Politicians as Products: Using MaxDiff to Optimize Political Messaging

This month marks the peak of the midterm election season, with contentious races in many states for control of Congress and state government. With more than $6.4 billion in ad spending already surpassing both the 2018 and 2020 election cycles (according to AdImpact), overall ad spending is estimated to hit $9.7 billion by Election Day in November.  In light of rising pressures and perceived stakes, candidates and campaign consultants are increasingly turning to the techniques used by major ad agencies to hone the message of their “product” — the candidate — to reach the most receptive voter audience. 

 

At response:AI, we have pioneered the use of one of the most powerful techniques for advertising message research, MaxDiff, to help campaigns zero in on the issues and attributes that are most important in the eyes of the voters. 

 

MaxDiff or “Maximum Difference” is a research technique where respondents are shown lists of attributes (product features / candidate messages) and asked which they like best and which they like least. This has an advantage over standard importance scales because it forces the respondent to tell you what is not important, or least appealing to them.

 

Standard techniques often result in a situation where everything is seen as important, whereas with MaxDiff, a clear rank-order preference of attributes emerges from the analysis. Positivity bias is a very real problem in survey research, particularly in the United States, and a rank order technique often helps to ensure honest negative feedback, with can be as important, if not more important, as a means of refining messaging. 

 

MaxDiff is not new. The technique was first described by MW Richardson in “Multimentional Psychophysics,” an article published in Psychological Bulletin in 1938. But the ability to use computer software to design and analyze MaxDiff exercises emerged in the 1990’s from companies such as Sawtooth Software, and it is now built into several survey platforms. Advertisers have used this method for many years to test taglines and product attributes to best highlight in commercials and product packaging. But until recently, it was relatively unknown in the political arena. 

 

Using the MaxDiff methodology, we can have respondents evaluate 10-20 attributes of a candidate and their key issues to identify which are most important to likely voters. These can be both personal attributes of the candidate (family oriented, successful business owner, lifelong resident of the district where they are running, etc.) and the key issues they will confront (tough on crime, focused on health care access, etc.). Rank order preferences can vary tremendously between primary and general elections, and between segments of the voter base (e.g. suburban moms, older Americans, first time voters), making it important to run MaxDiff surveys at appropriate times in the election cycle and with samples segmented to the right populations. 

 

One example of the use of MaxDiff was in the primary campaign of Joe Holland, a candidate for Attorney General of New York. Mr. Holland is a Harlem-based Republican, a Harvard Law School graduate who previously served as State Housing Commissioner in the administration of Governor George Pataki and who has worked on housing, homeless programs, and economic opportunity for over 35 years. The Holland campaign engaged response:AI early in the primary season to understand the candidate’s personal attributes that would best differentiate his candidacy in a crowded primary field. Personal integrity, equal enforcement of the law, and a commitment to reducing regulations on business emerged as the most appealing attributes of a candidate for Attorney General in this survey of Republican likely primary voters in New York state. 

 

According to Holland, “response:AI’s innovative approach to understanding voter concerns helped to position my campaign’s messaging around what mattered most.” 

 

As the 2022 midterm election season draws to a close, we are all bombarded with campaign ads. We will see how this plays out in the shape of Congress and state legislatures, and which candidates got their message to resonate most with the voters. $9.7 billion is a lot to spend without data to back up the campaign’s judgement as to what matters most. This research technique can help us understand the future of our country; imagine what a tool like this can do for your business. The possibilities are maximized.

Contact us today to learn how MaxDiff can help your campaign to achieve its best result.

Politicians as Products: Using MaxDiff to Optimize Political Messaging

Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to top