The Life and Death of Data Products
This is an approximate transcript of a Pecha-Kucha style talk I gave at a DataBeers Madrid event in the Google Madrid Campus on June 27, 2016. A Pecha-Kucha presentation recommends 20 slides for 20 seconds each, forcing a concise, fast paced, and entertaining 7 minutes talk.
My talk was about data products. That is products characterized by an experience that evolves according to user behaviors and constantly updating algorithms fed by streams of data (e.g. Fitbit, Spotify, Fintonic, or Helios, your hypothetical first self-driving car). I argued these living products — that inevitably also have a tendency to die — need different design considerations at the crossroad of data science and user experience. I showed how Design Fiction could be used in data product teams to unveil, discuss, and evaluate opportunities, challenges and other unknown unknowns before a project gets started or even funded.
Intrigued? This is my 7min Pecha-Kucha style article on “The Life and Death of Data Products”.
Let’s begin.
Hello Madrid! My name is Fabien. I am a researcher, engineer, creator and co-founder of the Near Future Laboratory.
I am fascinated by the co-evolution of humans and technologies, about how technology is changing and how it is changing people. Practically, this means I constantly observe how humans appropriate technologies and I love to question, design and create the future of this relation.
I currently work at BBVA Data & Analytics where — among other things — I collaborate with data scientists to transform their algorithms into experiences that are future forward. My responsibility is to bring a holistic experience design to the organization and make it an essential part of the development of data engines applied for instance in customer intelligence, customer advisory, risk, fraud or new data products. Some of these engines are already massively used and others belong to the speculation phase.
What attracted me to this job was the field of opportunities that emerged from the banking industry that has suffered from years of marketing myopia and that now wants to recover clear sight on people’s expectations on and interactions with their bank. The competition is now global, fast-paced, technological and named Google, Amazon, Apple or Alibaba. The growing number of people familiar with these brands expect a personal and digital experience with data as facilitator. It is a fascinating moment — for somebody with a career in tech, design and academia — to be part of that evolution inside a 159 years old institutions known for its culture of innovation.
In that landscape, the transformation of an idea into a product based on data must work the tensions between three dimensions: profitable, possible and desirable. Acting as bridge between teams of data scientists and designers, I am particularly involved in helping strike the balance between the technically feasible and the desirable for society, a company and customers.
The conceptualization phase requires that visions live not just as flat perfect things for board room PowerPoint, but as tangible things co-existing with all of the dynamic tensions and forces in the world. In consequence, I get my hands dirty programming and creating prototypes. A prototype is useful to experience the desirable and gauge the possible.
However, with products powered by machine learning and artificial intelligence (i.e. data products), the experience is no more linear or based on static business/design rules. The experience evolves according to human behaviors with constantly updating models fed by streams of data . Each single product becomes almost like a living, breathing thing. Or as people at Google would say: It’s a different kind of engineering.
I would argue that it is also a different kind of design. Beyond considering the first contact and the onboarding experience, a data product requires considerations for the user experience after 1 hours, 1 day, 1 year, etc.
To stretch this idea to its extreme, I inspire from the “philosopher of speed” Paul Virillo who argues that the invention of the ship was also the invention of the shipwreck.
Translated to the world of data, the introduction of a new service, products or algorithms requires a responsible design that considers moments when things start to disappoint, embarrass, annoy or stop working or stop being useful.
However the practice of designing the offboarding experience is not common. For instance, allegedly a third of the Fitbit users stop wearing the device within 6 months. What happens to these millions of abandoned connected objects. What happens to the data they produced? Where are the data product shipwrecks?
My practice intrinsically aims at considering the evolution of data products, their potential problems and opportunities, including their death or any associated repair culture and hacker culture. At the Near Future Laboratory we apply a creative approach coined Design Fiction that unveils and make tangible the different moments in the life a products and services.
For instance, consider a self-driving car. It is the quintessence of a data product that will come near you sometime in the near future. It is the moon shot of many data scientists in the world. If you are one of these data scientist, how can you get today a detailed understanding of the self-driving car of tomorrow? A first obvious step is to create an image that communicates that future.
This is compelling, but this is not enough!
With this image it is impossible to understand what the product means to its owner. What are the implications of purchasing this type of vehicle? What can you do with it? What aren’t you allowed to do? How does a human interact with this self-driving car the first time, and then routinely after a month, one year or more?
Design Fiction is an approach to engage with these essential questions. It is a way bring the future to present. We traditionally work backwards. For instance, we start writing the press release of the product. We create the ads of any potential related service.
We write headlines and any type of curious stories that might happen with the product. For instance, what happens if a lost and drunk Justin Bieber is found in his self-driving looping around Hollywood Hills?
And we continue working backwards completing a list of Frequently Asked Questions. What happens if my car gets lost? Is there an emergency button? What is legal or illegal? How do I force the update of the car’s operating system?
With answers to this questions, we can start sketching use cases. What are the main features of the car? What do you do if you forget a bag of groceries after sending it into Uber mode? Will there be a geo-fencing mechanisms to control where the car goes — and how fast it goes — when you give the “keys” to your teenage son to take to football practice. How does the car pickup groceries — and how do you upload the list — when you send it on errands?
To finally collect all the material and communicate it in the vehicle’s Quick Start Guide that represents the features, attributes, characteristics and behaviors of the self-driving car and its requisite “ecosystems”. With this document, the vision of the future feels real and its implications are more detailed.
This type of Design Fiction bring the future to the present to bend visions and trajectories of projects. The user manual acts as a totem for discussion on what is necessary and what is not desirable.
This type of Design fiction allows users to evaluate the concept. They can feel, touch and get a practical understanding of the data product and its experience.
This type of Design Fiction helps convince the world, colleagues, executives and investors. The feedback of people with different perspective allows to anticipate opportunities and challenges.
To wrap up, at the crossroad of data-science and design are emerging living products with an experience that evolves according to human behaviors and constantly updating models fed by streams of data. Design Fiction is one way to approach the design of data products anticipating their evolution, the frustrations they produce, their potential death and why not their after lives.
Thank you!
Thanks to the organizers of DataBeers for giving me the opportunity to present my practice to an open and curious audience.
And thanks to the wonderful attendees of DataBeers for their interest and beer-fueled post-talk discussions.
Particularly I was asked to provide further examples of ‘working backward’. It is an approach that has been applied for more than 10 years at Amazon and other companies. Check out Verner Vogels original article. Design Fiction is a broader practice. For more on that check out the Near Future Laboratory Design Fiction channel, get details on Our Approach of Design Fiction, watch our Design Fiction Evening, go through our projects and take home the near future from our shop.
We also discussed what is a product in the digital world (I like to use Roman Pichler’s description) and what is a data product is and is not. Good friend and Senior Data Scientist at Skyscanner Neal Lathia provides as thoughtful categorization in What do we mean when we talk about data-driven products?
