Platform Metrics: Methodology and Rationale
What each figure means, how it is derived, what it connects to in the platform, and where the caveats apply.
Methodology & Rationale · M. Kyle Watson, PMP · June 2026
Every figure on this platform is pre-registered against a measured pilot design with a holdout. We don't ship vanity metrics. When a real pilot runs, the modeled benchmark below is replaced — not adjusted — with the measured result, and the sensitivity range is published with it.
Figures are modeled from published research, not measured customer results. Real customer pilots will replace modeled figures with measured results, including pre-registered holdout comparisons.
Sources cited: Analyst Institute, Gerber & Green (4th ed.), USPS NCOA data, NGP VAN published averages, Qomon canvassing benchmarks.
Contact Efficiency
Modeled from published canvassing benchmarks and practitioner estimates. Not a measured client result. Midpoint of the range is 3.4×, displayed on the platform splash page.
The claim
Intelligence-driven targeting, dynamic routing, and list hygiene deliver between 2.2× and 4.1× more quality contacts per volunteer hour than a standard walk list driven from a static voter file. Quality contact is defined as a completed two-way conversation of at least 60 seconds with the intended voter — the Analyst Institute standard, and the same standard NGP VAN uses in its published canvass-rate averages.
How it connects to the platform
- Intelligence — Community Intake Instrument
Produces the community-scored targeting layer that replaces static voter file walk lists. Smarter targeting lifts the door contact rate from the 18–24% industry baseline toward 35–45%.
- CT Field — dynamic turf routing
Re-cuts walk lists weekly based on field intelligence returns rather than running static precinct walks. Modeled to increase doors attempted per hour from 14 to approximately 19 — a 1.35× improvement.
- Intelligence — list hygiene
Prunes deceased, moved, and wrong-address records continuously using USPS NCOA data. Industry voter files carry approximately 1.2–2% monthly staleness. Reducing the wasted-knock rate from 22% to 8% contributes a modeled 1.18× improvement.
- CT Field — briefing-to-deployment cycle
Targets sub-24-hour turnaround from intelligence update to canvasser deployment. This is a modeling assumption based on information recency principles — no published field experiment isolates this effect cleanly. The contribution is estimated at 20–35% efficiency improvement and will be measured in the first paid pilot with a pre-registered holdout design.
Rationale and sourcing
| Driver | Baseline | Modeled lift | Source |
|---|---|---|---|
| Smarter targeting via Intelligence | Contact rate 21% | 38% → 1.6× | Analyst Institute benchmarks · NGP VAN canvass-rate averages |
| Dynamic routing via CT Field turf | 14 doors/hr | 19 doors/hr → 1.35× | Qomon canvassing benchmarks · practitioner routing estimates |
| List hygiene via NCOA refresh | Wasted-knock rate 22% | 8% → 1.18× | USPS NCOA: ~14–17% move annually, ~1.2–1.4% monthly staleness |
| Briefing-to-deployment <24 hrs | 3–7 days industry cycle | +20–35% (assumption) | Modeling assumption · to be measured in first pilot |
The driver estimates above are partially dependent — better targeting and better routing both reduce wasted travel and cannot be multiplied as fully independent factors. The modeled range of 2.2×–4.1× reflects this uncertainty. The 3.4× midpoint displayed on the splash page is the central estimate. We commit to measuring this against a pre-registered holdout group in the first paid pilot engagement.
Cost per Quality Contact
Component percentages are practitioner estimates of canvass cost structure, not published benchmarks. Baseline: $8.50 blended fully-loaded cost per quality contact.
The claim
Modeled cost per quality contact runs approximately 38% below the published industry baseline of $8.50 per contact — reducing to approximately $5.25. The savings come from recovering wasted volunteer effort, not from underpaying organizers or cutting training. Same budget, roughly 60% more conversations. In a competitive cycle, that is the operational wedge.
How it connects to the platform
- Intelligence — Community Intake Instrument
Eliminates the primary source of wasted cost: sending volunteers to wrong addresses, moved voters, and unresponsive households. Voter files carry roughly 1.2–2% monthly staleness (USPS NCOA data). Reducing wasted knocks from ~22% to ~8% of total effort is the single largest cost driver.
- CT Field — automated turf cutting
Replaces manual GIS-based list preparation with platform-generated weekly turf. Campaigns currently spend significant staff hours on list preparation that CT Field handles automatically. Estimated 80% reduction in turf-cutting labor cost.
- CT Field — canvasser UX and briefing automation
Reduces volunteer attrition and re-recruitment cost. Better-briefed volunteers with clear turf and updated scripts stay longer and perform better. Estimated 40% reduction in attrition-related cost.
- Command — coordinated briefing and debrief
Automates status reporting that currently requires staff coordination time. Estimated 40% reduction in coordination overhead.
