5 Lessons to Learn from a Disney Research Scientist

The magic of Disney may be partially attributed to Tinker Bell’s sprinkling of faith, trust, and pixie dust, but it can also be chalked up to the company’s reliance on data and research. Maarten Bos, a research scientist for Disney Research, is one of the people who makes this kind of magic (or as the organization calls it, the “science behind the magic”) happen.

Bos describes Disney Research as the “corporate lab” within The Walt Disney Company. The organization conducts research in a number of different areas — such as computer graphics, robotics, and augmented reality (see below) — and presents these findings to The Walt Disney Company to see if it can leverage these insights within different areas of its business, such as its studios or parks.

“It’s really cool when things get picked up,” Bos says, “but there’s never a guarantee that it gets used in the company itself.”

Bos, who holds a Ph.D. in social psychology and previously taught ethical persuasion, studies human behavior. For instance, he conducted a multipart study in August that looked at people’s personality types and how they respond to different images with particular features.

The study began with Bos and his colleagues automatically extracting features from a database of images through computational methods. These images included pictures of people, buildings, and aspects of nature, and the extracted features included things like symmetry, contrast, and lighting. Then, using Amazon Mechanical Turk, Bos and his colleagues asked 745 people to rate how much they liked an image on a seven-point scale (one representing a strong distaste for the image and seven representing an extreme fondness for it). According to Bos, each participant rated a subset of 52 images. These participants were also asked to fill out a personality questionnaire to determine which of the big five personality traits they possessed: openness, conscientiousness, extraversion, agreeableness, or neuroticism. 

Based on this data, Bos and his team built a model that helped them identify which personality types preferred which kinds of image features. To test the model, they showed a new group of people a new set of images with extracted features. Like with the first group, they asked these participants to rate these images and fill out a personality questionnaire. Bos says that they could predict “above chance” which personality types matched the liking of which image features.

“If we know people’s personality,” he says, “[we] can show them an image that they are more likely to like than another image — or likely to like more than another image.”

This insight may seem golden to advertisers and marketers who would be keen to use this data to create targeted campaigns. However, at the time of this interview in April, Bos said that the findings were still “very much” in the research phase and that he had “no idea” whether The Walt Disney Company would pick up this research. 

In an interview with Loyalty360, he said:

“Our studies at Disney Research showed that tailoring advertising could be useful. We did not test advertising effectiveness at Disney (which is our parent company), but we created a model of which personality type might prefer which kind of image features. With that, we can predict above chance level which images will be liked by which kind of personality type. The effects are small, but in a sufficiently large advertising campaign, this could be useful. For our work, you’d need to have access to people’s personality scores, though, and that is a big conditional.”

Bos also says that researchers can look at how people respond to images based on other factors, like age or gender. He declined to share which of his other research findings had been applied to The Walt Disney Company’s various business segments, stating that these kinds of proprietary findings and data often “stay within the house of the mouse.”

Still, there are a number of lessons data-driven professionals can learn from Bos. Here are five takeaways researchers and marketers alike can benefit from.

1. Collect only the data that you need. Aside from knowing people’s personality types and image preferences, Bos says he and his colleagues didn’t collect any “super personal” information. And while he says collecting more data can be beneficial for exploratory analyses, he also says that it’s important for researchers to have a clear idea about what they’re looking for and then focus on answering the questions at hand.

“We make sure that we have data integrity and data security,” he says. “But there’s one way that you don’t have to deal with it: It’s if you just don’t have the data.”

2. Rely on the data, not your intuition. Bos says that he wanted to be “completely data-driven” in the aforementioned study. And even though he hypothesized that he and his team could build a model like the one previously described, he says that he wanted to let the data speak for itself and not rely on his intuition.

“Intuition here doesn’t matter,” he says. “If the data tells you something, then you just go with that.”

3. Be careful how you frame a question in a survey. When it comes to asking questions, Bos says researchers need to be careful how they frame them because they can influence the kind of answers that they receive. If a researcher asks about the weather in one survey question, for instance, and then immediately asks about happiness in the next, the participant’s answer to the first question could sway his response to the second, Bos says. In other words, a sunnier forecast could influence a sunnier mood.

4. Be willing to collaborate. Bos is in favor of having people from different departments apply different types of thinking to a project. A professional with a background in text analysis, for instance, might have a different thought process than someone with a background in computer science, he says, resulting in different perspectives and insights.

“A lot of the gold in innovation comes from two people in different fields combining their efforts,” he says.

5. Believe that data can enhance creativity, not inhibit it. When asked whether data can help or hurt creativity, Bos cited a quote from Walt Disney Animation and Pixar Animation Studios’ chief creative officer John Lasseter: “Art challenges technology, and technology inspires art.” Bos agrees that data can “definitely” inspire creativity and says that it can be hard to make decisions without all of the proper data in place. As he puts it, “It’s kind of great to start with the right data.”

Updated May 18, 2017: The words “and data” were added to this sentence. “He declined to share which of his other research findings had been applied to The Walt Disney Company’s various business segments, stating that these kinds of proprietary findings and data often ‘stay within the house of the mouse.'”

Related Posts