Balancing the benefits and dangers of using digital data to understand human behaviour

Digital spaces are flourishing, with people generating vast amounts of data through conversations, interactions, and social media posts. This data can provide organisations with valuable insights to understand people and changes in their behaviour. However, there are some risks associated with using digital data to draw conclusions about your audience.

We discussed this during Observe Summit with Katie Hillier, Chief Digital Anthropologist at LiiV Center; Hussein Kesvani, Journalist, Media Producer, Researcher & Author; Adrienne Barnes, Founder of Best Buyer Persona; and Marianne Taudiere, Vice President of EMEA at NetBase Quid.

‍

We have access to enough digital behaviour data to see the segments of society clearer than ever. Although traditional segmentation methods were very valuable and lasted for a long time, people are no longer interacting with the world in just one dimension.

Adrienne explained just segmenting audiences by job title, gender, race, etc., is no longer enough. With this approach, you could be laying the foundation for bias. “A more modern way to segment is using the ‘jobs to be done method’ or even just their pain points–what is it that people are trying to accomplish when they’re using your product when they’re living out in the world? It’s actually a stronger way to segment,” she explained.

“To create a modern segment, you need to do it in a way that’s going to understand the human behaviour, understand how people behave, and understand why they made these choices to behave the way they did in the first place.”

Using thick data to understand the “why”

While the world has invested trillions in creating digital spaces, not enough investment has been made in actually understanding how those spaces are affecting what it means to be human. This is where digital anthropologists come in, using thick data to bring context to digital analysis and adding emotions, stories, meaning, pictures, and silence into our understanding of a community or group of people.

Katie explained how big data is incredibly important to tell us what’s going on, where it’s going on, and how things are happening in the world. However, it struggles to tell us “why.” Why is the world changing? Why are communities changing?

“And digital anthropology is really looking to bring those two worlds together,” she explained. “If we’re really going to understand people, we need to understand what’s going on outside at scale with big data and add thick depth in the digital space.”

Using big and thick data together can really help us to better understand communities and groups of people through deeper, more authentic, and probably ethical insight. “If we only look at people through a number, we’re missing so much of the context of who they are,” Katie added. Thick data is necessary to bring in those deeper stories to the information extracted through big data.

Possible risks of analysing online behaviour

Although digital behaviour can provide us with tons of valuable insights to better understand human behaviour, there are some limitations mainly because of the approach used. “The way in which lots of organizations analyse digital life, it can often be very reductive,” Hussein said.

He explained how during his times in the newsrooms, they had very complex tools for data scraping. And then they would use these tools to produce various graphs to anticipate human behaviour. For elections, for example, they would anticipate how people would vote based on these very reductive demographics and scrape lots of data to reinforce certain points. Or they’d use existing platforms or APIs to try and develop models based on that, and those models would usually have very limited effectiveness.

“For example, we’d go to a constituency, where we were told that the local people are concerned with this kind of particular political issue and you’d go down there and interview a bunch of people, and really what they’d say is, ‘I actually didn’t really know much about this political issue happening in Westminster; what we really care about is the roads having too many potholes’ or this very sort of hyper-local issue that wasn’t picked up in that big scrape,” Hussein elaborated.

The reliance on this scale of big data means that those types of small stories - the more nuanced examples of how people interact with each other, the stuff that they care about, and how they express that care - tend to get lost because it becomes an outlier.

“What we need is people who spend more time trying to understand the digital space and how communicative networks operate… Time-sensitive research is needed to extract those stories.”