Purpose. This is the operational scoreboard for the stewardship function. It defines the 8 metrics that get pulled every Monday, the baselines they are compared against (Blackbaud, M+R, industry medians where public), the red-lines that trigger escalation, and the end-of-Y1 targets. It also documents the retention forecast model — the shape the 48→56→64 Blackbaud curve should take when fitted to LWD’s actual first-cohort behaviour.
Ownership. AI retention analyst (weekly ownership; Monday pull; delta note filed by 11:00 AEDT) · AI CMO (monthly review; spec updates) · Carla (quarterly review + any red-line escalation).
Rhythm. Weekly (Monday) pull · monthly (last Friday) roll-up · quarterly (end March/Jun/Sep/Dec) trend read against Blackbaud curve.
1 — The 8 metrics
Y1→Y2 retention rate on recurring cohort
Source: CRM — count of recurring-active donors 365 days after first gift / count of recurring-active donors at first gift.
Baseline: 48% (Blackbaud Australian charity median for recurring giving)
Red-line: < 42% at end of Y1 (Dec 2027) — six points below Blackbaud median
Target: ≥ 48% by end of Y1 (Dec 2027); ≥ 56% Y2; ≥ 64% Y3
Weekly pull: Trending indicator only during Y1 (the number is not final until Dec 2027); measured as % of Christmas Appeal cohort still-active as of pull date.
Involuntary decline recovery rate (7-day)
Source: CRM + payment gateway — count of donors whose card was updated within 7 days of decline / count of donors who received a save-flow email in the pull-week window.
Baseline: 70% (Blackbaud median for Australian recurring-giving programmes with active save-flow)
Red-line: < 60% for 3 consecutive weeks — triggers save-flow template review
Target: ≥ 75% — enabled by Card Account Updater once payment gateway supports it
Weekly pull: Rolling 4-week average; single-week can be noisy on small cohort.
Voluntary cancellation rate (monthly, annualised)
Source: CRM — count of donors who actively cancelled in the last 30 days / count of active recurring donors 30 days ago, multiplied by 12 for annualised view.
Baseline: 15–22% annualised (industry range for recurring giving; M+R report)
Red-line: > 25% annualised — investigate whether a specific touchpoint is generating cancellations
Target: ≤ 18% annualised (the low end of industry range)
Weekly pull: Trailing 30-day rate; the rolling window smooths week-on-week volatility on a small cohort.
Stewardship email open rate (T1/T2/T3/T6 combined)
Source: Email platform — unique opens / delivered sends over the last quarterly window.
Baseline: 38–44% (M+R Benchmarks for charity stewardship emails; higher than acquisition because the audience is already engaged)
Red-line: < 30% on any single touchpoint type — template review
Target: ≥ 42%
Weekly pull: Only meaningful in months with an active send window (see calendar §1). Otherwise reported as “no send this week.”
Unsubscribe rate on stewardship sends
Source: Email platform — unsubscribes / delivered per send.
Baseline: 0.2–0.6% (M+R median for charity stewardship)
Red-line: > 1.0% on any single send — halt the template, investigate before next send
Target: ≤ 0.4%
Weekly pull: Per-send basis. Reported per T-type when a send has fired.
Complaints received (retention-related)
Source: /enquire inbox + FIA complaints log — complaints tagged as retention/stewardship-related.
Baseline: 0 (aspirational floor)
Red-line: Any complaint mentioning coercion, guilt, or unauthorised charge — immediate escalation to Carla
Target: 0
Weekly pull: Absolute count over trailing 7 days. Any non-zero week is discussed at monthly review.
Q-Impact micro-report click-through rate
Source: Email platform — unique clicks on the “read the full micro-report” link / delivered sends.
Baseline: 8–12% (M+R Benchmarks; stewardship reports typically CTR higher than promotional)
Red-line: < 5% on any single Q-report send — content review before next quarter
Target: ≥ 10%
Weekly pull: Reported in the send-week and the following 2 weeks (long-tail readers).
Downgrade rate (recurring-to-lower recurring)
Source: CRM — count of recurring donors whose scheduled amount decreased in last 30 days / total active recurring, annualised.
Baseline: 4–8% annualised (industry range)
Red-line: > 12% annualised — investigate correlation with campaign pressure
Target: ≤ 6% annualised
Weekly pull: Trailing 30-day rate. Downgrade is honoured immediately per retention playbook §7; the metric measures pattern, not incident.
2 — The retention forecast model
The Blackbaud curve (48→56→64) describes what best-practice Australian charities achieve. Fitted to LWD’s cohort, the curve should look like the below. The retention analyst plots actual vs. forecast every quarter; a persistent gap means the model is wrong, the stewardship is wrong, or both.
RETENTION FORECAST — Christmas Appeal 2026 cohort
==================================================
Baseline: Blackbaud Australian recurring-giving median (48% Y1 → 56% Y2 → 64% Y3)
Assumption: Christmas Appeal 2026 cohort as first-cohort baseline; new joiners
through 2027 anchored to their own T+0 and folded into rolling curve.
