Individual and organizational antecedents of strategic foresight: A representational approach

Paper highlights

Strategic foresight is the ability to identify a superior course of action, especially one markedly different from the status quo. This study measures foresight through the accuracy of predictions about startup outcomes and asks why some people and groups predict those outcomes more accurately than others. Its answer has two parts: individuals do better when their mental representations cover more relevant factors and align more closely with informed consensus; groups do better mainly because averaging their predictions cancels errors.

The group result is especially useful. The observed advantage did not come primarily from discussion producing a better shared model of the problem. Statistical groups created by averaging members’ earlier individual forecasts performed almost as well. In this setting, combining judgments mattered more than combining mental representations.

Study design

The study analyzes a strategy exercise completed by 358 MBA students. Participants evaluated startup opportunities, explained which factors shaped their judgments, made predictions individually, and then worked in groups. Those responses allow the study to measure foresight, the breadth and consensus of individual representations, and the separate effects of prediction aggregation and group deliberation.

Results at a glance

• Broader representations were associated with more accurate individual predictions.
• Representations that agreed more closely with the sample’s consensus were also associated with better foresight.
• Groups outperformed individuals, including groups of only two people.
• Most of the group advantage was reproduced by averaging individual predictions made before deliberation.

Why it matters

• Strategic foresight depends partly on what evaluators include in their representation of a problem.
• Statistical groups performed nearly as well as deliberating groups, suggesting that most of the group advantage came from averaging independent errors rather than discussion.
• A useful sequence is to collect forecasts independently, aggregate them, and then use discussion to examine disagreement.

How to use this paper

Cite this for

• An empirical study linking mental representations, prediction aggregation, and strategic foresight.
• Evidence that broader and more consensual representations are associated with more accurate individual forecasts.
• Evidence that statistical groups created by averaging individual forecasts can approach the performance of deliberating groups.

Useful for teaching

• How to measure strategic foresight with predictions made before outcomes are known.
• Why collecting independent forecasts before discussion can preserve useful disagreement.
• The difference between aggregating predictions and building a shared representation through deliberation.

Careful claim

In this MBA startup-prediction exercise, groups outperformed individuals largely through forecast aggregation; the result does not imply that deliberation lacks value in every strategic task.

Abstract

The ability to make predictions about strategic outcomes—what we term strategic foresight—is central to most theories of competitive advantage. This paper identifies individual- and organization- level antecedents of strategic foresight by analyzing an exercise taken by 358 MBA students. Among the individual antecedents, we show that two characteristics of mental representations (namely, their breadth and agreement with consensus) are positively related to strategic foresight. Comparing individual to group performance reveals that groups exhibit greater strategic foresight than do individuals. Finally, from comparing the performance of real-life groups with “statistical” groups (for which decisions are computed by averaging the predictions of individuals before they become group members), we find that the superiority of group performance is due mostly to aggregating predictions, not representations.

Other links

• SSRN preprint

Last updated 2026-06-21