Why Your Company Needs Social Scientists
They’re uniquely equipped to map how users interface with digital technology.
The tech industry has always had a curious relationship with social scientists. As a PhD student in sociology in 2010, I remember hearing professors discuss the “Xerox Parc days”: starting in the late 1970s, anthropologists helped technologists encounter their machines in more natural habitats beyond the lab. Since then, social scientists of all kinds — sociologists, economists, psychologists, you name it — began working alongside technology entrepreneurs. Startups who “made it” in the tech sector had done their research, so to speak.
As the industry reached another boom in the 2000s, perceptions of research in the tech sector started to change. Consider Silicon Valley, the HBO show many of us in the tech industry lament is often painfully on the mark. In one episode, the members of this startup were confused as to why people weren’t sticking around after trying their platform. So they held a focus group. The team watched behind a one-way mirror as people shared their opinions about using what they built. The CEO got so frustrated with the feedback that he left his hidden seat from behind the window and stormed in to try to convince them that they were wrong.
Now, on the one hand, many of us researchers found this episode funny because it hit on familiar experiences we go through. Entrepreneurs need to regularly hear from the people using their products but don’t always take this feedback well — and when feedback is ignored, products fail.
But something I’ve found myself spending a lot of time doing, both among technologists and with the broader public, is explaining what social science is and what it does. In Silicon Valley, they showcase focus groups, and another episode references usability testing. When surveys come up, people think of marketing research. In fact, some think focus groups and usability testing are also within the domain of marketing.
It’s important to set the record straight.
How companies actually use social science
Irrespective of how businesses classify research on their organizational charts, be it marketing, user research, or data science, social science is the study of human behavior. It refers to a scientific process of identifying research questions and matching those research questions to an appropriate research design (including what data to collect, how to collect it, and how to analyze it) in order to get the right answers.
Furthermore, the data we use can take many different forms. It can be responses to surveys sent globally; transcripts from in-depth, 1-on-1 interviews, or from focus groups; responses to experiments done in-person in a lab or instead at scale; logs of the use of digital platforms; observations; and much more.
Critically, certain forms of data are not inherently better or worse than one another — as I’ve had to tell some technologists who are convinced they only care about one type of data. These data help social scientists understand different aspects of human behavior. Our training prepares us to match the right data collection and analysis methods to the research questions we need to solve.
To give you an example of what this looks like, consider a common scenario that many researchers encounter in industry. We may be asked, on the spot, to design a study to solve a problem not too unlike what they were trying to figure out in that episode of Silicon Valley: “The product lead tells you that while people are using your new platform, they are dropping off like flies. They want you to hold a focus group to figure out what users are doing wrong, so they can launch a marketing campaign to help with user retention. What do you do?”
So, what does a social scientist typically do?
Well, first, they’d likely take a deep breath to avoid showing how frustrated they are by this all-too-common scenario. “Users” are human beings — and if they’re not using your product the way you intended, it doesn’t make them wrong. Focus groups are also not the solution to everything, and pursuing one, in this case, would complicate how people respond as they will likely react to interpersonal dynamics in a group environment. Further, assuming that a marketing campaign is an ideal outcome before you even know what is causing the dropoff is potentially a sign that the team doesn’t think the product itself needs some work.
After taking this deep breath, a social scientist would ask more questions.
Who, exactly, is dropping off? Is it a particular demographic? What does dropoff mean — at which step in using the platform are dropoffs happening most? Were there any recent changes in the platform that were associated with this dropoff?
A social scientist might then agree with the product lead that this is a research question centering on “why” — why people are people dropping off — and “why” questions can often be best answered through qualitative methods, like in-depth interviews. You could construct a protocol for 1-on-1 interviewing that asks people to walk through their process of using the platform and unpack why they stopped using it.
The product lead now agrees with you, “but we’re running on a short timeline so I think you should interview a few new hires at the company who have been using the platform for the first time. They’ll help us figure this out.”
Here, again, the social scientist takes a deep breath. They explain that thoughtful sampling is essential to ensure the data we get will be helpful. The tech sector has a terrible track record of studying themselves to test their products, which is not just lazy, but it’s a bad social science, and quite honestly it’s often racist, classist, and ethnocentric.
Substantively, this approach will not answer the research question either. You’d want to understand behavioral patterns in retention, including whether there are demographic differences in adoption, and then recruit samples of participants who reflect those differences. You’d always want to ensure you include people who dropped off as well as those who stayed. That way, you can compare interview responses along these categories and illustrate why dropoffs occur relative to those that are retained.
Critical for understanding users
That’s just one example of the sorts of research design puzzles social scientists tackle. What’s also important here, too, is that most social scientists are trained in certain data collection and analysis methods, as expertise in each requires considerable training. Designing an interview protocol and conducting an interview, for example, is really hard. Developing survey instruments that net valuable data is also really hard. Learning statistics to analyze quantitative data is really hard, too. We typically spend years apprenticing with professors to learn these techniques.
Hopefully, I’ve convinced you that there’s a lot more to social science than a focus group. Social science is a process by which we seek to explain human behavior, whether technology is involved or not.
Entrepreneurs need us. We can uniquely map how human populations interface with digital technology; what needs people have that we can reasonably try to solve; the extent to which technologies solve those problems or not, and why.
Hire social scientists. I promise you’ll solve important puzzles together.
Matt Rafalow is the author of Digital Divisions: How Schools Create Inequality in the Tech Era (University of Chicago Press). He is a social scientist at Google and a Visiting Scholar at University of California-Berkeley.