About the tool

How Recommendations Work

A plain-language explanation of how the diagnostic turns your answers into a reflective recommendation.

The diagnostic does not make decisions for you. It helps you reflect on whether a decision is better suited to one accountable person, targeted consultation, a representative group, or a full team decision process.

For the cleanest experience, complete the diagnostic before reading too much detail about how the recommendation is generated.

Why this page exists

This page explains, in plain language, how the diagnostic turns your answers into a recommendation. The tool is designed to support judgement, not replace it. It helps make the decision context more visible so that a team can choose a decision process more deliberately.

The recommendation is not a final instruction. It is a prompt for discussion, reflection and testing.

The diagnostic begins with context, not mechanism. It does not ask whether “the team should decide” or whether “one person should decide”. It applies a published, researcher-authored rubric to the full response pattern.

What the diagnostic looks at

The diagnostic considers several features of the decision context. These features help the rubric assign points and select a participation level that may be sufficient for the context.

The decision itself

Is the decision structured, urgent, risky, reversible, or ambiguous?

Where knowledge sits

Is the relevant expertise concentrated in one person, held by a small specialist group, or spread across the wider team?

Who is affected

Does the decision affect one role, the whole team, customers, communities, or other stakeholders?

Who must act on it

Does implementation depend on one person, a small group, the whole team, or multiple stakeholder groups?

What participation costs

Would involving everyone improve the decision, or create delay, fatigue, duplicated discussion, and decision drag?

What legitimacy requires

Do affected people need meaningful voice, representation, cultural consideration, or implementation ownership?

The basic logic

The diagnostic uses a deterministic, rule-based rubric. It compares the value of wider participation with the conditions that may make narrower participation sufficient.

Individual-decision point index

An additive point index (0–100) built from criteria that favour one accountable person deciding, such as clear accountability, concentrated expertise, concentrated information, low legitimacy need, and high reversibility.

Team-decision-value point index

An additive point index (0–100) built from criteria that favour wider participation, such as distributed knowledge, broad stakeholder impact, high implementation commitment need, high legitimacy need, and high risk.

Participation-overfit proxy

A separate reflection index that checks whether a full team decision may be more participation than the context requires. It is reported for discussion but does not select the participation level.

More participation is not always better. Less participation is not always better either. The aim is contextual fit.

Minimum sufficient participation

The diagnostic is based on the principle of minimum sufficient participation:

Use the least participative decision process that still provides adequate information, expertise, implementation commitment, legitimacy, cultural integrity, relational care and risk control.

This does not mean minimum participation. It means enough participation for the decision to be informed, legitimate, responsible and implementable.

A decision may be made by one accountable person while still requiring consultation, advice, challenge or communication. The key distinction is between having input into a decision and holding final decision authority.

The six participation levels

The recommendation may place the decision on one of six participation levels. These levels are not a maturity ladder. Level 6 is not better than Level 1. The best level depends on the decision context.

LevelLabelPlain explanation
Level 1DecideOne accountable person decides using available information.
Level 2Gather information, then decideThe decision owner collects specific facts or data before deciding.
Level 3Seek advice, then decideThe decision owner asks selected people for advice, challenge or alternatives before deciding.
Level 4Consult affected parties, then decidePeople affected by or responsible for implementing the decision are consulted before the owner decides.
Level 5Representative group decidesA deliberately selected group holds final decision authority.
Level 6Full group decisionThe whole relevant group decides together using consensus, consent, voting or another explicit decision rule.
The ladder should not be read as “more mature” or “less mature”. Each level is appropriate under different conditions.

How the level is selected

The participation level is chosen by the first matching rule in an ordered threshold set. The rules are evaluated from top to bottom, and the first rule whose conditions are met wins. This is a researcher-authored decision procedure, not a machine-learning prediction.

For example:

  • If the team-decision-value index is high, affected stakeholders are broad, and cultural/relational impact is high, the rubric selects Level 6 (full group decision).
  • If the individual-decision index is high, the team-decision-value index is low, and no strong counter-indicators apply, the rubric selects Level 1 or 2 (individual decision after targeted information gathering).
  • If none of the earlier rules match, a default rule assigns Level 3 (seek advice, then decide).
The participation-overfit proxy and the caution flags are reported for reflection, but they do not change the level selected by the ordered threshold rules.

