The problem with empathy at scale: partial connections in design research
The term empathy is overused and under criticised — it is often intended to mean to understand and/or advocate for customers but holds little conceptual ground if critiqued in the context of thorough social science research. Empathy is less a concept that helps researchers contextualise data — it is more a method of advocacy, than for reflexivity and critical thinking — , and runs the risk of only holding partial truths about interlocutors.
Empathy is considered one step in the design process that may be described as an approach aimed at understanding ‘the user’ and then advocating for their needs by creating emphatic thinking towards them by stakeholders. With the insights gained during the empathy phase, through existing data or some design research methods such as surveys and interviews, design teams may create things like user personas and move on to synthesise these findings and ideate solutions through product, service or program design.
As design processes scale up in complexity, ‘the user’ is no longer a singular entity, but potentially made up of diverse groups of people who are united in their participation in a social system in which a product or service only plays a marginal role.
With larger scale, we may begin to think of the diversity of people who interact with a range of products and services, so not only are the people diverse, but the range of interactions becomes more diverse. We may be interested in the context between media consumption, transportation and financial services, as commuters pay for their train tickets while listening to music and reading the news, all through their mobile phone.
In this context, it becomes increasingly important to critically think of the limitations inherent in the design and research process as it is likely impossible to understand all of the interactions that shape experiences.
This leaves two problems with empathy. First, rather than simply a quick step to gain insights in a design process, this phase should be considered as research that allows designers to make evidence-based decisions based on data. Second, it is also worth considering the limitations of the process at scale, the results or insights it can lead to, and the partial connections that it creates.
This leaves the challenge of gaining valuable insights into user behaviours, concepts and lived experiences at scale when interactions go beyond one product, to a multiplicity of behaviours towards several products and services in social systems.
Originally introduced by Marilyn Strathern, partial connections is a concept she uses to address the difficulties of scale, complexity and (re)presentation in social science research. She argues that researching and articulating complex societies generates knowledge that is always incomplete.
The complexity not only spans across different social groups (think of factors like age, gender, social status) but also geographically and chronologically. As places change over time, so the lived experiences of different groups of people in those locations change.
History and place relate to, and thus shape, contemporary articulations of self. Societies are made up of the complex exchanges and relationships between diverse groups of people, and it is nearly impossible to trace and articulate the complete complexity of those relationships.
So the question is about how to identify and represent such diversity at scale.
The answer: “simplify complexity enough to make it visible”.
To address this, Strathern proposes ethnographic description as a way to articulate findings in a way that goes deep enough into identifying the relationships that make up the complexity and making the articulation of these insights simple enough without compromising their representative credibility, while making them easy to communicate. The skill is to know when is enough, both in research and in articulation.
Acknowledging partial connections means admitting that we can never completely understand the entirety of relationships and concepts our interlocutors hold: we only ever get a small snapshot of the lives and lived experiences of the people we talk to.
Sometimes it is enough to make design decisions based on small user samples because their behaviour is indicative of wider problems. For example, it may be enough to test a prototype with six people if their behaviours become repetitive, or it may be necessary to expand the sampling strategy.
The partial connection is enough because it is very focussed on behaviour when interacting with one product or product feature. The complexity scale is quite low as the intention is to identify behaviours while interacting with one product and does not need to address more complex social concepts.
On the other hand, when conducting discovery research that seeks to identify structural problems on a social and government level within a city or a state, such as identifying the underlying causes of municipal plastic pollution, regional mental health or income-specific financial wellbeing, then the social complexity and scale are much greater because the diversity of people who interact with a number of different products and services has increased.
Emphasising that empathy is informed by data that were generated through a rigorous methodology that rationalises its methods and limitations may hold higher advocacy value than suggesting an emotional argument. This becomes increasingly important when scaling beyond user interactions with one product or service, to complex social systems across several products, services and social groups. The insights of the research have to be complete enough to articulate the complexity of user behaviours, concepts and lived experiences. Yet simple enough to convey insights accurately.
Thinking through partial connections allows teams to develop research approaches that acknowledge biases and limitations of insights by analysing the context of data collection and analysis. This enables researchers to think of the plurality of social groups that interact with designs at scale and focus insights on the partial connections that are evident and relevant.
Thanks for reading this far! If you liked what you read, get in touch with me here to talk about how we can work together and design research-based solutions for your organisation.
Illustrations by Stephen Kung. Find him on LinkedIn and Instagram.
