Skip to main content


Have something more specific in mind that our existing models don’t address? We work with campaigns and organizations all the time to develop predictive models that answer the questions most important to them.

From the start, our predictive modeling team will work with you to ensure we acquire the most useful data from a representative sample to train and validate every model. And, we provide 100% transparency along the way, supplying detailed validation reporting, showing accuracy by quartile and decile, and giving you confidence in the performance of each predictive model.

Some of our past custom predictive modeling projects:

Named Ballot Model

Likely Supporter audiences for each of the candidates in the election of interest. Custom complement to the national Affinity Model.

Anti-Establishment Model

The anti-establishment model identifies audiences that may be more likely to support candidates perceived to be against the traditional political establishment.

Ranked Choice Candidate Preference Model

Relative Supporter audiences for each candidate and rank in the election of interest. Useful for strategic targeting in ranked choice elections.

Impacted by Opioid Epidemic Model

Custom audiences of individuals likely impacted by the opioid epidemic in a specific geography.

Success Stories

See how predictive modeling makes a difference

i360’s ability to identify and target Ohio voters and accurately predict their leanings on hyper-local, timely issues was the difference in our ability to win over voters by showing we cared about the same things they did.

Corry BlissCampaign Manager, Portman for Senate

6.1 million

voter contacts

21 point

margin of victory

i360 Voter & Consumer Data

Reach the right audience every time