Predictive analytics is a suite of tools for voter targeting, issue targeting, and strategy development and optimization that leverages the latest computing, statistics, and data availability.
- We take advantage of the broader & less expensive availability of detailed consumer data & behavioral data on voters.
- We have developed efficiencies to create a lower cost of computing power to run very large datasets and complex models.
- Our team of data scientists constantly develop new, more efficient model fitting algorithms to handle large numbers of variables.
A few of the models we have been building for our clients using Predictive Analytics
We build statistical models to identify:
- Likely donor targets to our client or candidate’s campaign across a state or districts voter database.
- Likely philanthropic targets for our client's non profit initiatives.
We construct statistical models to identify:
- The preference of our client or candidate compared to others across the entire database or registered voters.
- In some cases - this applies to ideological candidates- in which case we evaluate beyond the scope of the voter characteristics - and build a strategy to increase voter turnout to register new voters.
Issue Based Message
We break down segments into groups of people with similar personalities. Messages themed around each group's core issues are designed to resonate with the personality and other unique characteristics for that group.
- To achieve your turnout goals or donation solicitation goals, we develop messaging for your Base Voters.
- GOTV messaging is tested for strategy development around turning out voter groups.
- We design and test compelling messaging strategies to engage your Persuadable Swing Voters.
We examine populations with similar characteristics, and categorize them into sub-groups.
Base Voters are those who are already likely to volunteer or donate to our campaign effort. These are the individuals with high likelihood to vote who select our candidate as their first choice candidate. Messages for these base voters encourage voters to donate to the campaign and/or to volunteer.
Registered voters with turnout probabilities which statistically fall in the swing voter region around who place our client very highly among top preferred candidates are our GOTV targets. Messages for these turnout targets encourage likely supporters to vote.
WPA specializes in developing messages to persuade behavioural changes. This translates to a vote. For instance, Voters with high turnout probabilities who rank our candidate highly but less favorably than other candidates, with our candidate being their second or third preferred candidate. Messages designed for these Persuadable Swing Voters increase the likelihood that certain voters will vote for our client rather than other candidates in accordance with their personality information and issue preferences.
Our Predictive Analytics Tools in Action.
- Here's how we partnered with a candidate to use analytics helped drive a cost effective and impactful media buying strategy.
- For this strategy, we were guided by three basic predictive analytics principles you apply to their campaign: talk to one audience; measure outputs, not inputs; and test and retest.
- Here's how we used Predictive Analytics to help Greg Abbot win his gubernatorial campaign.
- Read how we used Predictive Analytics to drive a field program and improve turnout efforts.