Weekly brief

Demoville — City Council Ward 3

Week of 2026-10-27

Progress against plan

6 of 14 plan tasks complete (42.9%).

Next week's focus

Shift more volume into the text channel — Text earns replies well above email in your own data.
Lead with the higher-converting message variant — One variant converts to support above another.
Confirm quiet-hours windows before the next phone shift — Compliance item a human must verify.

Contact & response summary

Every count is reproduced from the append-only event store.

Touches by channel (1053 total)

Channel	Touches
phone	50
text	883
email	80
mail	40

Responses

Response	Count
disposition	50
inbound reply	217
opt out	3
reply code	122
sentiment code	8

Phone contact rate: 40.0% (20 reached / 50 dial attempts).

Sentiment trend

model-assisted, 67% agreement on sampled review

Inbound replies classified deterministic-rules-first, then model-assisted; anything uncertain is routed to human review and shown as its own band, never a confident label.

PositiveNegativeNeutralMixedNeeds review

Benchmark position

Your rate beside the platform benchmark for comparable campaigns (same race type and region), pooled through the governed gateway. A benchmark appears only once the cohort is large enough to keep every campaign unidentifiable; below that it reads "not enough comparable campaigns yet" rather than a small-n number.

Metric	Your rate	Benchmark	Cohort
Contact rate	100.0%	26.9%	6
Reply rate	22.5%	24.1%	6
Opt-out rate	0.3%	4.6%	6
Conversion to support	13.9%	48.3%	6

Active experiments

Each experiment pre-registers its metric and holds out a randomized control. A "no measurable difference" is reported as a first-class result, never spun as a win.

Evening vs midday send windowMeasured lift
Status: analyzed
Measured lift: treatment 34.1% vs holdout 16.2%, an absolute lift of 17.9% (95% CI 9.4% to 26.4%, p=0.000, n=270/130).
Subject: question vs statementNo measurable difference
Status: analyzed
No measurable difference on the pre-registered metric: treatment 20.1% vs holdout 19.8% — the 95% confidence interval for the difference (-8.3% to 8.8%) includes zero (p=0.957, n=279/121). Reported as a no measurable difference result, not a win.

Next week's recommendations

Advice derived from your own event data. Every suggestion carries its evidence — your data, a gateway benchmark where the cohort permits, and a narrative signal where one is present — never an unexplained score.

Shift more of the mix into text
Your text channel earns replies at 24.6% versus 0.0% for email. Move marginal volume and budget from email toward text, then re-measure.
Why:
- own_data Your text reply rate is 24.6% (217/883) vs email at 0.0% (0/80).
- benchmark Platform benchmark: your reply rate is 22.5% vs 24.1% across 6 comparable campaigns (through the governed gateway).
Lead with the "warm-intro" message
The "warm-intro" variant earns support-coded replies at 30.0% versus 10.0% for "issue-forward". Make "warm-intro" the default and test a fresh challenger against it.
Why:
- own_data Variant "warm-intro" converts to support at 30.0% (18/60) vs "issue-forward" at 10.0% (6/60).
- benchmark Platform benchmark: your conversion to support is 13.9% vs 48.3% across 6 comparable campaigns (through the governed gateway).
- narrative Narrative signal (Transit: farecap, pilot, across): public discourse on this is active and leans unclear (awaiting human review) (volume 16, resonance 0.30) — frame around what is actually moving. (public discourse sources: Capitol Wire, Civic Signal, Demoville Ledger) [model-assisted, 100% agreement on sampled review]