Data Science for Design
by Stuart George and Matt Millington from Method
A pragmatist’s approach to using data science and AI as a tool, not a replacement for design.
The topic of Design and AI makes for a confusing read. We have seen everything from outlandish and uninformed opinions claiming that AI is here to take over the job of the designer to the more sensible (and practical) discussions exploring how to deal with bias and verify how accurate results are. All too often we see businesses jumping to solutions asking “how can I shoehorn AI into my project?” in an attempt to appear current without considering if there is a genuine human problem that requires AI to solve.
“If we’re going to really capitalise on Big Data, we need to get to human insight at machine scale.”
In this series, we take a more pragmatic approach to the discussion. We hope to give you a glimpse behind the scenes of how we (Method’s data design team) are embracing the opportunities offered by the proliferation of data both within and outside the organisations we work with. We will help you understand how bringing design and data science together can offer significant new advantages to anyone designing new products and services in the 21st century.
In this first post, we will set the stage by briefly talk about design, data science and data design, and what types of questions data design can help your organisation answer.
So, what do we mean by ‘design’?
To keep things simple, let’s define design as -
“… the intentional creation and execution of a plan, a specification for the construction of an object or system… the implementation of an activity or process based on or adapted to one’s needs.”
At Method we practice user-centric design which introduces the additional step of first gaining empathy for the end-user (or at least a deep understanding of their implicit and explicit needs) and designing specifically with them in mind to mitigate risk of failure and ensure the product, brand, service or experience is meaningful and successful.
The design process can broadly be broken down into two spaces; the first being the problem space where the goal is to gain an understanding of the problem and gain deep empathy for the users (along with understanding the business, brand and their capabilities). The output from this space are insights and a clear problem definition. The solution space uses these insights and problem definition to explore, select and refine the solution.
For a more in-depth definition of design, we encourage you to read Method founder, Patrick Newbery’s post ‘Thinking about Design Thinking’.
Now, defining Data Science
We will define data science as -
“… a multidisciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from data in various forms, both structured and unstructured, and largely synonymous to data mining and big data” (source)
Here we use an expanded and more relaxed definition to include using data for generative exploration, simulation and prediction. We define it as the practice of using data as a design material either using it to discover insights or to mould the shape of an output.
To be able to manipulate data, a data scientist typically lies at the intersection of multiple disciplines, as shown below.
Like strategic designers, a data scientist’s goal is less focused on execution and more on exploration where emphasis and workflow need to be iterative and experiential (and, often, messy) and ensuring they learn as they go, rather than jumping to conclusions and introducing personal bias beforehand.
The opportunity in Data Design
A few years ago Method saw an opportunity to apply creative thinking to data to unlock its potential.
Bringing a designer into the process brings the end-user into the discussion around the application of data. Working with the skillset of a data scientist enables the designer to identify potential hypotheses to explore within the data which can have real value for existing and potential customers instead of small incremental product and service improvements through traditional applications of analytics.
Method has already brought designers and data scientists together to explore how the practice and tools of data science can assist to augment, accelerate and influence design.
To date, we have used data design to help clients across diverse sectors such as consumer electronics, energy and gaming to find new answers and evidence to make strategic decisions such as:
- Understanding the behaviour and drivers of current and new target audiences to help them build more relevant products and services.
- Understanding what features their target audience might want before developing products and services.
- Understanding the needs of their customers through the service they currently use.
- Spotting emerging trends in their sector or adjacent sectors to inform and accelerate product innovation.
- Identifying pain points and opportunities in their current customer journey to increase customer engagement.
Innovation in design, just as in data science, happens through a universal mindset where intuition and deduction, ambiguity and simplification, all coexist. The two disciplines may seem worlds apart but in actual fact, they are both focussed on the same goal: finding answers in complexity to inform how a business connects with its customers in the most meaningful way possible.
In further posts, we will share some of this work, talk to other pioneers in the exploration of data and design, and discuss how we are beginning to use data to further augment and accelerate the design process.
If you enjoyed this read, please also check out Using data design to understand your customers (part 1) and follow Method for more!
Illustrations by Luke Thompson and Ana Soto (Method). Contributions by Chantal Schonbachler (Method).
