Managing High Uncertainty with Human-Centered Design and Agile Practices

The time span between two earth-shaking news with global implications has been constantly reducing. Humanity is facing the biggest historical challenges which will ultimately determine the sorts of our species. Black swans are the new normal.

On a micro level, the structures and processes established to reduce uncertainty and support calculated bets are becoming obsolete with the new conditions of extreme uncertainty. New management models that can cope with these new conditions are becoming crucial factors in determining companies’ survival⁶.

Surely, what both micro and macro levels have in common, is the need for quick actions and creative ideas. The following questions arise:

This article aims to provide ideas for project managers or anyone facing a complex challenge on how to deal with high uncertainties. The main target area towards which most considerations are drawn is software and web development and therefore some reflections might not apply to different sectors. Information about what might be important to consider before embracing agile and/or design thinking practices will be included. The described Dual Track Agile model will serve as a tangible example of a solution to deliver user-centered products with speed.

It is asserted that one-third of all software development projects fail due to bad management and organization⁵. Different studies confirm an average of an uncertainty factor of 4 in time estimation for software projects¹. That means that the actual time needed for a project might be four times longer or only a fourth of what was predicted.
Another more recent study based on a total of 50’000 projects over five years showed that approximately 52% of all projects were challenged, while 19% of them completely failed¹³.
A higher rate of failure is noticed with bigger and more complex projects. In the same study previously mentioned, only 6% of “grand” projects were completed with satisfactory results, on time, and on budget¹³. In fact, bigger projects might spell longer timelines and therefore higher uncertainty. Some studies demonstrated that the accuracy of forecasting over one year span declines rapidly until culminating in the realm of randomness⁷.

Human-centered design (HCD) and agile approaches are both focused on the real needs of users and strive for continuous improvement of products/services through iterations. Problems arise when the two approaches need to be integrated. For example, a popular dispute revolves around the collaboration between UX designers and agile teams. Indeed, the detail level used by UX designers to describe user cases does not perfectly align with the format of the agile user stories¹³.
Moreover, the common iteration cycles of agile sprints are too short for a design team to properly conduct the required research to solve problems. Among the possible solutions to these problems, the Dual Track Agile framework (design thinking with agile) has been selected for further analysis to answer the main question being a perfect example of a human-centered design and agile mix.

The Dual Track Agile model is a very good example of human-centered design and agile hacks combined. It includes two tracks of activities running in parallel: A discovery track in which concepts are developed and validated, and a delivery track whereby the solutions are implemented (see figure 2). The advanced agile sprint cadence followed by the delivery track does not impact the designers who can therefore take the required time to identify users’ needs¹³. The framework seems to allow scalability² and reduce the risk of additional costs⁸. Nevertheless, the model does not fit for smaller projects, and it requires close communication between designers and developers.

Design thinking can boost innovation by unleashing “people’s full creative energies, win their commitment, and radically improve processes”¹⁵. Design thinking, a human-centered framework, supports the process of innovation with tools for discovering customers’ needs, generating ideas, and testing solutions. Furthermore, these tools encourage conversation among diverse voices, filter the “portfolio of options” with the best ideas, and enable the involvement of multiple employees (not only designers).

It has already been mentioned that different types of projects require different approaches. Also, the market where an organisation is operating can be stable and more suitable for traditional management styles. Conversely, the same traditional approaches would not sustain in unstable environments. What also plays an important role are the structure and the culture of a company. According to the classification of organizational models proposed by Laloux¹⁴, companies with command structures, pyramid hierarchies, with a long-term focus are not fitting the ideal conditions that fuel innovation.
When it comes to the agile approach, not every business area of a company needs to be nimble to quickly react to changing customer behaviours or develop new products in tight timeframes. Sometimes bureaucratic systems are more effective when consistency and efficiency are required. Moreover, a lot of companies with a traditional structure, are exploring ways to exploit “agility hacks” by creating temporary teams that can bypass bureaucracy in order to create novel innovation⁴.

When selecting a process model, several aspects need to be examined before the actual project starts. The Stacey landscape is a model that makes this decision easier by evaluating the degree of uncertainty of technology, methods, approach, and requirements (see figure 3 below).

When facing challenges in conditions of high uncertainty, HCD and agile methodologies are considered powerful tools. In unstable environments, it may be important to reduce complexity and try to make sense of the faced challenge. One way to reduce complexity is to break down the challenge into smaller modules and experiment solutions with speed. When the challenge calls for innovation, an efficient toolkit is provided by the design thinking process model. Nevertheless, there are situations where human-centered and agile methods are not considered very suitable, and the Stacey Matrix can be used as a compass to choose the best fitting process model depending on the degree of uncertainty.