analyzer

Analyzer

Compare a wide array of options across carefully selected criteria using a structured matrix. Designed to turn fuzzy "which is better?" questions into clear, evidence-based side-by-side evaluations.

When to Use

• User needs to choose between multiple options (tools, strategies, markets, vendors, approaches)
• A research file or deep dive has surfaced several alternatives that need ranking
• Any decision that benefits from structured multi-criteria comparison

Relationship to Other Research Skills

The Analyzer can work standalone or as part of the research workflow:

• Typically follows research-md-maker — takes the segments from a broad landscape scan and ranks them
• Typically feeds into deep-researcher — the top-ranked options become candidates for deep dives
• If source research files exist, the analyzer must hyperlink to them in the output

Execution Flow

Step 1: Identify the Options

Ask the user:

• What are you comparing? (products, strategies, markets, tools, approaches, etc.)
• Do you already have a list, or should I pull options from an existing research file?

If pulling from existing research:

1. Locate the relevant .md file(s) in the shared drive
2. Extract the options/segments listed
3. Present them to the user for confirmation

Aim for 4-12 options — enough for meaningful comparison, not so many the matrix becomes unreadable.

Step 2: Use or Refine Criteria

If a parent research file (from research-md-maker) exists, it should already contain a Proposed Evaluation Criteria section. Start there:

1. Locate the criteria in the parent research file
2. Present them to the user — confirm, add, remove, or adjust as needed
3. Ensure criteria are differentiating (skip any where all options score the same)

If no pre-identified criteria exist, propose 6-10 criteria that are relevant, differentiating, measurable, and balanced across dimensions (financial, technical, market, strategic, risk, etc.). Present them to the user for confirmation.

Step 3: Ask Output Format

Before building the matrix, ask the user:

• Markdown (.md) — Good for quick reference, lives alongside research files, easy to scan
• Excel (.xlsx) — Good for sorting, filtering, weighted scoring, sharing with stakeholders

Step 4: Score and Build the Matrix

For each option × criterion cell:

1. Research the answer — Use web search, existing .md files, or the user's input
2. Assign a score — Use a consistent scale:
   • Qualitative: Low / Medium / High
   • Numeric: 1-5 scale (1 = worst, 5 = best)
   • Ask the user which scoring style they prefer
3. Add a brief justification — One line explaining why this score, with a citation where possible

Step 5: Produce the Output

Markdown Format (.md)

# [Topic] — Comparison Matrix

**Author:** [Name]
**Date:** [YYYY-MM-DD]
**Source Research:** [Hyperlink to parent research .md file(s) if applicable]

## Context

[1-2 sentences. What decision is this matrix informing? Link to relevant research files.]

## Criteria Definitions

| # | Criterion | Why It Matters | Scale |
|---|---|---|---|
| 1 | [Name] | [One-line rationale] | 1-5 |
| 2 | [Name] | [One-line rationale] | 1-5 |
| ... | ... | ... | ... |

## Comparison Matrix

| Option | Criterion 1 | Criterion 2 | Criterion 3 | ... | Total |
|---|---|---|---|---|---|
| [Option A] | 4 | 3 | 5 | ... | [Sum] |
| [Option B] | 2 | 5 | 3 | ... | [Sum] |
| [Option C] | 5 | 2 | 4 | ... | [Sum] |

## Justifications

### [Option A]
| Criterion | Score | Rationale |
|---|---|---|
| [Criterion 1] | 4 | [One-line explanation with citation] [1] |
| [Criterion 2] | 3 | [One-line explanation] [2] |

### [Option B]
[Same structure]

...

## Summary

**Top ranked:** [Option] with a score of [X]
**Runner-up:** [Option] with a score of [X]

### Caveats
- [Any important nuance the scores don't capture]
- [Criteria that were close calls]
- [What would change the ranking]

## References

[Vancouver style]
[1] Author. Title. Source. Year. URL.
[2] ...

Excel Format (.xlsx)

Create a workbook with three sheets:

Sheet 1: Matrix

• Row 1: Headers (Option | Criterion 1 | Criterion 2 | ... | Total)
• Rows 2+: One row per option with scores
• Final column: SUM formula for total score
• Conditional formatting: color scale (red=1, yellow=3, green=5)

Sheet 2: Justifications

• Columns: Option | Criterion | Score | Rationale | Source
• One row per option × criterion combination

Sheet 3: Criteria Definitions

• Columns: # | Criterion | Why It Matters | Scale
• Documents what each criterion means and how it's scored

Step 6: File Placement

Comparison matrices live inside the Research/ folder of the topic folder.

• If a topic folder structure exists ([Topic_of_Research]/Research/), place the matrix there:

  [Topic_of_Research]/
  ├── [Topic_of_Research].docx
  └── Research/
      ├── README.md
      ├── Landscape_Research.md
      ├── [Topic]_Comparison_Matrix.md   ← or .xlsx — goes here
      ├── Segment_A/
      │   └── Segment_A_Deep_Dive.md
      └── ...
  

• If the matrix is comparing options within a specific segment, place it inside that segment's subfolder within Research/
• If no topic folder exists yet, create the structure:

  [Department_Folder]/
  └── [Topic_of_Research]/
      └── Research/
          ├── README.md
          └── [Topic]_Comparison_Matrix.md
  

• Name: [Topic]_Comparison_Matrix.md or [Topic]_Comparison_Matrix.xlsx
• Follow naming conventions: underscores, descriptive name
• Follow author sign-off rules from contribution guidelines
• Update the Research/README.md to reflect the new file

Step 7: Hyperlinks

1. Link to source research files — If the options came from a research-md-maker or deep-researcher file, hyperlink to it in the header and Context section (use relative paths within the Research/ folder)
2. Link from source file back — Update the parent research .md to add a note: See also: [Topic_Comparison_Matrix.md](link) or [Topic_Comparison_Matrix.xlsx](link)
3. Cross-link deep dives — If any options have deep-dive documents in segment subfolders, link to them in the Justifications section

Sources & Citations (Vancouver Style)

Use Vancouver citation style — numbered references in order of first appearance.

Rules:

• Every score justification should cite its source where possible
• Scores based on the author's judgment must be marked [Author's assessment]
• Numbers assigned in order of first appearance
• Prefer quantitative sources (market data, benchmarks, metrics) over opinion
• Include URLs for all web-accessible sources