or a Community

A commissioner asks a link worker to reduce loneliness among people newly referred by GPs. The link worker refers them to a weekly walking group run by volunteers. Six months later the funder wants to know retention, satisfaction scores, and reduction in loneliness score. The walking group organisers want to know whether they can keep meeting without a facilitator once the grant ends. These are not the same question, and answering one does not answer the other.

That gap is the service-versus-community problem, and most funding applications paper over it because the same pot of money is expected to do both jobs at once.

Two different things wearing the same line item

A service delivers connection to a person. Someone is referred, assessed, matched, and moved through a defined intervention with a start and an end. The organisation running it is accountable for outcomes per participant. This is how most social prescribing is funded and reported: a systematic review of loneliness-focused social prescribing found that all nine included studies reported positive individual impacts, and some reported reduced use of GP or emergency services. That is service logic — input, output, measurable individual change.

A community is a group of people who keep showing up to each other, independent of whether a facilitator or a funder is still there. Ray Oldenburg’s account of the third place — the café, the barbershop, the pub — describes exactly this: informal places that generate their own repeat attendance because people want to be there, not because a programme is scheduling them. Robert Putnam’s account of the decline in American associational life is really a story about that self-sustaining layer eroding. Eric Klinenberg’s argument for social infrastructure treats libraries and parks as scaffolding that lets communities form themselves, not as delivery channels for a service.

The confusion happens because both can look identical from the outside — a group of people in a room, on a schedule, run by an organisation with a grant. The difference is what happens when the funding or the facilitator disappears. A service stops. A community, if it has actually formed, does not.

Why funders keep buying the wrong one

Funders default to service logic because it is measurable in a grant cycle. You can count referrals, sessions attended, and pre/post loneliness scores over twelve months. You cannot easily measure whether a group has developed the internal social capital to survive without you, and most reporting templates don’t ask.

This produces a specific failure mode: the facilitator cliff. The programme runs well, attendance is good, outcome scores improve, the funding period ends, the paid facilitator leaves, and attendance collapses within weeks because the group was never actually a community — it was a service that felt like one. A qualitative synthesis of social prescribing found that participants describe benefit as extending beyond mere social contact to restored purpose and meaningful participation, and that structured, purposeful activity tends to outperform contact alone. That finding cuts both ways: purposeful structure works, but structure supplied entirely by staff is structure that leaves when staff leave.

A test before you write the proposal

Before funding a connection programme, ask which of these is actually true, and design — and evaluate — accordingly.

Question	Service answer	Community answer
Who initiates contact between sessions?	Staff, via reminders or check-ins	Members, unprompted
What happens if the facilitator is sick for a month?	Sessions are cancelled or covered by another staff member	The group meets anyway
Who decides what happens in the room?	The programme design, set in advance	The members, evolving over time
What is the exit plan?	The service ends when funding ends	The funding is scaffolding for something meant to outlast it
What would you measure at 18 months?	Outcomes for the cohort served	Whether the group still exists and who’s still in it

Most proposals answer the left column throughout and then, in the “sustainability” section, claim the right column without having designed for it. Funders should read that section skeptically. Sustainability is not a paragraph; it’s a set of design choices made from month one — who holds the group’s calendar, who has the venue relationship, whether members have each other’s contact details independent of any app or system the organisation controls.

What this means in practice: if you want a self-sustaining community, budget for a handover, not just a delivery period. That means naming who takes over communication and logistics before the grant ends, timetabling the facilitator’s withdrawal in visible stages rather than a hard stop, and measuring group survival at a point after your funding — not just outcomes during it.

The evidence gap this creates

There is a reason so much of the loneliness intervention literature reads like service evaluation rather than community formation research: almost none of it is designed to answer the community question. The evidence base for social prescribing broadly is limited and heterogeneous, and even a favourable systematic review of its well-being effects flags that trial evidence is thin. Randomised evidence on interventions generally is rare enough that a single trial — the HEAL-HOA study testing volunteering against a control among lonely older adults in Hong Kong — stands out precisely because controlled designs are otherwise scarce. None of the available evidence tracks what happens to a group’s survival once paid support is withdrawn, because almost no programme is funded long enough, or evaluated at a distant enough follow-up, for anyone to check.

That is a genuine blind spot, not a minor caveat. If you are funding something on the theory that it becomes self-sustaining, you are funding on a theory nobody has actually tested at scale.

Evidence status of the core claims here

Claim	Evidence status
Structured, purposeful group activity produces better-reported outcomes than unstructured contact	Reasonably supported, qualitative synthesis
Social prescribing referral models improve individual-level outcomes over a programme period	Supported but limited, small heterogeneous trials
Third places / social infrastructure generate self-sustaining informal association	Well-argued conceptually, not experimentally tested
Programmes designed as services can convert into self-sustaining communities post-funding	Largely untested; no direct evidence found

What to actually put in the application

If you’re applying for funding and you genuinely intend to build a community rather than run a service, say so explicitly and budget three things separately: the delivery period, a tapering handover period of roughly equal length, and a follow-up check at 12–18 months after funding ends purely to see whether the group still meets. Funders should ask for exactly this structure and treat its absence as a sign the “community” language in the proposal is decorative.

If you’re commissioning a service — GP-referred, time-limited, outcome-tracked — don’t dress it up in community language and don’t be surprised or disappointed when it doesn’t outlast the grant. It was never designed to. Fund it, measure it as a service, and be honest in the renewal conversation about what continued funding buys.

What this does not solve

None of this addresses who gets into the room in the first place. Both service and community models overwhelmingly reach people who were referred, who sought out the group, or who already had enough social confidence to turn up to a first session. The people most isolated — those without a GP relationship, without a link worker, without the social ease to walk into a room of strangers — are underrepresented in both models, and the service-versus-community distinction does nothing to fix that. It only clarifies what you’re actually paying for once someone is already there.