When Tricia Wang was working for Nokia in 2009 she had a big idea: the discount phone consumers across emerging markets like China were ready to pay up for better devices. Nokia was the leading mobile phone company in the world at the time and told her “no.” Their data said otherwise. They would neither make nor offer them a better cell phone.
Wang had done a significant amount of local research and produced some interesting evidence, but it wasn’t enough to change the business strategy on. Her hundred person samples paled in comparison to their million-person big data projects. The rest is history – Nokia missed the opportunity, lost their market share, and was acquired by Microsoft a few years later for a fraction of their prior valuation. What did Wang see that they couldn’t understand?
In “Why Big Data Needs Thick Data,” Wang explains how “big data” includes large sample sizes with a normalized, standardized, statistical outputs that often include assumptions of stability over time. Thick data is the opposite using small sample sizes with non-normal, non-standard, story-based outputs that change over time. Thick data looks to identify what might tip the big data’s assumptions in a new direction. When properly combined, thick data can help explain big data’s blind spots and vice versa.
In Wang’s work, she had observed the change in trend that Chinese discount phone consumers were undergoing. Because change starts small, it couldn’t show up on Nokia’s largest samples. The only way to understand the context of the emergent change and what it would mean for the company’s future was to combine the two insights. Unfortunately for Nokia, they weren’t willing to do so.
In our own businesses, we can apply Wang’s work as a reminder that we can’t use data without a story OR stories without data. Here are some sample questions to think through:
How do we decide when a client should stay the course and stick to a program despite current discomfort vs. make a change that might appear ill-advised?
How do we design and track client contact frequency if we have limited quality feedback?
How do we measure and test what content is having an impact?
How do we understand client preferences across various age demographics (ex. Boomers vs. millennials)?
There are data points and stories behind everything. We can’t talk about one without talking about the other. Big data and thick data are complimentary. Wang’s insights run far deeper than what we’ve outlined here. If interested, do spend some time with her work. As she says, “what is measurable isn’t the same as what is valuable.” Noting the differences and discrepancies in our own industries is a super power.