Complete firstComparable subject
Close scope, unit, WBS, included work, and estimation question in writing; appoint a coordinator who does not estimate.
Without a common frame, values are not comparable.
Method guideDecision Making
Wideband Delphi
A practical method guide: clarify prerequisites and materials, understand the complete process, and look up outputs, pitfalls, and stop criteria.
ComplexityHigh
Time1-4 h or multiple rounds
Participants4-12 Experten
FormatWorkshop + async
MaturityCanonical
01
Prerequisite
What needs to be finished first
Complete firstEvidence access
Shared scope/WBS, Definition of Done, historical analogies, private estimate sheets, and anonymous round summary
Missing data remains visible as uncertainty.
02
Preparation
What needs to be ready before start
Materials
Shared scope/WBS, Definition of Done, historical analogies, private estimate sheets, and anonymous round summary
People / roles
Coordinator facilitates · 3–7 expert estimators · scope owner answers · scribe records assumptions
Pre-read
Close scope, unit, WBS, included work, and estimation question in writing; appoint a coordinator who does not estimate.
Time needed
90–180 minutes for 2–4 anonymous estimation rounds; preparation separate.
Setup
Prepare a workspace with Brief · Private · Reflect · Clarify · Re-estimate; separate example from live data.
03
Core question
The one question this method answers
What estimate range emerges when experts first estimate independently, compare assumptions anonymously, and re-estimate with new information?
04
Flow
Phases with goal, actions, output and pitfall
Several experts estimate in parallel, compare their assumptions, and adjust after feedback. Ranges, uncertainty, and strong dissent become visible without forcing an apparent average. The approach strengthens planning discipline where numbers remain approximations.
Wideband Delphi · working structure
Fictional example
Fictional values show the actual method fields.
| Round/person | Estimate | Assumptions | Evidence | Anonymous feedback | Revision | Status |
|---|---|---|---|---|---|---|
| R1 · E1Add evidence and rationale. | 18 PD | No legacy migration | Token API exists | Range 18–30; migration disputed | 24 PD | Updated |
| R1 · E2Add evidence and rationale. | 24 PD | Five browsers, full tests | Last checkout item | Test scope confirmed | 26 PD | Updated |
| R1 · E3Add evidence and rationale. | 30 PD | Certification internal | Compliance experience | Certification external | 27 PD | Updated |
| OutcomeAdd evidence and rationale. | 24–27 PD | Migration included; certification external | Three private revisions | Residual spread from test depth | Review after migration probe | Stop rule met |
1. Estimate
- R1 · E1
- 18 PD
- R1 · E2
- 24 PD
- R1 · E3
- 30 PD
- Outcome
- 24–27 PD
2. Assumptions
- R1 · E1
- No legacy migration
- R1 · E2
- Five browsers, full tests
- R1 · E3
- Certification internal
- Outcome
- Migration included; certification external
3. Evidence
- R1 · E1
- Token API exists
- R1 · E2
- Last checkout item
- R1 · E3
- Compliance experience
- Outcome
- Three private revisions
4. Anonymous feedback
- R1 · E1
- Range 18–30; migration disputed
- R1 · E2
- Test scope confirmed
- R1 · E3
- Certification external
- Outcome
- Residual spread from test depth
5. Revision
- R1 · E1
- 24 PD
- R1 · E2
- 26 PD
- R1 · E3
- 27 PD
- Outcome
- Review after migration probe
6. Status
- R1 · E1
- Updated
- R1 · E2
- Updated
- R1 · E3
- Updated
- Outcome
- Stop rule met
All values are fictional.
- 1
Brief
Shared scope
Goal
Shared scope: Coordinator explains scope, unit, and historical evidence; answers are shared equally.
Inputs for this step
- Close scope, unit, WBS, included work, and estimation question in writing; appoint a coordinator who does not estimate.
- Shared scope/WBS, Definition of Done, historical analogies, private estimate sheets, and anonymous round summary
Actions
- Coordinator explains scope, unit, and historical evidence; answers are shared equally.
Output
Shared scope
Coordinator explains scope, unit, and historical evidence; answers are shared equally.
Quality check
Is everyone estimating the same work in the same unit?
