New technologies unveil unprecedented aspects on modern behaviours.
Ethnography is broadly defined as the “scientific description of peoples and cultures with their customs, habits, and mutual differences”. A staple of disciplines like sociology and anthropology, it invokes methods whereby a researcher mingles with study subjects to learn, as inconspicuously as possible, about how they behave. After enough observations are collected and interpreted, they form the basis for a generalised conclusion about who those people genuinely ‘are’.
Notwithstanding the value of in-loco data gathering, the growing share of online activity mirroring our real lives makes it natural to leverage other, more data-driven methods to grasp communities. It can, in other words, let us tap into a new digital type of ethnography.
“Like traditional ethnography, the digital type can examine as many aspects of our collective lives as there are life habits to be examined.”
Boosted by today’s ubiquitous connectivity, digital ethnography addresses a broad range of questions about our ‘collective selves’. It does so seamlessly, accurately and, more often than not, innovatively (delivering speed and depth that were beyond reach until recently).
With appropriate confidentiality measures in place (permission for, anonymisation and aggregation of individual data points being the non-negotiable ones), digital ethnography bears enormous learning potential. It can, for example, spot and longitudinally track granular pockets of attitudes, like in the illustrative analysis below, which highlights how citizens from different areas of a city are approaching the notion of borrowing money to complement their incomes:
Interest on money borrowing
Above or below country average per month
Add to that type of exploration the ability to set behaviours apart into finer-grained subgroups (men vs. women, high income vs. low income, high-income men vs. low-income women etc.) and digital researchers wield an extremely powerful tool to detect how different spheres of our lives work.
In the same vein, we can infer some of the most primal routines of our day-to-day endeavours. People’s sleeping habits, for example, are a symbolic phenomenon that can be observed though digital ethnographic studies:
% of HH dwellers going to bed in that window
Wake up time
% of HH dwellers waking up in that window
Ultimately, even more abstract interpretations about lifestyle attitudes can be derived, by integrating people’s online activities with other traits inferred about their profiles:
Predominant lifestyle vs. inferred household type
Split by type of household
The possibilities, if responsibly managed, are virtually infinite. Like traditional ethnography, the digital type can examine as many aspects of our collective lives as there are life habits to be examined.
Unlike other analytical fields in which advanced data science methods are required (i.e. behaviour prediction, which is a magnet for machine learning techniques), digital ethnography relies on rather traditional approaches, executed at scale. Intuition, structured thinking and well-guided descriptive statistics are the bread and butter of good digital ethnographic analyses. Those skills are, however, surprisingly difficult to find in today’s data scientists, who sometimes hastily jump into learning elaborate modelling before mastering a more fundamental kind of quantitative sense-making.
Abundant data, in the form of unusually large samples, makes digital ethnography particularly adept at large stratifications. Even after combining dimensions with huge cardinality and creating thousands of sub-segments, overall sample sizes still allow for statistically significant intergroup comparisons, enabling the same behaviour to be surgically examined across small pockets of individuals.
As a relatively young possibility, digital ethnography is still maturing the practical applications it enables. Its biggest advantage is not on allowing more societal investigations at lower costs (although it does do that). Its major strength stems from the new possibilities it opens up: the study of fresh behaviours, across newly-defined segments, with a degree of statistical robustness that was hitherto unattainable.
Better public policy design, finer-tuned business offerings, richer academic research and improved fulfilment of individual needs are a few of the benefits that digital ethnography can inform. It is, however, largely up to the data science practitioners exposed to it to keep asking the right questions and producing the right insights for it. Only that well-educated deployment of ‘common-sense science’ can steer the field towards the full impact potential it can still produce.