Q1'27 Q2'27 Q3'27 Q4'27 Q1'28 Q2'28 Q3'28 Q4'28
(Y1 progression) (Y2 progression)
Retained 100% 82% 72% 62% 56% 54% 52% 50%
% cohort . . . . ↑ . . .
Y2 anchor
target 56%
Red-line 100% 70% 60% 52% 42% ...
% cohort ↑
Y2 red-line
(Blackbaud -6pt)
Notes:
- Q1'27 (Jan-Mar) drop from 100 to 82 reflects the 4-week new-donor risk window.
Involuntary declines dominate this drop; save-flow recovers ~70% of those.
- Q2'27 (Apr-Jun) drop to 72 reflects natural card lifecycle + voluntary
cancellations spread across the second quarter.
- Q3-Q4'27 stabilises as the cohort's remaining donors are self-selected for
commitment (Stage 2/3 lifecycle).
- Y2 anchor is 56% (Blackbaud Y2 median). If Y1 actual lands significantly above
or below forecast, the Y2 curve gets re-anchored at Sprint 55 review.
The curve is a target, not a promise. Overperformance means we adjust the model
upward for the next cohort; underperformance means we investigate the touchpoints.
3 — Escalation triggers
Not every red-line breach requires a Carla escalation. Three categories:
| Category | Trigger | Escalation | SLA |
|---|---|---|---|
| P0 — Trust breach | Any complaint mentioning coercion, guilt, or unauthorised charge (Metric 6). Any unsubscribe rate > 1.0% on a single send (Metric 5). | Immediate to Carla + halt affected touchpoint | Within 1 hour of pull discovery |
| P1 — Systemic underperformance | Any red-line breach 3 consecutive weeks (Metrics 2, 3, 4, 7, 8) | 30-day stewardship review + AI CMO recommendation to Carla | Within 2 weeks of third consecutive breach |
| P2 — Trend drift | Metric worsening for 6 consecutive weeks even if within red-line (any metric) | Note at next monthly review + AI CMO recommendation | Reported at monthly review; no immediate action |
4 — Weekly Monday pull — template
The retention analyst files this template each Monday. It lives in the retention/monday-pulls/ folder in the operations drive. AI CMO reviews the last 4 weeks at monthly close.
MONDAY PULL — Week of [DATE]
Filed by: [retention analyst] at [TIME] AEDT
==================================================
M1 Y1→Y2 retention (trending):
Cohort still-active: [N] of [N_start] = [%]
Delta vs prior week: [+/- N]
On track for 48% Y1 target: [YES / NO / MONITOR]
M2 Involuntary decline recovery (7-day):
Rolling 4-week avg: [%]
Delta vs prior week: [+/- N pts]
Baseline check (70%): [ABOVE / BELOW]
3-week red-line status: [OK / BREACH-1 / BREACH-2 / BREACH-3 → REVIEW]
M3 Voluntary cancellation (30-day annualised):
Rate: [%]
Delta: [+/- N pts]
Baseline (15-22%): [WITHIN / ABOVE / BELOW]
M4 Stewardship email opens (last send):
Send: [T-type + date]
Open rate: [%]
Baseline (38-44%): [WITHIN / ABOVE / BELOW]
M5 Unsubscribe rate (last send):
Rate: [%]
Red-line (>1%): [NO / YES-INVESTIGATE]
M6 Complaints:
Count this week: [N]
Any P0 flag: [NO / YES-see-escalation]
M7 Q-Impact CTR (if send in window):
Rate: [%]
Baseline (8-12%): [WITHIN / ABOVE / BELOW]
M8 Downgrade rate (30-day annualised):
Rate: [%]
Baseline (4-8%): [WITHIN / ABOVE / BELOW]
==================================================
ONE-LINE SUMMARY: [analyst free-text, max 40 words]
FLAGS: [none / P0 / P1 / P2 - with detail]
ACTIONS TAKEN THIS WEEK: [bullets]
ACTIONS PROPOSED NEXT WEEK: [bullets]
==================================================
5 — Data hygiene rules
The tracker is only useful if the underlying data is clean. Four rules govern data hygiene:
- Attribution must survive. Every donor in the CRM has an
attribution_sourcefield populated at first gift (from the Sprint 48 attribution infrastructure). Retention is measured per cohort by attribution source. Undefined attribution = data quality bug, not a legitimate cohort. - Involuntary vs voluntary must be separated. Metrics 2 and 3 are never combined into a single “churn rate.” The CRM tags every cancellation with either
reason:involuntary-[gateway-code]orreason:voluntary-[category]. Reports that mix them are wrong. - Test sends are excluded. Internal test-send addresses (marked
test:truein the CRM) do not count in open/click/unsubscribe calculations. Weekly pull filters these out at query time. - The Monday pull is timestamped and immutable. Once filed, the pull is not edited. Corrections go in the following Monday’s pull with a note. This is an integrity discipline — retrospective adjustment of numbers erodes trust in the tracker.