Rationale and sourcing
| Cost driver | Share of baseline | Modeled reduction | Platform feature |
|---|---|---|---|
| Wasted knocks — bad addresses, moved voters | ~22% | −65% | Intelligence list hygiene · NCOA refresh |
| Turf-cutting labor — automated vs. manual GIS | ~8% | −80% | CT Field automated turf generation |
| List acquisition and cleaning | ~12% | −50% | Intelligence data layer — in-platform vs. vendor |
| Volunteer attrition and re-recruitment | ~15% | −40% | CT Field canvasser UX · automated briefing |
| Staff coordination — briefing and debrief | ~14% | −40% | Command coordinated operations layer |
$8.50 baseline × (1 − 0.383) ≈ $5.24 per quality contact. Fully-loaded baseline sourced from Green and Gerber cost-benefit analysis (~$19 per vote generated, implying lower cost per contact), NGP VAN published canvass averages, and practitioner estimates. Component percentages are an internal illustrative cost allocation model — not published benchmarks.
The percentage breakdown above reflects practitioner estimates of where campaigns lose canvass budget — they are not drawn from a published cost-allocation study. These figures will be replaced with client-specific cost analysis in any production engagement. The strategic point holds regardless of precise allocation: approximately 30–40% of canvass effort historically lands on the wrong door or the wrong voter. Recovering half of that wasted effort produces the 38% cost reduction.
Counties Covered
South Carolina has exactly 46 counties. Not a model. Live deployments will display real-time coverage — counties, legislative districts, and precincts under active turf.
The claim
Every county in South Carolina is in scope for the demonstration engagement. Intelligence community profiles are configured across all 46 counties for the statewide demo. Live client deployments will surface real-time coverage counts from the platform itself — showing which counties, legislative districts, and precincts are under active Intelligence intake, which are developing, and which are at baseline.
How it connects to the platform
- Intelligence — community coverage map
The Project tab displays a statewide SC county map with intel maturity tier color-coding: navy for vetted, lighter blue for developing, gray for baseline. This visual is powered by the county coverage count. As client engagements add communities, the map fills in and the count grows visibly.
- Intelligence — community profiles
Each of the 46 counties maps to at least one community profile in the demo. Full Intelligence deployments add ward-level and precinct-level profiles within each county, making the 46 figure a floor, not a ceiling.
- CT Field — turf table
The analyst-facing turf table shows community-level intel maturity and voter contact progress for each assigned ward. The 46-county figure anchors the statewide picture; the turf table provides the local view.
A DLCC director managing a portfolio of SC legislative races needs to know which communities have been profiled and which are dark. The 46-county count signals statewide coverage readiness. In production, the platform surfaces this as a live dashboard — the director can see exactly which of the seven target districts (SD-26 Orangeburg, SD-29 Sumter, SD-30 Marion, HD-74 Richland, HD-96 Lexington, HD-99 Berkeley, HD-116 Charleston) have vetted community intelligence and which need intake work before canvassing begins.
Turnout Lift — Why We Do Not Claim a Number
The platform does not display a turnout lift claim on the splash page. This is deliberate.
| Source | Finding | Why it matters |
|---|---|---|
| Gerber & Green meta-analysis (51 experiments) | 4.3 pp average turnout lift from canvasser contact | This is the literature average, predominantly from lower-salience elections. |
| Cambridge / Green-McGrath-Aronow (2013) | ~2.5 pp face-to-face canvassing | Consistent with a conservative 1–3 pp range for high-salience contexts. |
| Salience-adjusted analysis (ScienceDirect 2024) | Effects attenuated 33–76% in high-salience elections | Presidential primaries are high-salience — real effects are smaller than the average. |
Position when asked about turnout lift
The meta-analytic literature shows 1–3 pp lift in high-salience contexts from quality canvassing. We do not claim a specific figure because we have not measured it. Command's race forecast and Intelligence's targeted group analysis are designed to support the measurement infrastructure — pre-registered holdout groups, treatment and control turf assignment, and outcome tracking through the election results pipeline. The first pilot will generate the measured figure that replaces the modeled one.
Sources
| Source | What it supports in this document |
|---|---|
| Analyst Institute — published canvass benchmarks | Quality-contact definition (60-second completed conversation). Conversation rate baselines 18–24%. |
| Gerber & Green, Get Out the Vote (4th ed.) | Turnout lift context: 4.3 pp meta-analytic mean across 51 experiments. Conservative 1–3 pp range used for high-salience primaries. |
| Green, McGrath & Aronow meta-analysis (2013) | Face-to-face canvassing ~2.5 pp. Confirms conservative primary-context range. |
| USPS National Change of Address (NCOA) program | Basis for voter file decay estimate: ~14–17% of Americans move annually. Implies ~1.2–1.4% monthly address staleness. Industry convention adds deaths and purges to reach ~2% monthly. |
| Qomon canvassing benchmarks (2026) | 15–20 doors per hour in standard U.S. neighborhoods. 10–15 factoring travel and breaks. Confirms 14 doors/hour baseline used. |
| NGP VAN — MyCampaign published averages | Canvass-rate averages and quality-contact standards used across Democratic campaign operations. |
| Eitan Hersh, Hacking the Electorate | Targeting-vs-mobilization framing. List quality effects on contact efficiency. |