About two years ago, IBM announced that its artificial intelligence platform, Watson, would be adding “psychologist” to a résumé that includes physician’s assistant and Jeopardy! champion. Here’s how it works: Watson looks through the last 200 posts for a given social media account and uses natural language processing and machine learning to score the user on five different criteria – the Big Five personality traits. These traits are formed by early adulthood and typically remain stable throughout an individual’s life.

It’s easy to write this off as a gimmick – a supercomputer reading palms – but there’s a serious amount of peer-reviewed evidence that shows that individuals’ scores on these traits is a strong predictor of future behavior. Large-scale personality analysis was unfeasible in the past, chiefly because studying an audience on a statistically significant scale would be absurdly expensive and time consuming. Today, however, audience intelligence platforms like People Pattern make it possible for marketers to quickly scan through thousands of social media accounts and create a detailed personality profile for each individual in their audience. This is hugely helpful for marketers who want deeper insights into their customer base. Before we get into the different ways that marketers can put personality data to use, here’s a quick overview of the personality traits that psychologists consider so important.

The Big Five

Openness: People with high scores here love novelty and are generally creative. At the other end of the scale are those who are more conventional in their thinking, prefer routines, and have a pronounced sense of right and wrong.

Conscientiousness: This dimension measures a person’s degree of organization. Those with high scores are motivated, disciplined and trustworthy. Irresponsible and easily distracted people are found at the low end of the scale.

Extraversion: Those who score high for extraversion are companionable, sociable and able to accomplish what they set out to do. Those with low scores tend to be introverted, reserved and more submissive to authority.

Agreeableness: This trait describes how we deal with others. High values show that someone is friendly, empathetic and warm. Shy, suspicious and egocentric individuals score low on the spectrum.

Neuroticism (also called emotional range): This scale measures emotional stability. People with high scores are anxious, inhibited, moody and less self-assured. Those at the lower end are calm, confident and contented.

Definitions from Scientific American

So, how can marketers use this information to make their campaigns more effective?

1. Craft Your Content Strategy

Your content strategy is a vital part of how your customers see you and whether or not they engage with your brand. When marketers tailor their content strategy to the personalities of potential customers, they’re able to create much more efficient campaigns.

Cambridge researcher Sandra Matz recently conducted an experiment to study the effectiveness personality-based marketing. She took a sample of female Facebook users and examined their post histories, using the information to identify introverts and extroverts within the sample. Then she created two different ads for the same fictional beauty product:

Matz targeted both groups with both ads, monitored their responses, and found that tailoring advertisements to personality type significantly improved return on investment for a marketing campaign. ROI doubled when introverts were shown the ad on the right versus the ad on the left, and targeting extroverts with the “extrovert ad” increased ROI by 30 percent.

Still not convinced? Matz and her colleagues confirmed the results in another study, this time using tailored travel package ads across all five major personality traits.

2. Predict Consumer Behavior

Marketers can use links that have been identified between personality traits and purchase habits to predict their audience’s behavior. A recent Eastern Kentucky University study, for example, found a strong positive correlation between impulse buying and a high neuroticism score, and a strong negative correlation between impulse buying and high scores in agreeableness and conscientiousness.

How is this information useful to marketers? Let’s say you’ve identified a segment of your audience that is neurotic and not particularly agreeable. You can reasonably assume that a majority of these individuals are impulse buyers, and with this knowledge, you can implement features (like credit card scanning and internal site search) that have been shown to encourage and facilitate impulse purchase decisions, confident that you’re spending your time and money wisely.

3. Improve Product Design

Often, product design decisions are made with the mindset of establishing or reinventing a company’s image rather than catering to what the audience is looking for. It can be hard to get a handle on what the average customer wants, but personality analysis can give design teams some insight.

A study conducted by Paris Descartes University (the Sorbonne’s medical branch) demonstrated a correlation between personality traits and what people value in a product. For example, individuals who score high on openness tend to “focus more inquisitively on other aspects of products, leading them to disregard aesthetic characteristics,” while those with a high agreeableness score prioritize eye-catching design over anything else.

4. Build More Accurate Data Models

Marketers use dozens of different classifiers to segment audiences, from location to interests to extended demographics like marital status and political orientation. Rarely, though, do people make all of this information publically available online. That’s where probabilistic data comes in: using sophisticated modeling, data scientists (and by extension, marketers) can fill in the blanks with reasonable certainty. Because the Big Five is such an academically established concept, there’s a quite a bit of research linking them to other personal attributes. For example, according to The Economist, individuals who score high in openness and low in conscientiousness are more likely to identify as politically liberal, while the reverse is true for individuals who identify as conservatives. When data scientists are aware of these correlations, they can use personality data to train probabilistic models to generate the most accurate possible predictions about an individual.

Wrapping Up

Personality data are a useful addition to the marketer’s toolbox, and when combined with other information types (expressed interest, demographics, etc.), they can generate actionable insights into any audience.

Example of personality analysis: a look inside the mind of our founder, Ken Cho.