Setting Goals with Uncertain Variables

Unfortunately, goals rarely take into account the environment they’re supposed to elucidate. This is despite goal-setting in conditions that reflect VUCA, a phrase the U.S military invented near the end of the Cold War: volatility, uncertainty, complexity, and ambiguity. VUCA was meant to articulate the circumstances of war, but it also portrays the situation that many early-stage startups find themselves in.

During my time at ConsenSys Labs, I facilitated goal-setting workshops for several portfolio companies, and tracked the investor updates of more than fifty. Goals that had not been thoroughly examined would often be disconnected from actual progress, or worse, orienting the team in the wrong direction. It did not help that startup literature is woefully lacking in this area. Frameworks like Objectives & Key Results and Scrum provide the scaffolding to organize goals and metrics, but not the foundation they should be built upon.

Although the easiest conversation to leap into was what quantity to choose for their metrics, or how to coordinate their goals, it was always more meaningful to talk purpose. Teams found it most valuable when I dug into why they chose the goals they did. This wasn’t the spreadsheet or graph analysis some were expecting—I even had one client describe the workshop as “therapy”. Goal therapy isn’t a bad way to describe my work. There is certainly treatment and intervention.

An element I would sense for would be the type of learning environment they were operating in. This was connected to many factors, such as the experience of the team, the state of their market, and current traction. However, these factors all fit into the mental model of kind and wicked environments1— a way to conceptualize the usefulness of feedback and how it should factor into decision-making. By applying this framework to goal-setting, we can contain the effects of VUCA.

No Rest for the Wicked

Kind and wicked learning environments describe the level by which inputs to a system, map to outputs.

In a kind environment, it is easy to learn. There is plenty of information and feedback loops for each action. This ‘kindness’ allows for accurate inference. Golf and chess are examples of kind environments. The accuracy of a golf swing can be determined immediately. A chess move shows its value as the game progresses. In both situations, the player can take full ownership for their outcomes.

This luxury does not exist in wicked environments. In these, feedback is unreliable, delayed, or completely absent. Consider the game of poker: players can make the ‘right’ call—one that is statistically favorable—and still lose. The emergency room is another example. In any given situation, the details and medical history may be absent, and outcomes may occur long after the operation. Since there is no pattern matching for actions and outcomes, inference is useless. Decisions here are bets made at the mercy of the environment.

In a kind environment, you are determining how to best answer questions; in a wicked environment, you are determining the best questions to ask.

To better understand how to track progress, let’s go back to the strategic play of poker. Since poker players are not directly responsible for their outcomes, they focus on their process instead. In a long time horizon, a series of ‘right’ calls bring the probability of success onto your side. We can leverage this practice to navigate wicked environments.

When an environment is wicked, you set goals based on process. When an environment is kind, you set goals based on outcomes. Thinking of this as a dichotomy is not useful. Instead, goal-setting requires a temporal dimension: environments begin wicked and transform to kind.

In a startup context, your learning profile will constantly be developing. It changes when you hire a market expert, when you acquire more customers, when you begin user interviews. Isolated points of feedback give way to larger, triangulated insights. In turn, this earns you the ability to make goals that reflect outcomes.

It’s tempting to jump immediately to the last step and measure exactly those outcomes. After all, results are what business are built on, not process. However, an adherence to process—or in the day-to-day verbiage, tasks—provides the initial standardization necessary to pursue results.

Let’s show this with the example of a startup that is beginning to engage in outbound sales. The Sales Development Representative, or SDR, could have their goal expressed as a task or result. As a task, to make fifty cold calls per day. As a result, to generate three new leads. Framed as a task, it does not guarantee the SDR will obtain the leads the business needs. However, there’s also no telling how many cold calls they’d have to make in order to meet the designated result.

For the sake of consideration, let’s bring up some general variables that are common. Since this is the start of the startup’s outreach, the SDR has an untested script and sales pitch. Furthermore, they haven’t honed in on their ideal customer profile yet—there’s no sense of what types of recipients may be amenable to their calls. We can assume there’s going to be a lot of time spent experimenting and figuring out what works. This is a wicked environment.

Of course, there could be initial factors that contribute to kindness. These will be internal advantages. The SDR could be skilled at cold calls or have expert knowledge of their target customers. This is a hypothetical scenario, an example of information you can use to modulate your goals—although it’s always better to err on the side of process, to begin.

Beginning with process supports a longer time horizon. When the goal is three new leads at all costs, it opens the possibility of a widely inefficient process. Imagine hundreds of cold calls being made, when many could be avoided with a tighter sales pitch. I’ve heard opponents of this type of goal-setting claim that strictly measuring results will, in some sort of Darwinian fashion, force individuals to learn and cull the ones that don’t. The SDR will get their pitch right after hundreds of calls or not be the one for the job.

Notice the difference here: situating around process is a sign of respect for the larger environment. It acknowledges the variables at play and the role of learning. Setting goals around results is a Randian statement that the individual will overcome the environment; it is their sole responsibility. The confidence in the latter makes sense, but only once inference can be established. By treating the environment as wicked, we are kind to the individual. We can always begin with this principle.

Goal-setting is no easy task. It requires the ability to step in—introspection—and to step out, or get meta. Although the tempo of progress is enticing, it’s worthwhile to begin with process and move toward outcomes. The concept of learning environments can help you goal-set in situations that reflect VUCA.

  1. concept researched by Robin Hogarth.