Decisions: The Units of Productivity of Cognitive Work

Widget –a generic “thing” representing something a company makes and sells.

For decades, business and economics textbooks have used the term ‘widget’ as a unit of productivity, a generic term for the tangible output of industrial work, regardless of industry. Originating in early 20th century factory models and refined through Frederick Taylor’s scientific management, the widget represented productivity you could count: how many units a worker produced per hour and how much labor it took to make one more. It was simple, measurable, and concrete, the perfect KPI for a manufacturing economy (Taylor 1911).

Cognitive work does not produce widgets. Its raw material is thought, not steel or plastic, and its true output cannot be stacked, stored, or shipped; however, for decades, we’ve still tried to measure cognitive productivity with industrial-age metrics (e.g. hours worked, documents produced, meetings attended).

These metrics don’t measure the productivity of cognitive work because they don’t reflect what the output of cognitive work actually is.

If the output of cognitive work is not widgets, what is it?

It’s decisions.

Decisions are the units of productivity of cognitive work.

In the world of cognitive work, of thinking work, there is still associated task-based work. This often looks like creating reports and dashboards, scheduling and attending meetings, building prototypes and giving presentations. While useful and often needed, these “deliverables” are artefacts, or intermediate steps in producing the real output of cognitive work.

The output of cognitive work is not the things these task-based work elements deliver. The true output of cognitive work is the steady conversion of uncertainty into choices, and the ultimate selection of one of those choices, a decision made, that enables the organization to move forward.

In cognitive work, decisions are what is made. Decisions are what is produced. Progress happens when someone makes a decision.

If a widget is a generic term for a unit of productivity of manual work, a decidget could be a term for a generic unit of productive of cognitive work.

Example

An analyst at Company ABC produced a report, following a request from their manager, to show the fourth quarter sales numbers for the last 5 years. The report is a deliverable—an artefact and intermediate step—not the true output of cognitive work.

The manager asked for this report to take to the steering committee to inform next steps on sales targets for the coming year. The report (an artefact) becomes an input into cognitive work.

If the manager brings this report to the steering committee and, through meeting (another artefact) and discussion with the committee, arrives at a decision to increase fourth-quarter sales targets by 10% this year, a decidget, (i.e. a unit of cognitive work) has been produced.

Not a new idea

The term ‘decidget’ is new (and a little tongue-in-cheek); however, relating decision making to productivity is not a novel idea. It connects to the idea of attention scarcity in organizations where it’s argued that the more efficiently a group directs attention toward making and recording clear choices, the more it accomplishes. (Simon, 1971).

Seeing decisions as units of productivity also explains why cycle time matters. Long delays between “we need to choose” and “we chose” raise the cost of delay: opportunities age, dependencies pile up, and rework grows. Queueing theory gives a simple anchor: when a decision backlog grows, lead time grows with it (Little’s Law), so throughput falls unless you shorten queues or decide faster. In turbulent settings, research on executive teams shows that speed and quality can coexist when teams use real-time information, consider multiple options, and integrate choices tightly with action. (Little, 1961; Eisenhardt, 1989).

While neither Little, Simon nor Eisenhardt state decisions are the units of productivity, their relation to the measurements of productivity shows this to be true. The evidence has been there all along, we just haven’t taken a step back to consider why these things are the case.

Not an either/or

No job is purely cognitive work or task-based work. While thinking and task-performing are always part of the work equation, the consequence of computing, automation and generative AI, is that increasingly larger proportions of people’s jobs, are now classified as cognitive work.

You may spend part of your day making cognitive work input artefacts – a presentation, a report, an analysis, a meeting, and also spend part of your day engaged in discussion, thinking, considering, often using those input widgets, to produce one or more decidgets – a decision made – such as a decision to launch the new product next month.

Arguably, the higher up the corporate ladder one goes, the greater proportion a person’s job is cognitive work versus task-based work. An EVPs job is almost exclusively to think, with the purpose of making a decision based on that thinking. In fact, most task-based elements of senior leaders’ roles, are delegated to executive assistants, acknowledging the demand and capacity limits imposed by the cognitive workload on these individuals.

Why does this matter?

It matters that we acknowledge that the role of every person in an organization, has some portion of cognitive work attached to it (if it doesn’t, this work is likely soon to be automated or done by AI) so that we can look at productivity outputs differently for that work.

No longer is the presentation a stand-in productivity output measure for the thinking work of the job. What decision did you arrive at, if any, given or regardless of the presentation input?

This is what matters for the cognitive work part of your job.

When we don’t appreciate that decisions are the productivity outputs of cognitive work, we create a situation where the productivity of our organization is slower than it needs to be, or lower quality than it needs to be.

Slower

Decisions drag when ownership is fuzzy, the stakeholder list expands without bound, or meetings multiply without a commitment to decide. The penalty is not abstract: windows of opportunity close, competitors take action, and teams accumulate work-in-progress that clogs the system. Field studies in “high-velocity” industries found that fast decision makers used more information and more alternatives, then coupled decisions quickly to implementation—speed from disciplined process, not shortcuts. (Eisenhardt, 1989; Brown & Eisenhardt, 1998).

