The First Four Weeks: Sequencing a Connection Programme So It Survives Contact With Reality

Most connection programmes do not fail at launch. They fail in week four, when the novelty has worn off, the facilitator has learned who is difficult to sit next to, and the people who came because someone referred them stop coming because nobody followed up. Call it the week four problem. It is predictable, it is documented in the social prescribing literature as high early attrition, and it is almost entirely preventable with the right sequencing in the first month.

This is not a piece about what activity to run. It is about the order of operations before, during, and after the first session, regardless of whether the activity is a walking group, a shared meal, or a men’s shed.

Week zero: build the exit before you build the entrance

Before you recruit a single participant, decide what happens to someone who stops attending. Most programmes design the warm welcome and skip the warm follow-up, because follow-up is unglamorous and doesn’t show up in a launch photo.

Concretely, before week one:

1. Assign one named person responsible for noticing absence — not “the team,” one person.
2. Decide the trigger: one missed session, or two. Two is more forgiving and usually more accurate, since one missed session is often just a clash.
3. Script the follow-up contact. It should ask how the person is, not why they didn’t come. A qualitative synthesis of social prescribing participants found that what people valued was feeling noticed and having a sense of purpose restored, not simply being reminded of a session — the tone of contact matters as much as the fact of it.
4. Set a data field to log the reason for absence, even informally. You will need this in week four to tell the difference between “lost interest” and “structural barrier” — transport, caring responsibilities, health.

If this infrastructure isn’t built before week one, it never gets built, because everyone’s attention moves to running the thing rather than tending it.

Week one: recruit fewer people than you think you need, and know why they came

The instinct is to over-recruit to buffer against no-shows. Resist this. A room that’s too full on day one and half-empty by day three reads as failure to everyone in it, including the people who did keep coming. Recruit to your real capacity, not your hoped-for capacity, and keep a waiting list instead.

More important than headcount is finding out, on day one, what people came for. The AARP Foundation’s national survey of adults 45 and older found that the strongest predictors of loneliness were the size and diversity of a person’s social network and their degree of physical isolation — not a single missing activity. That means two people at the same first session can have completely different problems: one has a large network but no one nearby, the other has proximity but no network at all. A five-minute conversation at intake — not a form — tells you which. It also gives your named follow-up person something to reference in week three that isn’t generic.

Do not run a formal outcomes survey in week one. It measures a baseline nobody has settled into yet and it makes the first session feel like an assessment rather than an event.

Week two: protect the structure, not the size

By week two, attendance usually dips. This is normal and not yet a signal of anything. What matters in week two is whether the structure of the session survived contact with a smaller or different room. A well-run session for eight people should not collapse into a different activity because only five turned up; if it does, participants correctly conclude the programme isn’t stable, and stability is what keeps people showing up for something optional.

This is where physical setting does real work. Eric Klinenberg’s argument about social infrastructure is that shared physical space itself shapes the rate and quality of contact — a room laid out for structured interaction produces different behavior than the same room laid out for people to drift past each other. If week two’s dip has shrunk the group, keep the seating arrangement and format identical rather than improvising a looser version. Consistency of structure, not headcount, is the signal participants are reading.

Week two is also when you should have your named follow-up person make first contact with anyone who missed week one — not to chase, but to keep the door open. Waiting until week four to notice an absence from week one means you’ve lost them for good.

Week three: introduce the low-stakes task

Programmes that ask people to do something together — however small — tend to hold better than programmes built purely around presence. A randomised trial of volunteering among lonely older adults in Hong Kong found that structured prosocial engagement reduced loneliness more than a control condition, and it is one of the few controlled trials in this literature rather than an uncontrolled evaluation. The mechanism worth borrowing isn’t “volunteering” specifically — it’s giving people a task with a small amount of mutual obligation attached, which is a stronger tie than shared attendance alone.

By week three, introduce something that requires two participants to coordinate with each other outside the facilitator’s direction: bringing something for someone else, pairing up for a piece of the session, checking in on someone who missed last time. This is also the point at which a qualitative meta-synthesis of social prescribing found that participants described benefit less in terms of contact and more in terms of restored purpose and participation — task, not just presence, appears to be what produces that.

Week four: this is your first real read, not your last chance

Week four attendance is diagnostic, not fatal. A drop here does not mean the programme has failed; it means you now have enough data to know who is at risk of dropping and why, assuming you logged absence reasons from week one. Use week four to triage: structural barriers (transport, timing) need a programme-level fix; disengagement needs a personal follow-up; and if the drop is broad and unexplained, the format itself may be wrong for this group.

Resist running a full outcomes evaluation at week four. It is far too early to measure loneliness reduction and doing so risks treating an early operational read as a summative result, which the wider social prescribing evidence base already suffers from — systematic reviews of social prescribing consistently note that most evaluations are short, uncontrolled, and heterogeneous, and a 2025 systematic review protocol flagged that only one peer-reviewed randomised trial exists specifically for social prescribing in older adults. Week four is for programme management decisions, not for the funder report.

What this means in practice: name one person responsible for absence before you recruit anyone; recruit to real capacity, not hope; keep the session structure fixed even when the room shrinks; add a small mutual task by week three; and treat week four as your first diagnostic checkpoint, not your evaluation point.

What this does not solve

This sequence improves retention among people who already walked through the door. It says nothing about who doesn’t show up at all — the referral that never converts to attendance, or the person too isolated to respond to an invitation in the first place. Reach is a separate design problem, upstream of everything here, and no amount of good week-four practice fixes a recruitment channel that only reaches people already somewhat connected.