Decisions
A decision is a commitment to a course of action taken before the evidence is complete. This page covers what every decision has in common; the seven pages beneath it carry the detail, from how an expert decides alone to how a group decides badly.
1. Three levels of decision
The level sets how much machinery a decision earns.
- Operational — everyday, quick, little structure. Example: approving a pull request.
- Tactical — medium-term, directing resources in support of strategy. Example: choosing the project’s testing framework.
- Strategic — long-term direction, goals and values. Example: adopting Agile across the organization.
2. Analysis paralysis
The characteristic team failure is not deciding badly but not deciding at all. Buridan’s ass starves between two equal bales: options too alike to separate prevent any being taken. Overweighting drawbacks does the same, since every option has some. The cost is delay and lost opportunity; the question underneath is never fully answerable — how much information is enough?
3. What a decision is made of
A decision has four inputs — Problem, Experience, Assumptions, Knowledge — and two outputs: the solution, and the assumed risk [1]. Teams reliably produce the first output and skip the second, which is what leaves a decision impossible to review later. See the PEAK model.
4. Which page to read next
| Page | What it covers | The core claim | How well evidenced |
|---|---|---|---|
| The PEAK model | Inputs, process, outputs | A decision has two outputs — solution and assumed risk | A framework, not a finding |
| Intuition and expertise | Recognition-primed decisions | Experts recognise one action rather than compare options | Coded by its own research team |
| Thin slicing | Snap judgment, and where it misfires | Short exposures carry the signal — and the bias | Popular retelling, not research |
| Group decisions | Styles, techniques, the seven-step method | Agree how you will decide before you decide | Extension leaflets: guidance, no data |
| Groupthink | Janis’s construct, and the Abilene paradox | Hierarchy, not friendliness, silences dissent | Weakest — only partly confirmed |
| Cognitive bias | 37 biases mapped across software engineering | The measured biases are individual, and land on estimation | Strongest: 65 studies, lab-heavy |
| Deciding with AI | Reliance on model recommendations | Miscalibration runs in both directions | Two small studies, one a preprint |
How solid is this?
- Where it comes from. The levels, analysis paralysis and the input–process–output model are lecture material [2] — vocabulary, not research findings.
- The “decision-driven organization” advice is consultancy practice. It comes from Rogers and Blenko’s Who Has the D? and Blenko, Mankins and Rogers, The Decision-Driven Organization — widely used, but published without measurement, so its gains are still to be verified.
- What is contested. On ambiguity CHAOS 2015 reports the opposite — vague objectives 38% successful against precise 22% [3] — from a crosstab with no size control and no significance test, so it is no evidence either way. See Why projects fail [4].
Acknowledgments
This content is heavily inspired by and adapted from lectures by Eduardo Miranda and David Root [2] on software project management. The structure, examples, and pedagogical approach reflect their teaching materials and frameworks.
References
- C. L. Hoover, M. Rosso-Llopart, and G. Taran,Evaluating Project Decisions: Case Studies in Software Engineering. Addison-Wesley Professional, 2010. [Online]. Available: https://www.oreilly.com/library/view/evaluating-project-decisions/9780321685629/
- D. Root,Managing Software Development. Carnegie Mellon University / Innopolis University lecture series, 2014.
- The Standish Group International,“CHAOS Report 2015,” The Standish Group International, Inc., 2015.
- J. L. Eveleens and C. Verhoef,“The Rise and Fall of the Chaos Report Figures,” IEEE Software, vol. 27, no. 1, pp. 30–36, 2010, doi: 10.1109/MS.2009.154.
Disclaimer: AI is used for text polishing and explaining. Authors have verified all facts and claims. In case of an error, feel free to file an issue.