1) Run a compensated relational program that reached more low-propensity voters than the margin in 2019

2) Execute a head-to head test against paid canvassing to evaluate relative cost and efficacy in reaching targeted voters

Mississippi 2023

PROGRAM GOALS

by the numbers

1,729

Presley Ambassadors hired

72,501

contacts mapped in rally

$1.93

cost/relational touch

$11

cost/unique matched voter

participants

87% Black

88% BIPOC

network

82% Black

83% BIPOC

52% did not vote in 2019

52% did not vote in 2022

phase 1: recruitment

Target universe: Black voters with Dem support scores higher than 70, turnout scores lower than 60

Homophily effect: recruit the voters you want in your network

Phase 2: training

Virtual onboarding session: Included civic literacy training, education on the governor’s race, training on relational organizing

Participants joined Rally, our relational platform

phase 3: mobilization

Map 40 contacts in Rally and match themto voter file; make vote plans

Attend a Get Out the Vote Palooza

Optional: Attend an Election Day Bonus Palooza (1279 mobilizers attended!)

results: program reach

This program reached more low-propensity Black voters than the margin in 2019, contributing to the closest gubernatorial race in MS since 1999. Our network was 82% Black, and over 50% didn’t vote in 2019 or 2022.

A fun fact:

Brandon Presley flipped 3 counties that Dem candidate Jim Hood lost in 2019: Grenada, Forrest and Lowndes

  • Margin of victory across flipped counties: 1,982 votes

  • Our relational network across flipped counties: 2,434 voters

Learning: compensated relational continues to prove effective for reaching low-propensity voters and people of color.

results: relational vs. canvass

Relational program:

  • Cost per unique voter reached was $11

  • Cost per unique targeted voter reached was $28

Paid canvass:

  • Cost per unique targeted voter reached was $38

Learning: Compensated relational reached targeted low-propensity voters at a 26% cheaper cost per voter than paid canvass.