What is IBM Watson™ Personality Insights?

IBM Watson™ Personality Insights is a service that pulls psychological insights from unstructured customer communication data, using IBM’s proprietary machine learning technology.

Overview

What it is: IBM Watson™ Personality Insights is a cloud-based service that uses machine learning to analyze the emotional content behind unstructured customer communications like emails, social posts, and blogs.

What it does: When given a person’s communications, it delivers a detailed psychological report on their traits and interests. This includes commercially relevant details, such as consumption preferences and consumer values.

Why it matters: By using real communication samples, the service quickly gets a sense of the authentic essence of your current or potential customers. This is an improvement on traditional survey-based analysis, which is more time-consuming, as well as prone to errors in survey design. What follows is more easy and accurate personalization.

What to do about it: If you are an organization heavily dependent on customer transactions, evaluate incorporating Watson™ Personality Insights or similar technologies into your marketing activities. Use it to make sure that you’re effectively addressing your customers’ emotional needs, and to increase the precision of your targeted messaging.

What Insights Does the Service Deliver?

Watson™ Personality Insights provides a reading of each consumer’s psychology based on the Big Five personality model, which measures the traits of Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. It also provides data on consumer needs, decision-making values, and consumption preferences. This is weighed against data from a background population sample, so that you can see how different your subject is from the average.

Implementation Details

Personality Insights isn’t an app in and of itself, although a stand-alone demo version is available for small-scale testing. Instead, it’s a service that can be incorporated into apps. Thus, some custom development is required to put Personality Insights to work.

Things to Consider

Any analysis will only be as good as the data being analyzed. If you use Personality Insights on a language sample that’s unrepresentative of your general customer base, you’ll end up with misleading conclusions. Even though machine learning algorithms do a lot of the work, good statistical sampling practices are still required.

You should consider any technology aimed at improving the accuracy of your marketing activities in the context of an overall marketing strategy, for example taking into account how advertising spend is currently allocated, integration between systems and so on.

Pricing Model

Personality Insights is pay-per-use, which makes it viable for a wide range of enterprises. You’re charged for every time the service is “called,” which is to say, used by an application. Considerable volume discounts are available: after 100,000 calls per month, the price is cut in half, and then it’s cut in half again after 250,000.