Data Science and the Language of Politics

Print Friendly, PDF & Email

As the nation’s presidential election cycle heats up, we’re bombarded with a glut of ads, radio and TV commercials, emails, and direct mail solicitations, all imploring us to vote for one candidate or another and, even more urgently, to give, give, give donations to their campaigns.

With the political season in full swing, I am struck by the use of polls and focus groups to test and refine messages from the candidates – it’s a slow and labor intensive process that relies heavily on the imprecision of human judgment. In an age of social media and digital communications, the selection of language and messaging seems a decade behind.

While politicians of every stripe stake out their positions and hope that they can convince a sufficient number of voters to take them over the top to victory, there’s one thing missing in everyone’s efforts: an understanding of how words work in the human psyche.

The emotional response behind any appeal—whether for a political campaign, to buy a luxury car, or even a certain brand of toothpaste, differs based on the product that’s being pitched, the geographical area of the potential customer, or even in which media the appeal is being made.

This is not an idle observation, but rather something that can be learned by analyzing words and impressions through the use of sophisticated computer algorithms.

The research shows that emotional response is highly important. We’ve discovered that the proper emotional reaction to a message contributes as much as 70 percent to whether that message will have the desired outcome. In fact, the difference between picking the right and the wrong words to convey one’s thoughts can see up to an 8x improvement in positive effects.

Advanced algorithms can develop the most emotionally compelling content for digital communications, words designed to drive action. Think of it as the optimization of content.

We looked at 19 primary categories of emotion, scoring and tagging hundreds and thousands of words over a dozen years to build a database of codified emotional language. We further categorized emotions by intensity. (For example, does a particular word or phrase evoke a quiet or intense emotional response?) We then developed an advanced algorithm that has the ability to calculate the effectiveness of every possible message combination against any target audience, a process that until now was not humanly possible.

To state the obvious, it can get really complex. Emotional response is not just tied to the words themselves; the success of using those words are subject to individual industries and social goals, from toothpaste to insurance, politics or healthcare. Emotional response outcomes are also sensitive to different geographic areas, and even presentation modes.

As just one example, in the finance and insurance industry marketers achieve much better results when they use words that evoke intimacy—words or phrases that imply some sort of relationship—than they would if they used ones designed to eliminate worries or doubts. That’s counter-intuitive, because one would logically think that those interested in buying securities or insurance would want to feel good about the soundness of their investment.

A response to an ad campaign that leads to a positive outcome in the U.S., may have the opposite effect in another country. Or words used in a TV commercial may not work so well in persuading the populace of your position if you’re presenting those same words in a print ad.

When it comes to political outcomes, each candidate in the current election season is trying to make his or her case as to why their ideas are the best. Some see the state of the nation as poor, while others are more sanguine in their outlook. Energy levels differ from bombastic to lethargic. Facts and figures are used to bolster some arguments, while others make broad unverifiable statements.

Email messages can be particularly effective, but with the glut of marketing messages in inboxes, it’s difficult to stand out. To maximize that opportunity, all candidates need to create a compelling subject lines—for example, words that convey “luck” in the subject line tend to perform better than “exclusivity.” Very few are using this particular channel to its potential.

There’s been a long debate as to what sways voters to favor one candidate over another: is it their ideas? Their personalities? Our research points to a third alternative: much like top brands seek to inspire an action from a consumer by appealing to hearts and minds, aligning just the right words and other content, the candidate that is able to tap into true emotional intelligence when communicating with voters, is the one that is likely to win.

assaf-300x300Contributed by: Assaf Baciu, Co-Founder and SVP Product and Engineering of Persado, is responsible for the progression and foresight of Persado’s growing product portfolio and the management of all product advancements. Prior to joining Persado, Assaf was VP of Product for Upstream, where Persado’s core technology originated. Assaf previously worked for speech and imaging solutions supplier Nuance Communications as a senior director of product strategy where he was responsible for developing on-demand and mobile solutions. Assaf joined Nuance from BeVocal, following its acquisition in 2007. Assaf holds an MBA from the University of San Francisco and a Masters in Social Psychology from the Sorbonne in Paris.


Sign up for the free insideAI News newsletter.

Speak Your Mind