Design as a Probe for Innovation
Design is a tool to imagine possible future states.
Working at an innovation lab broadened my understanding of what Design can bring to the table for interdisciplinary innovation.
Innovation v Invention
Design is relevant for innovation. This becomes more clear when we differentiate “Innovation” to (rather technical) work on “Invention”.
If an “Invention” can be understood as a truly (technical) new idea, “Innovation” adds the notion of a “use” of such an idea [wired, 2015]. Success of an innovation cannot be determined purely by technical tests or benchmarking. It needs to be evaluated through context and usage. How do people use this innovative idea? How does it support their tasks and goals? Does it feel right?
Suddenly we start to talk about “use cases”, “users”, their “goals” and “hopes” and “fears”. We need (enough) context and user feedback to validate this more fuzzy part. This is where user-centred design methods, such as rapid prototyping, user testing and early qualitative feedback come into play.
So which roles does Design play in this process?
The Probe (Exploring Uncertainty)
When approaching a new project, so many questions are unclear. What are the use cases? How do people work currently? How can we support this work? Which technology makes sense to solve the problem? First, and foremost, what is the actual problem we are trying to solve?
Bringing many different people to the table brings challenges. Lingo, different terminology and domain expertise. It is hard to describe pain points and ideas verbally. And even harder to imagine other people’s ideas, discuss assumptions and deal with the uncertainty about added value to a customer or user.
We want all these different perspectives … and we want to make it tangible. Design Thinking methods help explore the problem space and alternative solutions by externalising knowledge and ideas.
The Explorer (Problem + Solution Space)
Creative problem solving can only be successful when based on a proper exploration of the problem space.
In workshops and upfront user research we can explore stakeholder maps, personas, customer workflows and written user stories to capture different facets of the context of a problem.
Moving into co-creation with an interdisciplinary group, it might be hard to describe new features and workflows verbally. And even harder to imagine different mind’s ideas. However we want all of that … and we want to make it tangible. We want data journeys, a sketch (which is a fantastic tool by the way) or Desktop Walkthroughs. The idea to help a team align and express ideas through physical artefacts is fairly simply — yet, it can feel new and (slightly) uncomforting.
Even a simple paper prototype can already be used to collect user feedback in a more tangible way than a verbal pitch. This allows exploring and ruling out inappropriate use cases or designs early on. Such rapid prototyping keeps cost to play through an idea low.
The Translator (between Technology + Users)
Wrapping data science innovation into a user-centred design brings up different questions for design and evaluation.
Often it is hard to imagine how automation might impact knowledge work — this is true for stakeholders, users and even system designers alike. An early mock-up — informed by technical capabilities — can help to translate abstract terms such as machine learning, classification and semantic search into more concrete examples.
A user might for example expect a tool to find “more” results when more keywords are added to a search query. Combining such terms with a boolean AND however would rather narrow the search (this is also known as the Linda Problem). When it comes to semantic search, that finds “similar” sentences or expressions, user expectation and system behaviour clash even more.
Another example, while (some) data sets can be clustered rapidly, it is really hard to automatically assign meaningful labels. This might be obvious to anyone working in the field — but quite unexpected to an end-user. When working on such solutions we therefore need to come up with ways to present such clusters — different from labelled groups. Spotify for example shows uses a design pattern where it shows artists within a cluster — rather than a label.
It is necessary to translate such new unexpected behaviour into new objects the user can interact with. However when designing a search that “feels” different to keywords search, we very quickly touch emotional issues like perceived accuracy, reliability and trust into the results.
The Vision
While good ideas are hard work, they also come most often when being relaxed the most. “Design” per definition is an unfinished state. So why not create a playful environment in design workshops to play through, unfinished and imperfect ideas — and iterate until we get to a more appropriate state?
Again, expressing ideas visually early on helps to establish a shared vision. Making an abstract project a visual thing helps to set a goal for a future state.
For example, while it is very easy to put some kind of a “score” or learning system on a wireframe — this might require hell of a lot of data science work. However it enables user testing of two things — first, if we need such a “score” from a user’s perspective, and second what a user would expect to go into such a score. This won’t help the data science team to actually develop the algorithm, but provide information about the context to tune it or represent it accordingly.
…
What are your thoughts on Design in innovation teams? There are certainly many more ideas and roles to discuss …
(All icons have been created by Nook Fulloption — and no, UX is not rocket science, but I loved the icons anyways)