Lower quality

When decisions are made without clear objectives, with thin or one-sided information, or with no explicit alternatives, quality suffers. Add group pressures toward conformity and you get classic failure modes: groupthink suppresses dissent and alternative generation, and outcome bias later confuses luck with quality. Empirical research highlights the importance of comparing alternatives through a sound process rather than rushing to a solution and selling it (Janis, 1972/1982; Baron & Hershey, 1988).

How Cognitive Work Productivity Slows or Worsens

Cognitive work productivity slows or worsens when the outputs of thinking—the decisions—are obscured, avoided, or misunderstood.

1. Confusing task-based outputs for cognitive outputs. Central to this article,organizations often mistake task-based work artefacts such as reports, presentations, dashboards, or prototypes for the productivity of cognitive work. These deliverables are input into decision output, not output in their own right. True cognitive productivity occurs when this information is converted into decisions—decidgets—that enable organizational action.

2. Avoidance and deferral of decision-making. Decision-making is uncomfortable and carries accountability, so it is often deferred, delegated, or escalated. People tend to avoid choices when action carries real or perceived risk, when options are poorly structured, or when uncertainty favors the status quo. The result of decision avoidance is misalignment, resulting in wasted effort and rework. Clarity in decisions improves performance. Making decisions visible as short records capturing the choice, alternatives considered, rationale, and owners builds accountability.

3. Misunderstanding the nature of a decision. Many so-called decisions are really proposals or wishes phrased as “let us explore” or “should we not…” or buried under an abstract metaphor (e.g. we need to build a bridge) with no clear statement of what will be done, by whom, and by when. Without a clear decision, explicit ownership over actions to be taken as a result of a decision doesn’t occur resulting in slow down and quality impacts.

4. The AI-driven urgency of decision-making. Generative AI has transformed cognitive work by making information and expertise instantly accessible. Organizations can no longer justify delays by claiming they are waiting for data or analysis. The speed of decision-making has become a critical competitive factor. Those who hesitate risk falling behind competitors who convert information into decisions rapidly. Treating decisions as measurable, repeatable outputs is now essential to maintaining productivity and staying competitive in a world where fast, high-quality choices define success.

The good news

The cognitive work revolution reframes productivity. The true output of cognitive work is not a report, a presentation, or a completed task; it is a decision. By treating decisions as the units of productivity, organizations can make the invisible work of thinking observable, measurable, and improvable. While decision-making can lag when ownership is unclear, stakeholders multiply, or meetings proliferate, this is not an unchangeable or unavoidable problem. With clear framing, documented processes, and attention to alignment, decision quality and speed can improve dramatically.

Treat decisions as the output. Make them visible, owned, dated. Then measure cycle time and quality. That’s how thinking turns into progress.

Authored by Bronwen Jones, for the Cognitive Work Revolution.

References

Anderson, C., 2003. The psychology of doing nothing: Forms of decision avoidance result from reason and emotion. Psychological Bulletin, 129(1), 139-167.

Baron, J., & Hershey, J. C. (1988). Outcome bias in decision evaluation. Journal of Personality and Social Psychology, 54(4), 569–579

Brown, S. L., & Eisenhardt, K. M. (1998). Competing on the edge: Strategy as structured chaos. Boston, MA: Harvard Business School Press.

Chernev, A., Böckenholt, U., & Goodman, J., 2015. Choice overload: A conceptual review and meta-analysis. Journal of Consumer Psychology, 25(2), 333–358.
Christensen, C., 1997. The Innovator’s Dilemma. Harvard Business School Press.

Dean, D., 2024. Decision avoidance in organizations: A contemporary review. Tilburg University Research Portal.

Eisenhardt, K. M. (1989). Making Fast Strategic Decisions in High-Velocity Environments. Academy of Management Journal, 32(3), 543–576

Janis, I. L. (1972/1982). Victims of Groupthink: A Psychological Study of Foreign-Policy Decisions and Fiascoes (2nd ed.). Boston: Houghton Mifflin.

Little, J. D. C. (1961). A Proof for the Queuing Formula L = λW. Operations Research, 9(3), 383–387.

Simon, H. A. (1971). Designing Organizations for an Information-Rich World. In Martin Greenberger (Ed.), Computers, Communications, and the Public Interest (pp. 37–72). Baltimore, MD: The Johns Hopkins Press.

Taylor, Frederick W. (1911). The Principles of Scientific Management. New York: Harper & Brothers.

Locke, E. A., & Latham, G. P., 2002. Building a practically useful theory of goal setting and task motivation. American Psychologist, 57(9), 705-717.

Nygard, M., 2011. Architecture decision records. ResearchGate.
Tushman, M. L., & O’Reilly, C. A., 2006. Ambidextrous organizations: Managing evolutionary and revolutionary change. California Management Review, 38(4), 8-30.