3.8 How to combine design thinking and data analytics to spur agility
The job profiles and roles in our companies are changing across the board. There is a multitude of new job profiles today. Until recently, Peter thought he had the coolest job in his company. After all, as the Co-Creation and Innovation Manager, he shaped the innovations of tomorrow. Then, some time ago, he read in the Harvard Business Review that being a data scientist is the “sexiest job in the 21st century.” In the future, data scientists will generate innovations, solve problems, satisfy customers, and get to know more about the customers’ needs through big data analytics. In his blog on digital transformation, the CEO of Peter’s company had also written about a data-driven business and that, nowadays, all business problems are solved with the new technologies.
To benefit from big data analytics, we need a procedural model that combines design thinking with the tools of data scientists. The “hybrid model” (Lewrick and Link) is a suitable way to do so. This model has been developed based on the design thinking components. It promises to boost agility and ultimately result in better solutions. The hybrid approach gives companies the opportunity to position themselves as pioneers and become data-driven enterprises.
The model consists of four components: (1) the hybrid mindset; (2) a tool box filled with the existing design thinking and new big data analytics tools; and other key elements are the collaboration of (3) data scientists with design thinkers as well as a hybrid process (4) that can give orientation to all parties involved. Thus the hybrid model is another possibility for expanding the design thinking mindset and generating better solutions from the combination.
The advantage of the hybrid model: We create a mindset that gives us superior arguments when dealing with skeptics in traditional companies. One frequent point of criticism is that design thinking generates information on the needs only through ethnographic and sociological methods such as observation and surveys. With the hybrid approach, we can eliminate this vulnerability. Expanded by tools for the collection and analysis of big data, the quality of the design thinking process is heightened throughout.
Because the hybrid model is following the design thinking process, we primarily want to point out what is added. As in design thinking, the customer need and a problem statement (pain) to be solved mark the beginning. It can be a more rational or else an emotional problem. In the end, the solution may be a newly defined physical product, a digital solution in the form of a dashboard, or a combined solution that encompasses both elements.
The first phase is understand: we develop in common an understanding of the problem. It is important that data scientists and design thinkers already collaborate here. Some facts can be determined through the analysis of social media data, for instance, which has a broader base than data gathered from traditional user surveys.
The observe & data mining phase is dedicated to the collection of “deep insights” and “deep learnings.” “Deep insights” arise from our traditional observations of customers, users, extreme users, and the like. To obtain “deep learnings,” data must be collected, described, and analyzed, which allows us to identify initial patterns and visualize them. We recommend discussing the insights from both observations together and reviewing the next steps.
In the define phase, we combine the “deep insights” and “deep learnings.” A more exact point of view can be defined this way. The PoV describes the need a specific customer has and on what insights the need is based. The combination of both sides helps to get a better picture of the customer. The stumbling block again here is the definition of the PoV. We already talked about it in Chapter 1.6. The hybrid approach yields more “insights” that confirm the PoV but can also result in even bigger contradictions.
The aim of the ideate phase is to continue to generate as many ideas as possible, which are then summarized and evaluated by us. Several ideas are available at the end of this phase that are used in the next steps.
Then comes the prototype & modeling experiments phase. In this phase, we develop prototypes and carry out experiments with models. Prototypes make ideas palpable and easy to understand. As we know, a prototype can take many different forms; an algorithm, for instance, is also a simple prototype. The insights from the data experiments are best represented with models in the form of visualizations; in data science, this is the best solution to make something tangible.
In a test & proof of value phase, the prototypes are tested together with the potential user in order to learn from the feedback and adapt the solutions to the needs of the customer. This includes models, visualizations, and dashboards from data science, which constitute the basis for the prototype.
In the final phase, realize, we transfer an idea into an innovation! This includes integrating the models in operations. While data solutions usually evolve from data science projects and design thinking develops products or services, in the hybrid process, combined solutions from data science and design thinking can emerge. This can refer to a service-plus business model that presents added value as a result of the aggregation of various data sources; an example would be changes in the behavior of drivers to avoid traffic congestion in combination with an app.
For the successful combination of big data analytics and design thinking, a mindset should prevail that reflects the work in a hybrid model. Because we now have a group of data scientists on board in the projects, it is useful to add the corresponding components to design thinking. A possible mindset can be described as follows:
You need an interdisciplinary team to work with the hybrid model. It is made up of design thinkers, data scientists, and those responsible for implementation.
