discover-discuss-decide-framework - SKILL.md Agent Skill

name: discover-discuss-decide-framework description: A structured process for group decision-making that eliminates groupthink, surfaces hidden disagreements, and increases hit rates. Use this when building product roadmaps, hiring for key roles, or setting strategic priorities.

The "Discover-Discuss-Decide" framework transforms meetings from coercive "alignment" sessions into high-signal information exchanges. By decoupling the generation of opinions from the discussion of those opinions, you surface the true spread of team knowledge and reduce the influence of the loudest or most senior voices.

The Three Phases

1. Independent Discovery (Asynchronous)

Collect data points, forecasts, and opinions independently before anyone speaks to each other.

Identify the Core Question: Define what you are trying to solve (e.g., "Which feature should we prioritize for Q3?").
Solicit Independent Input: Send a prompt to every stakeholder. They must respond individually without seeing others' answers.
Required Outputs:
- Forced Rankings: Ask for a prioritized list of options.
- Rationales: Require a 3-5 sentence "why" for the rankings.
- Forecasts: Ask for point estimates (e.g., "70% probability this funds at Series A") and range estimates (Lower Bound/Upper Bound).
- Ratings: Use a 1-7 scale for specific criteria (e.g., "Rate founder quality 1-7"). Define what a "4" vs. a "7" means beforehand to ensure shared definitions.

2. Focused Discussion (Live Meeting)

The meeting is strictly for discussing the variance in the independent inputs.

Map the Disagreement: Share the gathered data (spreadsheets/charts) showing the range of opinions.
Prioritize Differences: Do not spend time on areas where everyone agrees. Focus 100% of the time on the outliers and the widest spreads in ratings or forecasts.
Exercise "Curiosity, Not Coercion":
- The goal is to convey information, not to convince others you are right.
- The Reflection Technique: As a facilitator, reflect back what a participant says without agreeing: "So what I hear you saying is [X] because of [Y]. Did I get that right?"
- Prohibit Interruptions: Silencing others through interruption destroys decision quality.

3. Independent Decision

Once the discussion is over, move to the final decision.

Avoid "Alignment": Do not force a room-wide consensus. Alignment is often a mask for coercion.
The Single DRI Model: Ideally, one Directly Responsible Individual makes the call after hearing all arguments.
The Partnership Model: If voting is required, stakeholders vote independently and asynchronously after the meeting so their final tally is not influenced by "following the leader."

Shortening the Feedback Loop

When the final outcome is years away (e.g., long-term R&D or seed investing), create "Kill Criteria" and "Surrogate Markers" during the Discovery phase:

Identify Correlated Signals: What must happen in the next 6-12 months for the 10-year goal to be possible? (e.g., "Must hit Series A funding," "Must achieve 20% MoM growth").
Set Kill Criteria: Define specific signals that would cause you to pivot or stop.
- Example: "If the customer refuses to demo and only asks about price, we kill the deal."
Make it Explicit: Record these markers at the time of the decision so you can judge the decision quality months later, rather than waiting a decade for the outcome.

Examples

Example 1: Product Roadmap Prioritization

Context: A PM team is deciding which three features to build next quarter.
Discovery: Each PM submits a forced rank of 10 features with a 3-sentence rationale for their Top 2.
Discussion: The lead PM displays a heat map showing Feature A was ranked #1 by two people and #10 by two others. The meeting focuses exclusively on why those four people see Feature A so differently.
Decision: The Head of Product reviews the rationales and the discussion notes, then selects the final three features.

Example 2: Project Timeline Forecasting

Context: An engineering team is estimating a complex launch.
Discovery: Every engineer independently submits: "How many sprints will this take?" (Lower bound, Upper bound, and Confidence level).
Discussion: The team discovers the Lead Engineer said 4 sprints, while the Junior Engineer said 12. The facilitator asks the Junior Engineer to explain the risks they see that the Lead might be overlooking.
Decision: The Project Lead sets the official deadline based on the aggregated range.

Common Pitfalls

The "Alignment" Trap: Believing everyone must agree before moving forward. This leads to watered-down decisions and hidden resentment. Aim for being "heard," not "agreed with."
Discussing Agreement: Spending 40 minutes of a 60-minute meeting having people "double-click" on things they already agree on.
Implicit Models: Relying on "gut feel" or "I know a good founder when I see one." Force the intuition into explicit 1-7 ratings or percentage forecasts so it can be audited later.
Resulting: Judging a decision solely by its outcome. A good decision can lead to a bad result (luck), and a bad decision can lead to a good result. Focus on the process of gathering and discussing information.