Comscore Adds Voter Data To Digital Media Targeting, Can Segment 'Behaviorally'

Audience measurement giant Comscore, which has enabled TV advertisers to analyze TV viewers based on their political affiliations for years, this morning announced it will enable advertisers to segment and target consumers in digital media based on their political affiliation, as well as “behavioral political data.”

The deal, which licenses explicit voter data from L2 Inc., is drawn from state and/or county boards of election and it includes the name, address and voting history of the voters they’ve registered.

Comscore said the “behavioral” data is based on models that are used to “help micro-target” the voter files, including each individual’s “propensity to support/oppose an issue, candidate or behavior.”

Comscore said that modeled data can be used  to create audience targeting segments that “go beyond party affiliation to capture behaviors and opinions, such as opinions on canceling student debt, climate change and COVID vaccine as well as behaviors such as church attendance, charitable-giving and voting in-person or ballot.”



The data is being offered as part of Comscore’s “Predictive Audiences” service to “create first-of its-kind cookie-free political audience” segments that could be used to target consumers via digital media such as desktop and mobile internet, CTV (connected TV) and podcasts available through demand-side platforms (DSPs).

Asked if the data could be used to target consumers by advertisers in non-political categories, a Comscore spokesperson said, “Yes, there are many applications for these segments outside of just supporting political advertisers,” and offered three examples:

  • “A transportation brand can reach users who have ‘resumed commuting to work’.”

  • “Streaming providers looking to acquire new users with their movie selection can reach ‘movie theater avoiders’.”

  • “A mission-driven brand looking to appeal to consumers who are likely to ‘give to charity’.”

6 comments about "Comscore Adds Voter Data To Digital Media Targeting, Can Segment 'Behaviorally'".
Check to receive email when comments are posted.
  1. Dan Ciccone from STACKED Entertainment, January 13, 2022 at 10:05 a.m.

    Name, address, party affiliation is public. "How" we vote is supposed to be secret/private. What kind of "modeling" does Comscore use to figure out how people are leaning on a subject or do they just fall back on "proprietary algorithms" like every other data company without disclosing how they come across this info?

    According to C/net, 41% of US adults use adblockers as well, so how reliable is this information and how is it actually evaluated and checked?

  2. Colleen Jones from Martin Communications Inc. replied, January 13, 2022 at 11:19 a.m.

    You're right about how we vote being secret. But I believe it's public to see the person's voting history (e.g., "voted in last 5 years," "voted in most recent local/national election," etc.) But then they'd have to layer on other behavioral models for that other data.

  3. John Grono from GAP Research, January 13, 2022 at 4:41 p.m.

    I struggle to see how applying "macro" data will help with "micro targeting".   Smart marketers and media specialists will be wary as to the accuracy of such modelling.   I look forward to Comscore's detailed explanation of how this will work with the precision that micro targeting requires.

  4. Ed Papazian from Media Dynamics Inc, January 13, 2022 at 6:35 p.m.

    John I'm surprised at your comment. Why not think big first, then bore down into the more "granular" detail. For example what about whether the consumer is left- or right-handed--or ambidextrous? That could be important. Or whether the consumer is  blond, redhead or brunette---or balding---I can see why that might add some broad based context to an analysis.

  5. John Grono from GAP Research, January 13, 2022 at 9:20 p.m.

    Good point Ed.   Sorry to be so amiss.

    I reminds me of that report in the 70s or 80s that found a relationhip during the fashion and clothing boom.   Males were harder to predict than women.   But a correlation between inside leg measurement and your BMI was reported.   It had correlations in excess of R*2 of 0.80 for some of the least expected factors such as whether what they wanted in fabric type and colour, belt-or-no-belt, and most surprisingly whether they wanted cuffs.   Maybe we should add that in.

    P.S. I made that up.

  6. Ed Papazian from Media Dynamics Inc, January 14, 2022 at 2:31 a.m.

    I seem to remember something along those lines,John. For example men who preferred double breasted jackets in their suits were  more likely to be egotists than those who favored single breasted jackets---of course you had to factor in their age, weight and occupation in the correlations. This, combined with other variables, was found to be an important influencer for brands selling mirrors to consumers as well as those marketing certain types of hats---derby hats, if  I recall correctly.

Next story loading loading..