How a Bath sleep coach turned a leaky 4-week programme into one members finish

Claire Whitfield ran the same 4-week group sleep programme over email and a shared spreadsheet for three years, and watched it finish, on average, at 38% completion. Her first Bitir cohort — same content, same four weeks, same coach — finished at 89%. The NHS recommends most adults aim for seven to nine hours' sleep a night; the whole point of a four-week programme is to move someone toward that and make the change stick past the end. Email was losing most of her members before the change could take hold. Bitir is not.

Where she started

Claire trained as a behavioural sleep coach in 2020, after fifteen years as a midwife in the NHS where, she says, she watched exhaustion wreck good people. She set up Sound Sleep Collective from her home in Bath in 2021 and now works entirely online, with clients from Cornwall to Aberdeen.

Her flagship offer is a four-week group programme for adults with chronic poor sleep who are not in acute clinical need. Each cohort takes 10 to 14 people and costs £140 per place. She runs one most months.

The programme was sound. Four weekly themes — wind-down routine, light and caffeine, the bedroom environment, and getting up at a fixed time — each grounded in the behavioural principles she draws from Professor Matthew Walker's Why We Sleep and standard CBT-for-insomnia practice. What kept failing was the delivery.

• A weekly themed email, sent Sunday evening, with that week's focus and a short audio.
• A shared Google Sheet where members logged a nightly sleep score out of ten.
• A single group video call each Wednesday.

On paper it was organised. In practice, the spreadsheet was a graveyard by week two.

The breaking point

The moment Claire knew she had to change came in October 2025. A member messaged, embarrassed, to say she had stopped filling in the sheet because "everyone can see I had three terrible nights and I didn't want it sitting there."

That one message named the whole problem. The spreadsheet made a private struggle public, so members hid their worst weeks — exactly the weeks Claire most needed to see.

"My members weren't dropping out because they didn't care. They were dropping out because the place I asked them to be honest was the least safe place to be honest."

She spent that weekend trying two wellness platforms (both wanted to own the video calls and neither did a simple per-member score) before a former client mentioned Bitir. She rebuilt the four weeks the same evening.

What she set up in Bitir

The build took an afternoon. Claire followed the spare, short-format structure we describe in our guide to short coaching programmes, compressed to four weeks.

1. One private group per cohort. Invite-only, members on first-name handles rather than full names, because sleep is personal and handles lowered the bar to logging an honest bad night.
2. One group goal. "By week four, every member has a fixed wake-up time and a wind-down routine they actually use."
3. A per-member sleep score. Each member's goal card tracks a nightly score out of ten, visible to that member and to Claire — not to the rest of the group. Members watch their own seven-day trend line instead of comparing raw scores with strangers.
4. A nightly check-in poll. Three taps, under twenty seconds: last night's score, hours in bed, and one word for how today felt. It opens at 8am and closes at noon.
5. One weekly assignment. The week's behavioural change, a short audio, and a Saturday deadline — following the completion patterns in our assignment design guide.
6. A two-night private nudge. If a member misses the check-in two mornings running, Claire sends one private line. The private-nudge, public-celebration model comes straight from our accountability group guide.

What changed in the numbers

The cohort that started 3 November 2025 had 13 members. Claire ran it exactly as before — same four themes, same Wednesday call, same audio — and compared the results at the end of week four.

The completion number is the one she watches, because a member who quits in week two of a four-week programme has, in behavioural terms, received almost nothing. Halving the drop-off was worth more than any single piece of content.

What she puts the change down to, in order:

1. The private per-member score. Members logged honestly because a bad night was no longer on display. Diary completion nearly tripled, and an honest diary is the raw material of the whole programme.
2. Self-visible trend lines. Each member could see their own seven-day score climbing. "It's the thing that keeps them in week three," Claire says — the week a short programme usually loses people.
3. The two-night nudge. She sent nine across the cohort; seven of those members were back the next morning. On the old spreadsheet, a member who went quiet stayed quiet.

What she would tell another wellness coach

Her advice to other UK practitioners is blunt and specific.

• Keep sensitive self-tracking private to the member and you. A shared, all-visible log punishes honesty on exactly the bad days you most need to see.
• Let members watch their own trend, not each other's raw numbers. Self-comparison over a week motivates; peer-comparison on a bad night shames.
• Keep the nightly check-in under twenty seconds. Claire tried a six-question morning version and watched compliance fall the same week; she cut it back and it recovered.
• Use handles, not full names, for anything as personal as sleep. The same coach might use real names for a professional skills cohort, but the privacy floor matters more here.

What's next

Claire now runs a permanent low-touch "graduate" group on Bitir for members who have finished the four weeks, where a fixed wake-up time is maintained with a weekly poll and a monthly call. She is also piloting a workplace version for a Bristol employer's wellbeing programme, running three parallel cohorts of staff through the same four-week structure.

Related reading: how to design a short coaching programme people finish, and why async check-ins outperform live calls for busy groups.