Pitfall
Some experts receive extra corridor information. Distribute every answer in writing to the whole panel.
Timing: 15–25 minutes
- 2
Private
Round 1
Goal
Round 1: Each expert independently estimates total or WBS parts and records assumptions and uncertainty.
Inputs for this step
- Shared scope
Actions
- Each expert independently estimates total or WBS parts and records assumptions and uncertainty.
Output
Round 1
Each expert independently estimates total or WBS parts and records assumptions and uncertainty.
Quality check
Were values submitted independently with rationale?
Pitfall
Early group discussion creates authority anchoring. Submit the first round privately and simultaneously.
Timing: 15–25 minutes
- 3
Reflect
Anonymous summary
Goal
Anonymous summary: Coordinator shows distribution and rationales anonymously without making the mean a target.
Inputs for this step
- Round 1
Actions
- Coordinator shows distribution and rationales anonymously without making the mean a target.
Output
Anonymous summary
Coordinator shows distribution and rationales anonymously without making the mean a target.
Quality check
Are range and diverging assumptions visible?
Pitfall
Only mean 24 is shown and differences disappear. Show individual values, range, and assumptions anonymously side by side.
Timing: 10–20 minutes
- 4
Clarify
Clarified assumptions
Goal
Clarified assumptions: Panel discusses reasons, updates shared scope, and identifies missing evidence without defending people.
Inputs for this step
- Anonymous summary
Actions
- Panel discusses reasons, updates shared scope, and identifies missing evidence without defending people.
Output
Clarified assumptions
Panel discusses reasons, updates shared scope, and identifies missing evidence without defending people.
Quality check
Did the discussion create information rather than consensus pressure?
Pitfall
Coordinator pushes a compromise value. Clarify factual assumptions only; protect private judgement until the next round.
Timing: 20–30 minutes
- 5
Re-estimate
Delphi range
Goal
Delphi range: Estimate privately again; after stable spread, document range, assumptions, and review condition.
Inputs for this step
- Clarified assumptions
Actions
- Estimate privately again; after stable spread, document range, assumptions, and review condition.
Output
Delphi range
Estimate privately again; after stable spread, document range, assumptions, and review condition.
Quality check
Is residual spread explainable and the stop rule met?
Pitfall
Rounds continue until every value matches. Stop at a stable explained range and preserve uncertainty.
Timing: 15–30 minutes
05
Artifact
What comes out at the end
Sketch: what the artifact looks like
Form
Traceable Delphi estimate with individual rounds, anonymous rationales, shared scope, residual spread, and review trigger.
Versioning / ownership
Record date, team, data state, and review date; preserve prior versions.
Tool alternatives
- Scope and WBS sheet
- Private estimate sheets
- Anonymous round summary
- Assumption log
wideband-delphi-working-template.md
Set subject, data, roles, flow, and review.
Wideband Delphi: Worksheet
Question
To be confirmed
Desired outcome
To be confirmed
Scope
Complete run Time: 90–180 minutes for 2–4 anonymous estimation rounds; preparation separate.
Complete this worksheet on paper or in your own document during the work.
Preparation for this scope
Method setup
- Materials: Shared scope/WBS, Definition of Done, historical analogies, private estimate sheets, and anonymous round summary
- Roles: Coordinator facilitates · 3–7 expert estimators · scope owner answers · scribe records assumptions
- Advance information: Close scope, unit, WBS, included work, and estimation question in writing; appoint a coordinator who does not estimate.
- Overall time needed: 90–180 minutes for 2–4 anonymous estimation rounds; preparation separate.
- Setup: Prepare a workspace with Brief · Private · Reflect · Clarify · Re-estimate; separate example from live data.
Prepare for the work steps
Brief: Shared scope
- Close scope, unit, WBS, included work, and estimation question in writing; appoint a coordinator who does not estimate.
