A Small Framework for Decision-Making

As a CEO, the primary work product is decisions, and for any leader more generally, decisions remain at least one of the primary outputs.

The Time Frame

When I join a group of people to discuss a decision, the first question to ask is when the decision needs to be made by. We all know a mediocre plan violently executed now is better than the perfect plan executed just too late. Clarity about time helps avoid unnecessary pressure to make a decision too early, as well as paralysis analysis (often motivated by fear).

Part of figuring out the timeframe is asking what might happen if we don’t have a decision made by a certain time. Will we miss a window of opportunity? What are the costs and benefits of taking more or less time, and when do we reach declining marginal benefits of more information?

Critically, the question of timing comes before I even deeply understand what the decision to be made is, or what the topic is. If I don’t have time to understand the problem, then my role is to assist the subject matter experts in reaching a consensus and plan, and to empower them to execute.

Understanding

After figuring out the timeframe, the next step is to define and understand the problem. In some cases, this can take many meetings. Understanding the problem is critical to identifying stakeholders who should be involved in the discussion or in the execution, and to avoiding scope creep.

The Usual Suspects

Next we can start outlining courses of action and performing cost-benefit analyses. This is where most people mistakenly start, and it’s the best-understood step in general, so I won’t say much about it.

One helpful trick is to classify courses into reversible and irreversible, or more and less reversible courses. Easily reversible choices should be entrusted to more junior people (“make a call and let people know”), and should occupy less decision-making time.

It’s also helpful to consider what choices now will show us what decisions to make later; sometimes we can wait to solve part of the problem later, with a more mature environment, if we take a generative course now.

Deciding Who Decides

For many decisions, at this point (or sometimes right after figuring out the timeframe for the decision), we decide who should decide and how. This is another under-discussed step, and most people have not considered the range of available options, although getting this right can dramatically reduce friction.

In social life, we often operate by consensus, which is another way of saying by universal veto: anyone who hates the restaurant choice gets to veto it. In a startup, at times we operate at the other end of the spectrum, with a unilateral dictatorship.

In between those extremes, many government agencies must seek public input, by announcing a deadline and process for giving input, but then take decisions by a small group of stakeholders. And in direct democracies, a plurality or majority or supermajority take a decision.

Different cultures tend to moralize these choices, but any of these gradations may be appropriate for a particular decision. Hierarchical structures lead to clear, fast decisions, and horizontal structures capture more information and increase buy-in.

In general it’s simply helpful to remember that there are options beside the Scylla of rigid hierarchy and the Charybdis of the tyranny of structurelessness. I will note that the “agency model” in which a small group retains control (no voting from outside the group, no veto) but must seek input more broadly does seem underused in most companies.

Documentation

Finally, in a writing-based company culture, we document everything about this decision, consider where to store it, and whom to notify about the documentation. This can range from a Slack message to an email to an essay carefully archived. Usually the decision-makers are responsible for this step.

TLA: Acronym Needed

That’s it: time frames, understanding the problem, the usual probabilistic decision tree and cost-benefit analyses, choosing the deciders, and documentation. I haven’t yet figured out a helpful acronym for these five steps.

A framework like this one sometimes will be honored more in the breach than in the observance, but even then its assumptions help guide the discussion.