People Are Your Data

Tricia Wang is done facing the blank stare when introducing herself to people. 

“If I’m networking, or at a party, and I tell someone I work in ethnography—they don’t know what it is,” says Wang. “They think I’ve said ‘ethnology’ or ‘anthropology.’” 

It’s part of the reason Wang decided to invent her own vernacular to describe her work: she says she’s a tech consultant who deals with big data and “thick data.” Thick data, Wang says, is essentially her rebranding of qualitative data, and a call for people to be open to the unknown, to get to the insights that matter. The other major reason for coming up with her own term? Credibility—and a maybe a dash of audacity. 

“I wanted to have credibility within 10 seconds in a room of data scientists, to make them want my data,” says Wang. “And not call it puny or a small data set. So I rebranded it thick data. I get an immediate reception to the term, people always ask ‘What is that?’ There’s a little bit of a sexual undertone too if you think about it. ‘Your data is big but mine is thick.’ It’s catchy, it’s memorable, and anthropologically, it goes with this idea that data has layers, and that if you don’t make an effort to understand all of them, you’re probably missing something important.” 

Wang runs a consulting firm, Sudden Compass, that provides strategic planning services and facilitates on-site workshops, training clients’ employees to embrace ethnographic methods. She’s worked with everyone from Kickstarter to Spotify to P&G. She’s also the co-founder of Magpie Kingdom, a firm dedicated to providing insights about the changing Chinese market. Magpie publishes a free weekly newsletter that showcases surprising and timely stories about China’s digital landscape and burgeoning social media subculture, and helps translate those cultural insights into strategic recommendations for corporate clients. dscout met up with Wang on a recent sunny afternoon in Brooklyn to talk about the importance of thick data in the digital age, and the implications it holds for diversity, privacy, and personhood.

dscout: Let’s start with the million-dollar question: do you consider yourself a People Nerd?

Tricia Wang: Yes definitely. I think I've always been a People Nerd. When you’re a child of an immigrant family, you don’t really have a choice but to be a People Nerd. You’re often the one doing all of the work for your family, so you have to understand adult relationships and how to navigate them. I had to learn how to read situations at a young age because I had to—I was the one who was filing paperwork, opening up a bank account. So I was confronted with the realities of society very quickly and learned how to read the nuances of situations.

It seems like that actually ties in pretty closely with the overarching thesis of your work—that in order to fully understand what’s going on in the world, we need to deepen our awareness of what’s going on around us.

There are absolutely parallels. Society and institutions both normalize. They're always trying to figure out how to fit people and things and ideas into boxes. That’s essentially what quantitative data does. It gets rid of context to normalize a data set to make it fit into a set of numbers. Institutions do that too. They take a set of things that are fuzzy, like human interactions, and they try to put them in a box.

There can be a lot of value in normalizing, but it can also be really dangerous, especially when you don’t stop to ask: why are we doing this? What happens to the things that don't fit in the boxes? The people who don't fit into the institutions or who aren't represented in the data set. Are they made invisible? Who and what is being excluded? 

Both individuals and companies fail to ask these questions all the time. When we don’t see the ramifications of what we’ve excluded, we aren’t open to the unknown. And to get to the unknown, we have to embrace thick data.

So what is thick data? It’s a term you coined, right?

Thick data is simply my re-branding of qualitative data, or design thinking, or user research, gut intuition, tribal knowledge, sensible thinking—I’ve heard it called so many things. Being open to the unknown means you will embrace data that is thick. Ostensibly it’s data that has yet to be quantified, is data that you may not even know that you need to collect and you don't know it until you're in the moment being open to it, and open to the invitation of the unknown in receiving that.

It’s a challenging concept, being open to the unknown, but it’s one that comes up in research over and over. We have a pretty good handle on what’s right in front of us, but we’re not always asking questions about what could be happening, what's coming, what might happen. And those are really hard questions to ask. How do you start to dig into that? And not only anticipate the issues that are going to come up, but start to understand how people are going to feel about it?

