Online anthropology: audiences and analytics

It ain’t easy being an arts graduate in digital. In the increasingly distant past, I trained as a social anthropologist. Our first week as undergraduates was punctuated with talks about our future career possibilities: lots of mention of the fact that all sorts of companies like Microsoft and Intel employed anthropologists in their head offices in HR roles to better understand their workforces and that anthropology was the future of business.

Aside from a few hundred ethical points about what anthropologists have done actually done for businesses, it is a pretty good argument. Of all the social sciences, anthropology is the best possible training for understanding groups of people, their motivations and trying to make sense of messy and complicated patterns through looking at minutiae and quotidian detail.

Working online, the most obvious way that these skills are most useful is in trying to get an understanding of your users. In all of the organisations I’ve worked in, content teams try their hardest to make sure that they create the right words/sounds/pictures for their audience. But too often they aren’t actually aware of who their audience is and tend to keep users as one homogeneous group instead of breaking them down. When people stereotype large groups, nuance gets lost and that nuance is absolutely key to understanding human behaviour.

There are two really key tools to understanding your users and what they want. The first is to ask them. Set up a good survey on your site, sit down and take time to analyse and read the results, iterate the questions depending on the answers. Nothing better than a whole heap of qualitative data as your users will tell you exactly what’s wrong with what you’re doing.

The second tool is analytics. Aside from being something that you are very careful to type correctly into your address bar, analytics are the cornerstone of evaluating what you’re doing online. Too often though, people get stuck in a loop of only reporting headline figures for management dashboards instead of taking a deep dive into their data.

Last year the NAO very kindly hosted AnalyticsCamp where we talked about quite a lot of these issues, and discovered that to a greater or lesser extent, most organisations have trouble getting exactly what they want from their analytics set-up.

What do we learn?

You can’t tell everything from numbers. You don’t know a user’s name, their story or their thoughts, but you do get a sense of their tracks through the site. It helps you see trends, and to segment your data by characteristics that could make a difference.

Most useful in my experience is looking through keywords from external search, internal search terms, the device split (are you testing for the devices that are actually being used to access your site?) and what networks people are coming from.

This means you can start to build an idea of jobs, times of day and technologies visiting the site. It means you can start to user test with the right people based on evidence rather than hunches.

This isn’t anthropology with a trek through a rainforest, this is anthropology and research design that is focused on the everyday, focused on understanding through qualitative and quantitative data. And it will give you the best start in understanding the people that you’re talking to.

We’ll be coming back more to analytics over the next few weeks as we explore the practical implications of the Service Design Manual on how to build and run a website. Stay tuned for our interpretation of the Discovery Phase┬ánext week.