- Shared scope/WBS, Definition of Done, historical analogies, private estimate sheets, and anonymous round summary
Private: Round 1
- Shared scope
Reflect: Anonymous summary
- Round 1
Clarify: Clarified assumptions
- Anonymous summary
Re-estimate: Delphi range
- Clarified assumptions
Wideband Delphi · working structure
| Round/person | Estimate | Assumptions | Evidence | Anonymous feedback | Revision | Status |
|---|---|---|---|---|---|---|
| R1 · E1 | ||||||
| R1 · E2 | ||||||
| R1 · E3 | ||||||
| Outcome |
- R1 · E1: Add evidence and rationale.
- R1 · E2: Add evidence and rationale.
- R1 · E3: Add evidence and rationale.
- Outcome: Add evidence and rationale.
The blank worksheet contains prompts only.
Reference · original teaching example
Brief: Shared scope
Expected artifact: Shared scope
Coordinator explains scope, unit, and historical evidence; answers are shared equally.
Entry:
...
- Is everyone estimating the same work in the same unit?
Private: Round 1
Expected artifact: Round 1
Each expert independently estimates total or WBS parts and records assumptions and uncertainty.
Entry:
...
- Were values submitted independently with rationale?
Reflect: Anonymous summary
Expected artifact: Anonymous summary
Coordinator shows distribution and rationales anonymously without making the mean a target.
Entry:
...
- Are range and diverging assumptions visible?
Clarify: Clarified assumptions
Expected artifact: Clarified assumptions
Panel discusses reasons, updates shared scope, and identifies missing evidence without defending people.
Entry:
...
- Did the discussion create information rather than consensus pressure?
Re-estimate: Delphi range
Expected artifact: Delphi range
Estimate privately again; after stable spread, document range, assumptions, and review condition.
Entry:
...
- Is residual spread explainable and the stop rule met?
Open questions and next steps
...
Method guide: https://methodatlas.meierhoff-systems.de/en/methods/wideband-delphi/run-sheet
06
Example output
Concrete filled scenario, fictional example
Fictional example for orientation. Counts and timings describe this case and are not universal requirements.
wideband-delphi-beispiel.md
Concrete filled scenario, fictional example
Fictional teaching example: Wideband Delphi · working structure
| Round/person | Estimate | Assumptions | Evidence | Anonymous feedback | Revision | Status |
|---|---|---|---|---|---|---|
| R1 · E1 | 18 PD | No legacy migration | Token API exists | Range 18–30; migration disputed | 24 PD | Updated |
| R1 · E2 | 24 PD | Five browsers, full tests | Last checkout item | Test scope confirmed | 26 PD | Updated |
| R1 · E3 | 30 PD | Certification internal | Compliance experience | Certification external | 27 PD | Updated |
| Outcome | 24–27 PD | Migration included; certification external | Three private revisions | Residual spread from test depth | Review after migration probe | Stop rule met |
07
Pitfalls
Recognize symptoms and steer against them
Quick overview, the full text is in the matching phase in the flow
- Shared scopeSome experts receive extra corridor information.View in phase
- Round 1Early group discussion creates authority anchoring.View in phase
- Anonymous summaryOnly mean 24 is shown and differences disappear.View in phase
- Clarified assumptionsCoordinator pushes a compromise value.View in phase
- Delphi rangeRounds continue until every value matches.View in phase
Trap
Shared scope
Symptom
Some experts receive extra corridor information.
What to do
Distribute every answer in writing to the whole panel.
Trap
Round 1
Symptom
Early group discussion creates authority anchoring.
What to do
Submit the first round privately and simultaneously.
Trap
Anonymous summary
Symptom
Only mean 24 is shown and differences disappear.
What to do
Show individual values, range, and assumptions anonymously side by side.
Trap
Clarified assumptions
Symptom
Coordinator pushes a compromise value.
What to do
Clarify factual assumptions only; protect private judgement until the next round.
Trap
Delphi range
Symptom
Rounds continue until every value matches.
What to do
Stop at a stable explained range and preserve uncertainty.
08
Stop criteria
When to pause, resolve prerequisites, or choose another approach
Stop Brief: Is everyone estimating the same work in the same unit?
Stop Private: Were values submitted independently with rationale?
Stop Reflect: Are range and diverging assumptions visible?
Stop Clarify: Did the discussion create information rather than consensus pressure?
Stop Re-estimate: Is residual spread explainable and the stop rule met?
Finished the method guide?
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.