What is a corporate wellbeing programme?

An employer-funded service combining mental, physical, financial and social health support for staff. MEM Academy pairs coach-led corporate fitness sessions with a self-serve wellbeing curriculum and ESG/SROI reporting so the spend is provable at year-end.

How much does it cost?

Programmes start at a fixed per-seat monthly fee. A 100-person pilot typically lands between £6 and £14 per seat per month depending on coach hours, on-site cadence and reporting depth.

How quickly can we launch?

Two to four weeks from contract — Week 1 diagnostic, Week 2 design, Week 3 launch comms, Week 4 first sessions delivered. First board report at end of Q1.

What ESG evidence do we get back?

A quarterly board-ready report with workforce KPIs (absence, wellbeing score, participation) plus a verified SROI figure showing social value per £1 spent — suitable for the Social-S section of your annual report.

04 — Client Impact Dashboard — Product Spec

Audience: Engineering (frontend + Supabase), Product, Account Managers Status: Spec for build. Pair with 05-data-model-and-events.md for schema.

Purpose

Give the client HR sponsor a live, anonymised view of the engagement so they can:

See what's happening this week without waiting for a report.
Spot dropping engagement early and intervene with us.
Show their exec / board the value mid-engagement, not just at Week 12.

Audiences & access

Role	Access scope	Auth
Client HR sponsor	Their org's cohorts only	Magic-link email, expires 7d
Client exec sponsor	Read-only summary view of their org	Magic-link email, expires 7d
MEM Account Manager	All client orgs they manage	Standard staff auth + RLS by `account_mgr_id`
MEM Delivery Lead	All cohorts	Staff role
MEM Admin	All data	Admin role

Never accessible: individual line managers within the client org. They get the manager pocket card (Sprint 1 §03) and the W11 debrief — not dashboard access.

Layout (HR sponsor view)

┌─────────────────────────────────────────────────────────────────┐
│  Header: Org name · Cohort · Week N of 12 · Account Mgr contact │
├─────────────────────────────────────────────────────────────────┤
│  ROW 1 — HEADLINE                                               │
│  [SROI Row 1 tile — populated from W12, placeholder until then] │
├─────────────────────────────────────────────────────────────────┤
│  ROW 2 — ENGAGEMENT (live from W2)                              │
│  [Workshop attendance] [Coaching utilisation] [Nudge engagement]│
├─────────────────────────────────────────────────────────────────┤
│  ROW 3 — SENTIMENT & BEHAVIOUR (live from W6, n ≥ 5 only)       │
│  [Confidence delta] [Behavioural deltas] [Pulse response rate]  │
├─────────────────────────────────────────────────────────────────┤
│  ROW 4 — STORIES (curated, anonymised)                          │
│  [2-3 quote cards from coaching, anonymised, coach-approved]    │
├─────────────────────────────────────────────────────────────────┤
│  ROW 5 — WHAT'S NEXT                                            │
│  [Next milestone] [What we need from you] [Download W4/W8/W12]  │
└─────────────────────────────────────────────────────────────────┘

Tile specifications

Row 1 — Headline SROI

Source: sroi_rollups table (one row per cohort per snapshot).
Display before W12: Placeholder text — "Headline impact will land at Week 12. Track progress in Rows 2-3."
Display from W12: Single sentence — "For every £1 invested, this cohort returned £X.XX in value — driven by [top 2 drivers]." Plus link to PDF (Sprint 3 §03).

Row 2 — Engagement (3 tiles)

Tile	Metric	Formula	Threshold
Workshop attendance	% of invited who attended	`attended / invited`	Green ≥80%, Amber 60-79%, Red <60%
Coaching utilisation	% of expected sessions delivered	`delivered / (participants × expected_sessions_by_week)`	Green ≥75%, Amber 50-74%, Red <50%
Nudge engagement	% of nudges opened/clicked	`engaged / sent` (last 14d rolling)	Green ≥40%, Amber 25-39%, Red <25%

Each tile shows the number + a 6-week sparkline.

Row 3 — Sentiment & Behaviour (3 tiles, n ≥ 5)

Tile	Metric	Source	Display rule
Confidence delta	Mean change in confidence (pre vs Day-30)	Sprint 4 §08 pulse + workshop pre-survey	Hidden if n<5; show "+X pts on 0-10 scale"
Behavioural deltas	Top 3 self-reported behaviour changes	Sprint 3 §05 Day-30 pulse Q4-7	Bar chart, % who answered yes
Pulse response rate	% of eligible who completed pulse	`completed / eligible`	Show actual %; informs reliability of above tiles

Anonymisation rule (hard): Any tile drawing from <5 individual responses renders as "Building — needs 5+ responses to display." No exceptions, even if client requests.

Row 4 — Stories

2-3 quote cards. Source: coaching session notes flagged share_anonymised = true AND participant_consent = true.
Cards show: quote (≤30 words), role category (e.g. "Frontline manager, retail"), week of engagement.
Never: name, exact job title, team name, location more specific than country.
Curation: Account Mgr selects from queue; Delivery Lead approves; renders to client dashboard.

Row 5 — What's Next

Next scheduled milestone (pulled from W0 calendar).
Action list for sponsor (e.g. "Confirm W11 debrief attendees by [date]").
Download buttons: Dashboard PDF snapshot (W4, W8, W12).

States

State	When	Display
`pre-engagement`	W-4 to W0	Onboarding checklist only; tiles greyed
`live`	W1 to W12	Tiles populate per the W availability above
`wrapping`	W12 to W13	Board report banner; retro pending
`archived`	W14+	Read-only; banner: "Engagement complete. Contact your Account Mgr to start the next cohort."

Non-goals (explicit)

No real-time chat with coaches.
No participant-facing view (participants use the coaching portal, not this dashboard).
No comparison to other clients' benchmarks in v1 (industry benchmarks ship in v2).
No editable fields — this is a read surface. Edits happen in coach tooling.

Tech notes (for engineering)

TanStack Start route: /_authenticated/client/dashboard/$cohortId
Magic-link auth for sponsor roles via Supabase Auth signInWithOtp.
RLS: cohorts table policy joins on client_org_members for sponsor reads; staff role bypass via has_role(auth.uid(), 'staff').
All tiles are server-fn queries (createServerFn), cached 60s.
PDF snapshots generated server-side via Puppeteer-alternative compatible with Worker runtime (use HTML→PDF via external service; Puppeteer is banned per server-runtime constraints).

See 05-data-model-and-events.md for schema and event spec.