What the point index means

Some results show a point index out of 100, such as the individual-decision point index or the team-decision-value point index. This is not a statistical probability, percentile, or validated prediction. It is a rule-based point total built from your answers under the published rubric.

Use the point index as a prompt for judgement and discussion, not as proof that one decision process is correct.

DoDo not
Use the result to start a better conversation.Treat the result as a final instruction.
Review the reasons behind the recommendation.Treat the point index as scientific certainty.
Check whether important voices are missing.Use efficiency to exclude affected people.
Adapt the decision process to your context.Assume one mechanism is always best.

Ties and ranking

When the diagnostic also ranks decision mechanisms (such as directive, advice process, consensus, majority, delegated authority, or shura), each mechanism starts from the same 50-point baseline. Configured adjustments are applied, the total is rounded, and the result is bounded to a 15–96 range. Mechanisms are then ranked using competition ranking: tied mechanisms share the same rank, and the next rank skips ahead.

If two mechanisms end on the same point total, both are shown as tied at that rank. The display cutoff is generous: every mechanism tied at the cutoff rank is included, so a mechanism tied for third place is never silently discarded.

When the tool adds caution

The diagnostic may add caution flags when the context suggests that a narrow decision process could miss important knowledge, exclude affected people, or underestimate risk.

Caution may appear when there is:

  • High risk
  • Low reversibility
  • Broad stakeholder impact
  • High legitimacy need
  • Missing or under-weighted voices
  • Cultural, relational, ethical or wellbeing implications
  • Psychological safety or power concerns
  • Uncertainty about who is affected
  • Uncertainty about where relevant knowledge sits
A caution flag does not automatically mean the whole team must decide. It means the decision process needs more care, consultation, challenge or review.

How your responses are used

Your diagnostic responses are used to generate your recommendation. Where consent is provided, responses may also support this research project.

You can complete the diagnostic without providing your name or email address. If you choose to provide contact details, they are used only for the purposes you consent to, such as receiving your results or follow-up communication.

Research reporting uses anonymised or non-identifying data.

Responses are described as anonymised or non-identifying for research reporting rather than fully anonymous, because optional identifying details may be provided.

Scoring version and transparency

The current scoring procedure is v3.0-deterministic-rubric. Every recommendation is stamped with the scoring version, the rule-set identifier (decision-compass-rubric-v3.0), and a hash of the configured points, bounds, ranking procedure, and ordered threshold rules.

The meaning of a previously published scoring version is never overwritten. If the rubric changes, a new version constant is added and the old version remains available for reference.

The browser shows a provisional result as you complete the diagnostic. The authoritative result is recalculated on the server when you submit, and any client-supplied scores are ignored. This keeps the stored result, email, and PDF aligned with one shared scoring implementation.

These results are calculated by deterministic, researcher-authored points and thresholds. They are not probabilities, percentiles, statistical effect estimates, or validated predictions. The values support comparison and reflection within the current scoring rubric only.

Research basis

The diagnostic draws on research about participative decision-making, group performance, psychological safety, information sharing, coordination cost, team reflexivity and culturally responsible decision-making.

It also draws on the project white paper, “When the Team Should Not Decide”, which introduces the concepts of minimum sufficient participation and decision-participation overfit.

The concepts of decision-participation overfit, participation inflation and minimum sufficient participation are conceptual syntheses developed for this research. They are intended to support reflection and evaluation, not to operate as validated diagnostic constructs.

For examiners and researchers

A more detailed methodological explanation is available in the examiner-facing section. That page describes the diagnostic fields, scoring logic, thresholds, data storage, limitations and ethics in more depth.

Participants are encouraged to complete the diagnostic before reading the detailed examiner explanation, to reduce response priming.

Use the result as a conversation starter

The recommendation is most useful when it helps a team ask better questions:

  • Is this decision being made with enough information?
  • Are the right people being consulted?
  • Is the process slower than the decision requires?
  • Are we excluding affected people or relevant knowledge?
  • Does the decision need individual accountability, shared authority, or both?

The diagnostic is not trying to remove judgement. It is designed to make judgement more explicit.