Our survey research is powered by predictive analytics, which has allowed our firm to outpace the accuracy of the polling/data community at large as it relates to producing outcomes consistent with public opinion and eventually measured at the ballot box or through consumer results. Our approach to methodology and sample construction is dictated by advanced statistical guidelines as determined by our PhD-led team of researchers and data scientists.
“CVIC’s advanced analytics and modeling solution identified 1.6 million new prospective customers based on their degree of similarity to existing customer segments.”
“Public opinion research played an important role in rate case positioning and messaging, as well as making the case for approval of mergers and acquisitions by state regulators.”
Using predictive analytics will make every tactical effort in your campaign more efficient.
By measuring the likelihood of voting behavior and assigning probability scores to each individual voter, the campaign can spend its valuable resources—time and money – on the voters most likely to vote, who need persuading, or need to be turned out. The goal is not to exclude voters, but to prioritize your outreach.
We survey voters until we get enough responses to “train” a model.
Using machine learning our team assigns each voter with a probability score for that model.
If someone has a 1.0 they are very likely your person, 0.0 or negative score means they are very likely not.
We also test support and opposition to issues and messages, which is also on a 0 to 1 scale.
Our data science team finds the reasonable cut offs based on the distribution of the scores and creates universes to make it easy to use for all of the campaign’s voter contacts.
We develop a plan based on the program you choose and the unique characteristics of the organization’s mission. CVIC is powered by Bonfire which is in essence a warehouse of over 3 billion data points on adult-age consumers in the United States, with tens of millions more points added every month.
With Bonfire, we can measure the likelihood of individual consumer behavior. We can monitor issue awareness, issue affinity, and the likelihood of donating. This means we can pinpoint individual consumers who are most likely to donate to your organization.
As part of our modeling, we will gauge consumer response to specific marketing messages or themes and preferences for online or offline giving.
* Donors and customers are not all alike. In most cases there will be multiple donor models and marketing messages, each geared to reach a target group.
We create multiple targeted lists—unique to different marketing messages and based on the probability that prospects will take action.
When a prospect model is created, current donor contact information is entered in our database. We then enter names we to find information including but not limited to:
1. Voter file data – Age, Gender, Party Affiliation, voting history, and Census data (detailed ethnic data, family structure, immigration, key local industries).
2. Consumer data – Organic Food Purchases, Charitable Giving, Pro Wrestling Fan, Wine Lover, Heavy Junk Food Eater, Hunting, Fishing, Cat/Dog Owner, Sports Fans, PBS Donor,Online Gamer, etc.
3. Demographic data – Education, Income, Race, Religion, Marital Status, Social Science Data (local health data, crime rates, unemployment, inequality and social mobility, pollution, migration, religious institutions and affiliations).
Working with you, CVIC can develop an outreach program complete with a series of marketing messages geared to reach your recommended target audience. We enable you to reach out directly to those on the list via text message, or email—and identify which prospects to target first. Using our platform, you can contact these individuals directly, directing them to a specific link to donate online or directly through text.