I think there are a few key things here. The first is that we need to recognize that we’re susceptible to what I call “horizon deficiency.” That’s that inability to see anything other than what’s right in front of you—we can't see the horizon because we’re so set on what's right in front of us. Once we understand that we might be in a horizon deficient mode, we also gain an awareness of how easy it is to design products or make decisions that can lead to what I call “perspective collisions.” That’s essentially the idea that, because you’re in a horizon deficient mode, you’re only designing for a certain set of people and a certain set of boxes. If you design a product that way, once someone who is outside the boxes tries to use it, it doesn’t work.

Take the example of  why women tend to be colder in buildings. It’s because the “default” room temperature setting was calibrated to the average body temperature of males. And male body temperature tends to run higher than females’. It’s the same reason seatbelts and airbags pre-2011 weren’t safe for women—because the dummies in crash tests were always male dummies. If we look, we start to see these perspective collisions all around us, not only in our built environment, but embedded into our social structures and the digital tools that now mediate our lives.

We need to realize that we have to create room for the unknown. We need to widen our horizon and embrace what we don’t know to surface a wider set of perspectives, to prevent “perspective collisions.” And that leads to thick data.

One of the ideas behind your company Sudden Compass is that you don’t have to be a trained ethnographer or anthropologist to collect thick data.

Exactly. We train people to collect big data and thick data. We're not saying that we can turn you into an expert, but we can arm you with a baseline data literacy so you can confidently interact with variety of datasets and data experts. Ultimately that’s much more valuable to the business. Experts are easy to hire, but finding generalists who can work across all kinds of data and apply that skillset flexibly to business problems is much harder. If you think about it, we’re really all born ethnographers. We are all born with the ability to be People Nerds. But society stamps that out of us and tells us what we're doing is not right, and that people have to fit into these boxes. Essentially my role is to help unleash people’s innate ability to be attuned to the interactions happening around them, and then to develop insights based off the data they see. 

And honestly, that makes me a bit of a pariah in many qualitative communities, because I'm basically saying, your job isn’t just to do research. Your job as a researcher is to be a catalyst, a guide who enables the entire company—from the C-suite on down—to get as close as possible to the people the company is serving.

I want to change the infrastructure, and what it means to be an ethnographer or qualitative researcher inside companies, and that angers a lot of people.

Because they think you’re endangering jobs?

Yes—but I think it actually makes our jobs as researchers even more important. You go from just being the one doing the work of gathering the data to leading the team, equipping them with a common language and guiding them to be as close as possible to the customer, the human they’re designing and making stuff for. The outcome is that you’re operating at a much more strategic level for the business, as opposed to playing in your siloed researcher box. It means that our entire work as ethnographers, as thick data experts, is no longer in the downstream, transactional data ghetto. We get to move upstream. We become more  proactive and can help companies look beyond their existing business models. We can bring attention to newer human behaviors and social dynamics, the ones that haven’t been reflected in the market yet.

In a digital era, everyone at a company should be asking: how do I build data models that represents human interaction? And that means ostensibly a data scientist should also be out there collecting human data, interacting with humans. My favorite response from data scientists after they take our Unlock Labs© is when they say, “Wow, you've absolutely freed me to ask human questions.” The best data scientists are brilliant, but most of the time, when they get hired into companies, they're only being asked optimization-level questions. They're asked, "How do I find this group of people within this set"? It’s very, very downstream. They’re not asked emergent questions where they're able to look at data and surface new findings. I hear so frequently: “Wow, I didn’t know that I had the freedom to also talk to people.” And I say: “Your data is people.” Data comes from people at the end of the day. And so why shouldn't you be making sure that your data models are connected to human models?

It feels like we’re hearing so much now about the importance of that firsthand knowledge, of everyone being as close to the customer, to the human, as possible.

Exactly. You have to enable the people closest to the customer to be in these agile teams with the people making the decisions. Too often when a company adds digital technology, like big data or artificial intelligence, it adds a layer of distance between them and the customer. I cite companies all the time who have gone digital, and their stock went down, their company went downhill, because they couldn't use digital in a way that connected them more closely with their customers.

