Customer-Centered Innovations: Choosing Thick Data Over Big Data

You might remember this story from 2012 that underlines the power of big data: Based on the buying behavior of a teenager, U.S. department store Target concluded that she was pregnant and sent her an ad for baby-related products. Her parents angrily called the manager of the branch concerned. A few days later, however, the parents' apology came to the manager. It turned out that Target's data system knew their daughter better than they did themselves.

And indeed, big data has proven its value for many things related to sales and marketing, including for the personalization of promotions, recommendations and personalization of products. Big data creates unprecedented opportunities to quickly deduce "what" is happening from very large numbers of flat data and to make statements and/or predictions based on that. Due to an increasing belief in technology, data and machine learning, it seems attractive to also use big data for innovation.

But that's when it often goes wrong. While big data can tell a lot about "what" is happening, it has big shortcomings in telling you "why" things are happening. And that latter question is exactly the one fundamental for innovation. 

An example is the initially failed product Febreze from P&G. First introduced to test markets in the early 1990s, P&G used TV commercials to teach consumers what smells the product could eliminate and what it could be used on, such as fabrics, furniture and other household items. The company thought it had a winning product, but as months went by, sales dropped. Its leaders couldn't understand why until they talked to customers, who showed them that it was hard to market a product that neutralizes odors to consumers who don't believe odors exist in their homes. This shifted Febreze's marketing and the work its innovation team was doing. Now Febreze is known for its refreshing scents that act almost as a "reward" after cleaning. 

This product innovation and launch is just one of many examples where a different type of data is needed, providing insight into deep-seated human aspirations and needs. 

“Invest in thick data to understand the “why” rather than the “what.” ”

We call the beforementioned data "thick data." It concerns small but very well-picked samples, data collection in the user context and a focus on deeper insights. With thick data, it's about people, feelings, situations, behavior and the rich data that results from research on these aspects. Instead of relying on data analysis in large numbers, insight is gathered by observing people and doing in-depth interviews in which not the "what" but the "why" question is central — and in which the surprise is central by asking further questions.

In my own practice, I use an extensive set of techniques that involve thick data, each with its pros and cons and context in which they are best suited. Let me share some examples:

Job-to-be-done research exposes why a user buys a certain product so that you gain insight into what improvement means for the user. Doing this type of research, a fast-food company might found out that it should not improve or add taste options for its milkshakes but innovate on long-lasting taste, compatibility with cup holders and a straw that lasts.

Unmet need-finding exposes which problems and priorities customers have in addition to the ones solved by your product or service. Thick data provides direction for opportunities to broaden or create completely new services. This is why Netflix makes content now instead of only being a streaming provider. 

Customer journey/experience mapping exposes what feelings customers have when consuming your services. This thick data provides direction in opportunities for better customer experiences. At a large airport in the Netherlands, my company discovered that the experience was affected by a hostile security experience that we were able to revert into a hospitable experience. 

Pretotyping exposes whether (and which) customers understand your product idea and how they perceive the value by displaying a fake version of it in stores or digitally.

Thick data gives companies the ability to deduce "why" things happen from small numbers of rich data. Finding relevant pointers for innovation often requires this type of research and mindset.

Do companies have to forget about big data in order to innovate?

Not at all. The strength lies in knowing when to apply both disciplines and how they can reinforce each other. For example, one can use thick data to provide starting points for new customer problems or needs, which can then be validated and deepened via the support of big data available in systems. Or vice versa: Through constant NPS measurements, one can see trends in customer experience or certain moments when customer experience drops and then find out through interviews what caused them, making designing relevant interventions easier.