Design Thinking and Data Science (Marriage Confidential ...?)

Pablo Aguirre Solana
Feb 20, 2023
8 min read

Dedicated to MIT Professor Renee RichardsoN, who has been a source of inspiration and valuable professional advice.

In her bestselling book Marriage Confidential, Pamela Haag draws a pretty sober and realistic view on marriage, trying to wittingly and inquisitively disentangle the reasons and things that lead to what she calls an “ambivalent, semi-happy, low-conflict or melancholic marriage”. She does it capitalizing a central premise: for a marriage to work out one must demystify and be open to more possibilities and options outside the realm of science fiction and pure fantasy, that is, permission to choose alternatives within marriage.

But, what does this has to do with Design Thinking and Data Science, you might wonder. Honestly, I think it does in a very big way. In my professional experience, very recently the relationship between Design Thinking and Data Science among some clients and colleagues is sometimes seen as a romanticized “happily-ever-after” marriage, or sometimes as a “shotgun” marriage. Let me tell you why.

A few weeks ago, in a pitch for a consumer research project a C-level executive of a beverage multinational firm and his managerial team invoked the words “artificial intelligence” and “innovation” probably ten times in the meeting, presenting them as a solve-everything paradigm. The problem with this approach is that innovation and artificial intelligence are two very broad activities that can encompass many things (such as Design Thinking and Data Science, for example), and in this meeting they were mixed indiscriminately as if they were one big collapsible concept. This is what I call the “romantic version” of Design Thinking and Data Science.

Contrariwise, working with a Data Engineer business partner in solving a specific need for a client for which we were trying to develop a machine learning application, the conversation focused solely on the type of containerization (Dockers) we were going to use to split the clients huge SQL data base and then feed it to our API, disregarding conversations such as if the customer needs were properly addressed, who and how was going to use the application, which were the client’s expectations, etc. That is, what I would call the “divorced version” of Design Thinking and Data Science.

Thus, based on these examples and many others, a thought started to emerge in my mind: What type of marriage can exist between Design Thinking and Data Science? And what permissions can we take to choose alternatives within this marriage?

In order to be able to answer these questions, my first step was to ask myself: what do my clients and colleagues understand broadly about these two concepts? And secondly: what kind of marriage they expect between them based on what they know? Since I don’t know the answer to these questions, I think it is more useful to provide a definition, in such a way to have a common knowledge on what we mean by what we talk. Because, as I mentioned before, the business world is full of “buzzwords” that enhance a narrative, but provide little content and tangibility for the solution of problems.

The central idea behind these definitions is that what my clients and colleagues do know is that with leverage of data and design we can make better products, design better and innovative customer experiences and add value to them. As the saying goes: “Data is the new oil”, but if left unrefined, and undefined, is effectively worthless, I guess...

So, before we delve deeper into the types of marriage that can exist between these two concepts and the tensions that can arise, let’s refine ourselves with some definitions from Randal Elliot, and Tim Brown, that will help clarify the proceeding discussion and ideas ahead.

·Design thinking is a human-centered approach to innovation that draws from the designer's toolkit to integrate the needs of people, the possibilities of technology, and the requirements for business success.

·Data Science applies the scientific method to data as a means to infer valuable evidence that can be used to form new ideas and can encompass Artificial Intelligence, Machine Learning and Big Data.

· Artificial Intelligence (AI) refers to machines that can think in some ways like humans.

· Machine learning is commonly associated with AI and is a field of computer science that gives computers the ability to learn without being explicitly programmed.

·Big data refers to extremely large data sets that may be analyzed computationally to reveal patterns, trends, and associations, especially relating to human behavior and interactions.

Now, with these definitions at hand, my second step was to ask myself: what tensions and misunderstandings can Design Thinking and Data Science solve together?

As I dug deeper into the challenges that I have faced in recent projects, and as I put more thought to the above question, I found at least two very common big tensions, assumptions and misunderstandings that I have systematically seen in my professional experience that, I think, can be solved combining the best of these two worlds (Design Thinking and Data Science), as a way to open the alternatives within a marriage.

The first assumption, misunderstanding and tension is that Design Thinking (generically and indiscriminately called Innovation) is a separate world from Data Science. Typically, in large companies the Innovation department (if it exists) is separated from other departments such as IT, Revenue Management, Marketing, etc. And I refer not to a physical separation, but in terms of agendas and business objectives. For example, in a large multinational beverage corporation, if the Innovation department has to develop a new flavor, the communication and relationships that will develop with other departments such as IT, Revenue Management and Marketing will be instrumental and scant; that is, the probabilities that the Innovation department takes the development of a new flavor to the realm of other departments is very low, at least in my experience, thus generating a siloed vicious circle about information, insights, creativity and results, in which each department has its own data, its own research, its own information, and hence, its own way of doing things.

