An efficient frontier in organization design: Organizational structure as a determinant of exploration and exploitation

Paper highlights

Organizations face two decision errors: they can approve bad projects or reject good ones. A structure that reduces one error often increases the other. This paper maps that trade-off as an efficient frontier and shows how hierarchy, independent decision-making, and committees place organizations at different points along it.

This gives exploration and exploitation a concrete decision-process foundation. Structures that make approval easier explore more: they pursue more projects and miss fewer good ones, but accept more bad ones. Structures with several approval screens exploit more cautiously: they suppress bad projects, but also lose promising ones.

How the model works

The formal model treats each member as a fallible evaluator and defines structures by the number of favorable judgments required for approval. It compares many combinations of group size and consensus threshold, allowing designs to be evaluated by their joint omission and commission errors rather than by a single performance score.

Main results

• Organization designs lie on or behind an efficient frontier: some structures reduce one error only by increasing the other, while some are dominated on both dimensions.
• Adding decision makers can move the frontier outward, but the effect of size depends on the decision rule. Larger polyarchies explore more; larger hierarchies screen more aggressively.
• Majority-rule and other hybrid structures can achieve combinations of exploration and exploitation that alternating between highly exploratory and highly exploitative units cannot reproduce.

Why it matters

• A structure behind the efficient frontier makes more of both errors than another feasible structure; choosing a point on the frontier depends on which error costs more.
• Exploration follows from permissive decision rules, while exploitation follows from additional screening.
• Calls for “more exploration” are incomplete unless they also name the extra false positives the organization will tolerate.

How to use this paper

Cite this for

• An efficient-frontier view of organization design based on omission and commission errors.
• Decision structures as approval thresholds that shape exploration and exploitation.
• A process-level explanation of how hierarchy, polyarchy, and committees produce different error profiles.

Useful for teaching

• How to move from abstract calls for exploration to concrete decision rules.
• Why adding decision makers can help or hurt depending on the aggregation structure.
• The difference between choosing a point on a frontier and using a dominated design.

Careful claim

The model shows feasible trade-offs among organizational structures; choosing a point on the frontier depends on the relative cost of missed opportunities and mistaken approvals.

Abstract

This paper develops a parsimonious process-level theory that connects organizational structure to exploration and exploitation. Toward this end, it develops a mathematical model of organizational decision making that combines an information processing approach in the spirit of Sah and Stiglitz (1986) with elements from signal detection theory (Green and Swets, 1966). The model is first used to explore a “design space” of organizations and identify trade-offs and dominance relationships among alternative organization designs. The paper then studies open questions in the organization design literature, such as the extent to which exploration and exploitation can be produced by one organization and what is the effect of organization size on exploration. More broadly, this research speaks to calls for introducing more process-level explanations in the organizations literature. The paper concludes with testable hypotheses and managerially relevant insights.

Other links

• SSRN preprint

Last updated 2026-06-21