Stg Scoring Problems

Problem Scoring

Score problem candidates using four-property framework with compression-model elimination. Every score carries a tier label and cited evidence. Output is a problem hypothesis, not a canvas section.

Procedure

Step 1: Load Context [S]

Read: segment hypothesis (or candidates if segment not yet finalized), governor's problem space description, research signals from BUILD phase 1.

Produce: problem generation parameters anchored to segment.

Gate: context_loaded: bool -- segment context and problem space available.

• Pass: Step 2.
• Fail: Report missing inputs. If no segment data, proceed with governor input alone, noting reduced confidence.

Step 2: Enumerate Candidate Problems (Minimum 5) [K-grounded]

Grounded in: segment context, governor's problem description, research signals.

For the target segment, identify 5-7 candidate problems from:

• Governor input (what they believe the problem is)
• Public signals (forums, reviews, job postings, support tickets)
• Competitive analysis (what alternatives solve -- implies the problem)
• Adjacent segment patterns (problems common in related segments)

WebSearch for segment-specific pain signals.

Produce: candidate problem list.

Gate: candidates_enumerated: bool -- at least 5 candidates identified from at least 2 different sources.

• Pass: Step 3.
• Fail: If fewer than 5, broaden search to adjacent problem categories. If still 1M potential buyers | T1 if from market sizing; T2 if estimated |

| 4 | 100K-1M | Cross-reference with segment size from stg-segmenting-customers | | 3 | 10K-100K | | | 2 | 1K-10K | | | 1 | 80% satisfaction."

Produce: complete problem hypothesis in register format.

Gate: hypothesis_written: bool -- claim is solution-independent, kill condition references observable thresholds, possibility space records all candidates.

• Pass: Done.
• Fail: If claim contains solution language, rewrite as pain statement. If kill condition is vague, add specific thresholds.

Quality Criteria

• Minimum 5 candidate problems enumerated
• Each problem scored on all four properties with tier labels
• Every score cites specific evidence (not "likely high frequency")
• Governor's stated problem is evaluated honestly (not auto-promoted)
• Composite score used for ranking, not as false-precision metric
• Kill condition references specific observable thresholds
• Problem statement is independent of solution ("customers struggle with X" not "customers need a tool that does Y")

Failure Modes

| Mode | Signal | Recovery | |------|--------|----------| | Solution-shaped problem | Problem statement includes solution language ("need a tool", "need a platform") | Rewrite as pain: "teams spend 40 hours building design systems" not "teams need an automated design system generator" | | Governor's problem auto-promoted | Governor's stated problem has highest score despite weak evidence | Explicitly compare evidence quality between governor's problem and alternatives. If governor's problem has weaker evidence but higher score, flag: "Your stated problem scored highest but evidence is thinner than alternative X. This may be confirmation bias." | | All problems score similarly | Top 3 problems within 20% of each other | This is useful information, not a failure. Carry all 3 as alternatives. Recommend governor input to distinguish |

Boundaries

In scope: Problem enumeration, four-property scoring with evidence, compression-model elimination, possibility space recording, hypothesis writing in register format.

Out of scope: Market sizing (stg-sizing-markets), segment definition (stg-segmenting-customers), competitive mapping (stg-analyzing-competition), solution design (stg-designing-solutions).

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: BellaBe
• Source: BellaBe/strategy-os
• License: MIT
• Homepage: https://bellabe.github.io/leanos/

Install and usage instructions live in the source repository linked above.