Purgatory is a state of existence in which a person or object remains suspended between two distinct states of being.

Pilots are often considered the proving ground for technologies. The logical premise is that a technology can be “piloted,” the same way a television series is piloted, to test its applicability to the organization. Technology companies and consulting firms have invested heavily within this premise over the years to demonstrate the value of their solutions to potential customers.

A staggering number of these pilots, however, even when they meet the required benchmarks and advance beyond the pilot stage, never make it into production. While the specific data for this is difficult to track down (this is not a metric organizations tend to share publicly), it is commonly understood and referred to across industries, and is directionally represented by the statistic that only 13% of data science projects make it into production.¹

Within the context of a mechanistic worldview, a pilot is a test of the viability of a technology or solution in the same way that a machine would be piloted to demonstrate its value.

But it is more than that.

Pilots within Social Systems

A technology or solution pilot is a social engagement between two or more parties, in which one party is endeavoring to earn the right to do business with another party. Unlike a machine test, which relies solely on the performance of the machine, a technology or solution pilot relies on the interaction between those demonstrating and those testing the viability of the technology.

A data science pilot, for example, begins with the process of designing the experiment, which involves a number of social interactions internal and external to the organization while choosing the use case, measuring the existing performance benchmark against which the solution will be tested, building the team, and proposing a shared timeline. This is a social engagement between organizations in which the social dynamic, regardless of individual intentions, is fraught with undue tension. Fundamentally, an external organization is asking or being asked to prove that they are able to outperform the existing work with the same data and/or systems, but without the internal network of relationships or domain expertise. Furthermore, depending on the organization and the construct of the pilot, friction internal to the organization can present itself, such as if a technology is being piloted without the knowledge of the information technology (IT) organization or if the technology goes against the prevailing technology strategy across the organization.

Once the experiment has been designed, its success relies on the ability of the technology or consulting organization's data science team to access the necessary data, and just as importantly, to understand the data, the broader context in which the data is generated, and what it represents. This requires interaction with domain experts, business leaders, and managers within the potential customer's organization.

This challenge is compounded in situations layered with additional social dynamics, such as Data Science Taylorism, where data scientists and experts do not believe they need to speak with domain experts, and that they only need access to the data to know all that they will need to know to complete the pilot and demonstrate value.

This behavior can also stem as a defensive mechanism to the pilot construct. If a given data science manager has faced undue criticism due not to their work, but to the natural direction of the social construct of a pilot, defensive maneuvers that eliminate social interaction within the prospective customer organization can begin to take shape. This harms the likelihood of pilot success and of production deployment in the long term.

Riskless Experimentation

If an organization does not need a new capability in order to get to a desired future, a pilot becomes an interesting‐at‐best, riskless experiment conducted in a controlled environment with limited scope and resources.

This strategy hedges the organizational leader approving the experiment by removing the risk if the pilot does not achieve its goals and, ironically, also almost guarantees that the pilot will not achieve its goals.

Technology projects require domain expertise in order to create value. Domain experts typically do not have extra bandwidth to dedicate the necessary time to share expertise with technologists building a pilot that is not guaranteed to be successful and ultimately make an impact on the broader organization. This can lead to a self‐defeating cycle.

Efforts to lower risk also have a natural tendency to direct pilot initiatives toward use cases that are not essential to the present or the future of the organization, which places the pilot at the bottom of the list of organizational priorities. When pilots inevitably face the need for greater support upward or across the organization, leaders from across the company are naturally and logically incented to deprioritize dedicating their resources or social capital to support anything deemed “experimental.”

Piloting the Path to Promotion

Some organizational leaders have seen pilots as a means of securing promotion within their organization. This involves leveraging investment from one or more technology and/or consulting firms to demonstrate the value of their platform or solution across a set of challenges the organization is facing. When one or more of the pilots meets its targeted metrics, the leader or manager can present the findings to their leaders and the broader organization with a planned path forward, and claim the vision not only of designing the pilot but also leveraging external investment.

This design pattern is common enough that many external consultants have fostered symbiotic relationships with up‐and‐coming leaders, to whom they are happy to give all credit as long as there is a steady flow of billable hours.

When this works, both the organizational leader and the supporting external organization are happy and feel proud of the result.

There are several issues with this approach, however:

1. The nature with which a pilot with this goal is set up means that the only path to escalation in order to overcome a challenge is disclosing the pilot and diffusing its element of surprise and awe, which incents organizational leaders to try ineffective methods of overcoming challenges.
2. When peers learn about a pilot that has been devised in this manner and with this goal, they, at best, offer resources or help despite the incentive not to, and at worst, are incented to intentionally derail the pilot.
3. Even if a pilot is successful in technical terms, when the organizational leaders attempt to raise broader awareness and secure investment to move the pilot forward, it can be shut down by leadership, either directly within or adjacent to the organization in which the pilot took place.
4. A pilot of this nature, by design, cannot get buy‐in and collaboration from broader organizational stakeholders, which is a fundamental input to any successful organizational initiative.
5. The pilot could also aim in a direction that overlaps or contradicts other investments being made by the organization, and could therefore be shut down regardless of its ability to achieve its targeted objectives.

Human‐Centered Transformation

Because organizations are social systems, the path to overcoming pilot purgatory begins with a decisive step away from the mechanistic concept of a pilot, which focuses on choosing a problem to solve or a use case, deciding what metrics can be measured to determine if the problem was adequately solved, paired with a data‐driven analysis of a projected return on investment, a timeline, and so on.

An approach better suited to a social system—a human‐centered approach—begins with the forming or reforming of the center of the social system around which the future of the organization can begin to transform through a combination of both designed and organic butterfly effects.

The Executive Committee for Human‐Centered Transformation is that center. As demonstrated in Figure 19.1, the committee is designed to anchor on leadership voices from across the business, industry, and technology groups, led by a C‐level executive sponsor, with the inclusion of external voices as best fits the ecosystem of a given organization and the future it wishes to design, represented in the figure as a managing partner of an advisory firm and an academic leader, which could be replaced by other voices. The same holds true for the rest of the diagram, which is intended as a directional guide to be translated into the specific context of an organization.

By creating a miniature model of the organization at the executive level with joint accountability for shared objectives, leaders can begin solving for the future of their industry and organization, and any broad‐sweeping organizational changes that would be required in order to achieve that future can begin within this context, and be subsequently carried out throughout the organization until those changes accrue to a meaningful transformation across the overarching organization.

Transitioning from being a data‐driven to a reason‐driven organization (from Chapter 15), for example, would begin with a subset of organizational leaders such as this determining the future it wants to bring about for the broader organization, forming theories and hypotheses to be assigned across organizations to then be supported by the creation, testing, and validation of additional hypotheses, and so on, until the culture of the organization transforms from being data‐driven to reason‐driven.

This construct can and should continue to transform and evolve as the nucleus around which the organization pivots in leading and reacting to market forces and dynamics, which will be reflected by the people and organizations represented within the executive committee.

This committee, while most effective when chaired by leaders with the broadest purview across the organization, can begin at any level within the organization to begin to drive meaningful change.