Chapter 13: Introducing the Mathur et al. taxonomy
The taxonomy I prefer to use in my work as an expert witness is the Mathur et al. taxonomy (2019) because it is practical to use and strongly evidence-based. It was created by a group of seven researchers at Princeton University and the University of Chicago and was published in the paper, ‘Dark Patterns at Scale: Findings from a Crawl of 11K Shopping Websites’.1 In their research, the team used a machine-learning algorithm to analyse about 53,000 product pages on 11,000 websites, which revealed 1,818 instances of deceptive patterns. They then took these instances and analysed them in detail, which led to a taxonomy, shown in the table below. This taxonomy is mainly aimed at eCommerce, but it provides some broad categories that can be applied to almost any area of user experience, making it quite flexible.
You’ll notice that I am not proposing a new taxonomy based on the exploitative strategies that we covered earlier. This is because deceptive patterns usually don’t fit tidily under one exploitative strategy or another – they tend to be creatively combined in different ways. For example, a deceptive pattern like ‘trick wording’ as executed on a specific website might take advantage of ‘exploitation of expectations’ and ‘manipulation of comprehension’ together at once, while another designer might do it differently on another website.
When considering a taxonomy of deceptive patterns, it’s tempting to view them as prescriptive – that there are a certain number of types that can be implemented. In reality, there’s no limit to human ingenuity and exploitative behaviour. Even good practice guidelines can be used as inspiration for deceptive patterns, as they can be flipped from something helpful into something harmful.2 Deceptive patterns are often created in an opportunistic and pragmatic manner: if something works, the business doesn’t need to stop and carefully consider the exact reasons why. Think of it like a brawl in a kung fu movie. If a character sees a nearby prop – whether it’s an umbrella, a ladder or a mop – they’ll pick it up and give it a go. If it doesn’t work, they’ll move on quickly to something else, because results matter more than principles. Everyone tries different patterns in different ways.
I hope I’ve now explained why the landscape of deceptive patterns is so messy, and how taxonomies are always going to have some...