The companies who have figured out how to do that, to use digital to connect with their customers, are, not surprisingly, the ones who are combining those big data insights with human insights. One example is Netflix, and their being able to capitalize on that key insight that not only were people binge-watching, but they were enjoying doing it.

Exactly, and Netflix did a really good job of building that qualitative insight into their quantitative data model. They also recognized that their customer is what we at Sudden Compass call the “Network Customer.”  

Ultimately, people’s social networks are a better indicator of their behavior than the more traditional demographic data indicators marketers often look at—things like age, geography, education level. The whole idea of sorting people and predicting their behavior based on demographic data never really worked anyway. We’re just finally starting to see the diversity that’s always been there, and it’s showing cracks in the old ways of understanding people. The marketplace is finally catching up to that. In a traditional marketing model, you have one identity: you're an urban white male who works in engineering. But what Netflix figured out is that maybe that urban male actually shares more in common with a 70-year-old grandmother who lives in a rural community. Maybe that they appreciate and watch the same show. People have fluid identities. They express different identities depending on the network they’re in at any given moment. That’s always been the case, but social media has opened up a wider range of spaces for expression.

You've done some in-depth work on this, on how we’re starting to understand at a deeper level the way networks and social media are connected to our identities. You’ve done a lot of research on this particularly in China.

I wanted to understand how people use social media to express their identities in a case where privacy is not even a given anymore. In America, we have this concept that we have privacy, though clearly it's been torn down a bit. But in China, privacy is not a given. You grow up without it. But what I wanted to understand was, do they still have anonymity? Anonymity and privacy are two very different things. Anonymity is the ability to do things without people knowing who you are. Whereas privacy is when people know you who are but don't know what you're doing. To me, anonymity is a core interaction in human civilization, because without it, we wouldn't have all new kinds of collaboration. We wouldn't have resources like Wikipedia, we wouldn't have large scale collaborations that meld different kinds of social groups together.

What I found was people have a formal identity, the identity they express with people they know. Which is not actually a very wide range of expression, because often we’re the most scared to express ourselves around our family and friends, especially in places like China. But on social platforms, where they can be anonymous, they play with their identity all these different ways, experimenting with different forms of gender, sexuality, and hobbies. Finding ways to express themselves outside of top-down institutional norms. And that, to me is what’s at risk right now, not privacy. That train has left the station. What I'm most concerned about is personhood. That’s your ability to define your future, to have the agency to decide what kind of life you want to live, what you want to do, and where you want to go. That’s the most important thing to consider in our digital society. And that's where I'm very concerned around what the role of algorithms and machine learning is doing because the people whose personhood is most at risk are those who are marginalized, and those who don't have a say in the system, or whose data has been collected without their input. It’s similar to how institutions are taking fuzzy things and trying to fit them into a set of norms. The same thing is happening with machine learning algorithms, just at a faster rate. But we lack transparency in how algorithmic decision-making programs are translating those fuzzy inputs into quantitative data points.

ProPublica recently did a really great piece on machine learning bias and COMPAS, the software that’s being used in court systems to determine jails people are sent to and release times. The data that the system uses includes all of these people were were historically racially profiled. So its resulting in all of these proxies for race showing up even though race is not a data input. And so because of people’s racial or economic background, they're being determined as higher risk individuals. So they’re kept in jail longer, they're given a higher bail. It’s affecting their personhood and livelihood, and their ability to decide their own futures. That’s what I’m most concerned about.

Ultimately, we don't understand the terms in which we're giving up or sharing our data. People's identities are being negotiated and put at risk over data practices that even the data collectors or analysts themselves aren’t fully aware of.

You need to push yourself and the people at your company to think through how this type of data collection or data tracking affects people. What does it mean for their lives? That’s the currency any people nerd should be concerned about.

Any modern person who calls themselves an ethnographer or a qualitative researcher needs to be both a people nerd and a data nerd. Those things are interchangeable because data is people, people are data. Data and people are inextricably linked; to decouple them is a massive mistake. Researchers don't have the equivalent of a Hippocratic Oath, but if we did have one, I would add that our job is to empower people to understand how their data is being used and that our internal work inside organizations is to be the advocate for people's data, which ultimately means being an advocate for people’s personhood.