A facilitator who has the methodological knowledge continues to support the team. The team members can come from a wide variety of areas and contribute their differentiated background knowledge. Depending on the situation, the right specialist from data science can be used. The people responsible for implementation are part of the team.
We recommend having a combined toolbox ready that contains the usual methods of design thinking and the tools from data science. As in design thinking, the critical point is: use the right method at the right point in time. There are many useful methods in design thinking that are easy to apply and quick to learn for everybody. In data science, things are a little more complicated because many tools require expert knowledge. But there is hope that more and more tools are being established that are user-friendly and can be used to perform data analysis without programming skills and expert knowledge. In addition, an increasing number of companies train their employees to acquire these skills. We have had very good experiences with Tableau. This is an easy-to-use tool. It also has a “back” functionality, if something goes wrong in the data experiments.
The hybrid approach compensates for the weaknesses of the unified approaches. Introducing a combined mindset has better chances of success than introducing one after the other sequentially.
In our experience, both top-down and bottom-up work.
With a bottom-up approach, the exchange between employees who deal with the subject of design thinking and those who are into data is promoted. In workshops, the two groups can present their approaches and challenges to each other. It quickly becomes apparent that the two approaches are complementary. The goal is to find a common pilot project in which the collaboration can initially be tested.
In a top-down approach, the advantages and disadvantages of both mindsets are presented to top management with the goal of carrying out an initial pilot project using the method of the hybrid model. After the pilot project is completed, the experienced gathered and the advantages are reported to top management and the stakeholders. In general, the hybrid approach reduces a number of risk factors; for example, it lowers the innovation risk of early experiments. In interdisciplinary teams, not only are new skills brought to the projects but also different ideas, which broaden the perspective. The same applies to a combination of systems thinking and design thinking and to projects that link strategic foresight to design thinking.
The hybrid approach—paradigm shift reduces risks
Paradigm shift Focus on the overall picture (human being + data) New mindset New composition of the teams New hybrid process |
Risk factors that can be reduced Innovation risk/risk entailed in search field for ideas Cultural risk Skills risk Model risk |
Implementation principles Support, top management Part of the transformation toward digitization and/or data-driven enterprise |
Risk factors that can be reduced Implementation risk Strategy fit risk/management risk |
The usefulness of hybrid models became quite clear to us early on. Along with heightened agility, we can generate more insights with the combined approach and mindset, which allows us to increase the number of possible solutions. Top innovators go one step further with their mindset and switch between design thinking, systems thinking, and data analytics across the entire development cycle. The quadruple diamond ensures that the optimum mindset is applied at each point in the cycle. Especially with far-reaching and complex problem statements, the respective design teams, squads, or experimental labs can optimize their work and apply the different skills sequentially or in a mixed form. The respective experts come from the corresponding chapters or guilds and help ensure that the necessary skills are available for each phase. As a facilitator or as the leader of a tribe, this also means having a higher level of methodological expertise at their disposal and having a sense for applying the right methods and tools in each phase.
We benefit from the fact that the three approaches go through similar steps. Thus the quadruple is purposefully built on the “double diamond” of design thinking, which is augmented by data analytics and systems thinking. Depending on the project, the mindsets can be mixed in the respective iterations.
When applying them sequentially, one single approach is executed; in the reflection at the end of the iteration, the further course and the method to be used in the next iteration are determined.
As described in Chapter 3.1, using design thinking and systems thinking in every project is recommended.
In a project that is largely driven by design thinking (example 1), systems thinking should be applied at least once in the end so as to depict and classify all the insights systematically. In a project that is driven by systems thinking and in which the system has already been improved iteratively two or three times, the critical assumptions should be checked in design thinking experiments, thus validating the system (example 2).
At the end of the day, the point is to understand each and every aspect of the problem from all perspectives by working on mixed teams with mixed methods. In the second part of the double diamond, the right solution is then also found with combined approaches. Example 3 shows the combination of design thinking and data analytics.
You can also combine all three approaches. This should only be done by experienced teams together with a facilitator, however. Combining all approaches naturally requires know-how in all of them (see example 4). As in the hybrid model, it is important that mindset, team, and tool sets are combined and not only the process be considered.
As in the hybrid model, it is important that mindset, team, and tool sets are combined and that not only the process be considered.