At the bottom of this problem lies a non-revealed assumption of the apparent incompatibility between the Innovation department’s processes and those of other departments. What I mean by this is that, in this particular example, to develop a new flavor for a beverage the Innovation department will probably conduct some qualitative and quantitative consumer research and do some lab formulations in accordance, and finally test it with some sort of randomized trial of consumer product testing. At this point, all the information will be generated by the Innovation department and will stay in it. Perhaps the Innovation department will reach the Marketing department to ask for some consumer trends and sales information to contextualize its results, but they will remain siloed in the department. And what about the other departments, what about other data sources (i.e. real time consumer data), what about other information and creativity sources within the company and outside the innovation team, that can be crucial for developing a new flavor for a beverage?

I am not advocating for a massive and general integration of departments, nor for a “one-size-fits-all” kind of solution. On the contrary, I am thinking that those apparent different processes among departments, and those siloed- informational vicious circles can be broken with the integration of Design Thinking and Data Science.

Therefore:

First, companies can benefit from a framework that connect needs with methods. In the problem described above, the company new perfectly well what to do to develop a new flavor, but carried out the process separately, loosing perhaps some valuable touchpoints and sources of information, creativity and insights from other parts of the company.

Design thinking provides a framework that can connect the company needs, and data science can simultaneously provide the methods to satisfy them. Both share the same epistemological* structure, because, in the end, both are problem-solving processes. Both can be human centered, scientifically based, iterative, analytical and synthetical. From my perspective, the stages of Design Thinking and Data Science in many projects can run in parallel. The challenge here is to have the managerial sensibility to form multidisciplinary teams that can harness and potentiate this framework productively and proactively.

2) The second one is an assumption, an understanding and a tension regarding the fact that Design Thinking and Data Science are a source of answers and expert knowledge. Under this assumption, and somehow expectation, both are forced to work in an unsatisfied and frustrated marriage. Watch out: I’m not stating that there is not domain knowledge and expertise in them. What I am saying is that the sweet spot between data science and design thinking does not lie in the expertise they produce, obviously it is a by-product of them, but rather in the questions they generate and in the insightful creativity stimulated by them.

* Relating to the theory of knowledge, especially with regard to its methods, validity, and scope.

As my dear and most admired professor Hal Gregersen has stated and written about extensively: “It is about the questions we ask, not the answers”, and the mixture of the processes that Design Thinking and Data Science follow can harness that in a most productive way.

I have to confess that in my professional life as a consultant I have always felt this latent undisclosed and silent pressure of knowing everything, or at least to have a really intelligent or smart answer for my clients. Because, in some way, I guess that there is this unspoken expectation, as in marriage (sometimes, ha-ha), that you have and know all the answers because that’s the reason why they hire you, right?

Nothing falser and potentially toxic than that. Focusing only on answering specific questions can limit our perspective, scope and possibilities of widening and stretching a myriad of outcomes for a project or a business problem. Inquiry leads to insight, as Hal Gregersen tells us, and I believe that the mindset and framework that Design Thinking and Data Science provides can entice and provoke more insights than just answering a specific question.

This is why Design Thinking and Data Science articulate a process in which the outcome is part of a series of asking questions and inquiring in different stages and processes. It is the sum of the parts, rather than the part itself, what reduces the uncertainty and produces concatenated answers through a multiplicity of insights discovered at each stage.

Specifically, a typical Design Thinking project would follow, generally, five stages (Discovery, Immersion, Ideation, Experimentation and Evolution). Similarly, a typical Data Science project would follow these same five stages (Identify, Exploratory Data Analysis, Hypothesize, Model and Validate). In these five stages, a lot of things happen, and one for sure is that each stage provides some new information, and thus, some new questions. Hence, it is in the embedding of these findings along with the questions where we can find the intersection (sweet spot) between what is desirable, viable and feasible for a business problem.

Therefore, it is not about a single business question that needs to be answered in isolation, as in a typical consulting-client context or in a typical business scenario (developing a new flavor for a beverage). It is about a series of questions that at each stage provide pieces of information that can be concatenated, and thus provide a richer and a wider story, and a frame of answers to that same initial and isolated business question. And along this path, Design Thinking and Data Science can complement each other naturally, rather than compete with each other in segregation.

Understanding these two tensions that I have presented to you above, we might, perhaps, be in a position to conclude and asses what kind of marriage can exist between Design Thinking and Data Science. For this purpose, let me offer you a visual explanation of how Design Thinking and Data Science share the same nature and framework, even if operatively and technically they differ.

I believe that, combined, these tools can open new ways to solve complex problems and innovate with data. Consider the example that I presented before about developing a new flavor for a beverage, and imagine that other departments apart from Innovation, such as IT, Finance, Marketing etc., can ask some insightful questions, discover interesting relations and create a myriad of solutions with different sources of data and multidisciplinary teams in the different stages and different processes, like the ones that I show you below.