NEWWhat November 3 showed. Registration isn’t turnout. Read it →Sign in
Why Red Glare · the first turnout intelligence platform

The file sees an R. We see who’s fading, and win the vote.

A voter file tells you how someone is registered. It can’t tell you they’ve gone quiet, and an R on the file isn’t a vote in November. Registration isn’t turnout. Red Glare watches what supporters actually do all term, catches the ones going quiet while there’s still time, sends the person they’ll listen to, and lets the voter file say whether it worked.

0slipping, noticed0brought back
The thesis

Campaigns run on inference. Votes come from testimony.

For thirty years the stack has been a file and a model. Both describe a voter from the outside. Turnout intelligence starts from what the voter tells you, and what they do next.

THE FILERegistration

Who’s on the roll. Party, address, history. It tells you who could vote, and nothing about who will.

A list
THE MODELPrediction

A score that guesses who’ll show, built from last cycle, refreshed a few times a year, blind to anything said this week.

A guess
THE TESTIMONYTurnout intelligence

What people actually said and did this term: replies, doors, silences, ballots, plans. Read every day, and acted on by someone they know.

A record, and a reason to act
The evidence

Decades of field experiments point the same way.

Political scientists have run randomized trials on turnout for decades. The findings are consistent: who asks matters more than how often, and voting is social. Red Glare is built on that, not on a hunch.

CANVASSING VS. MAIL VS. PHONE

A person at the door moves people.

Gerber and Green’s New Haven field experiment randomly assigned voters to canvassing, mail and phone appeals. Face-to-face contact raised turnout far more than anything else.

Door, in person+8.7 pts
Each mailer+0.5 pts
Phone callnone detected
Effect on turnout, percentage points

Gerber and Green, “The Effects of Canvassing, Telephone Calls, and Direct Mail on Voter Turnout,” American Political Science Review, 2000. Source →

WHAT WE BUILT FROM IT

So Red Glare never discounts the door. The walking app works with no signal and in full sun, because the porch is where turnout is won.

Doors →
THE NEIGHBORS EXPERIMENT

Voting is social.

In a Michigan primary, about 80,000 households got mailers. The more a mailer made voting feel seen by the neighbors, the more people voted.

No mailer29.7%
Your own record34.5%
You and your neighbors37.8%
Turnout by group

Gerber, Green and Larimer, “Social Pressure and Voter Turnout,” American Political Science Review, 2008. Source →

WHAT WE BUILT FROM IT

Red Glare keeps the social and drops the pressure. A neighbor asks in person, and the page they see never shows party, score or vote history.

Neighbors →

These are published results from peer-reviewed field experiments run by independent researchers. They are not Red Glare results, and we don’t claim them as ours. Red Glare reports only what your own county file shows.

Our methodology

Red Glare runs on GLARE.

Gauge. Learn. Assign. Rally. Ensure. The GLARE Method closes the loop: it wins a commitment from the voter, then learns what worked from votes, not from activity.

The GLARE Method, in full →
The difference

What the old stack counts. What we count.

Red Glare isn’t a better list or a smarter score. It’s a different unit of work: the supporter, from the first silence to the ballot coming back.

Registration and a model scoreWhat people said and did this term
Activity reports: texts sent, doors knockedVotes: whose ballot came back after you acted
One channel, with the others bolted onSix channels on one record, each with its own rules
Software for one cycleA record that runs all term, and survives a new chair
Consent you have to rememberA send without consent doesn’t leave
A dashboard that guessesEvery number opens to its source

Stop watching them slip. Bring them back.

Thirty minutes with people who have run a ground game. Bring your voter file.