Research Synthesizer

Cross-analyze your market, competitive, and customer research into a unified strategy brief. No external tools — pure synthesis of research-memory/ data. Output bridges research → execution.

Purpose

Research Synthesizer is the bridge between research and action. It answers:

• What do all our research findings mean when connected together?

• Where do market trends, competitive gaps, and customer needs intersect?

• What should we do first, and why?

The output — research-memory/strategy-brief.md — translates scattered research data into a unified strategy that every execution skill (copy, SEO, email, lead magnet, etc.) can act on.

"Lots of research without synthesis is just a pile of data." — The Boring Marketer

Key distinction: This skill creates NO new data. It reads everything in research-memory/ and finds the connections that individual skills cannot see on their own.

Enrichment chain: This skill → expert-validator (adds expert consensus/divergence).

Modes

Mode When to Use Behavior

Full Synthesis No strategy-brief.md exists, or it's an empty scaffold Run all 5 steps from scratch

Refresh strategy-brief.md already has data Check research-log.md for files updated since last synthesis → re-run affected cross-analyses only

Auto-Load Protocol

On every invocation, BEFORE any analysis:

• Check research-memory/ directory

• If files exist → Read ALL .md files (except README.md)

• Verify required files exist AND have substantive content:

• market-landscape.md — REQUIRED (market definition, size, trends, structure)

• competitive-intel.md — REQUIRED (competitive set, positioning, channels)

• customer-insight.md — REQUIRED (segments, journey, pain points)

• customer-language.md — OPTIONAL (enriches Cross-Analysis 2 with real customer phrases)

• If any REQUIRED file is missing or empty scaffold:

• Name the missing file(s) and the skill that produces it

• Suggest: "Run [skill-name] first, then come back for synthesis"

• STOP — do not attempt partial synthesis

• Assess data richness for each file: Rich / Adequate / Thin

• Tag thin sections: "Cross-analysis may be limited here — consider re-running [skill] for deeper data"

• Check brand-memory/ (read-only) → If exists, use positioning and voice info to align recommendations with brand direction

• If strategy-brief.md has data → suggest Refresh mode

Input Gathering

This skill requires minimal user input — research-memory/ is the primary data source.

Field Required Description

Business goal Optional Current priority (growth, market entry, pivot, retention) — shapes recommendation priority

Analysis focus Optional Specific cross-analysis area (e.g., "pricing vs competitors", "messaging-market fit")

Constraints Optional Budget, team size, timeline — grounds Next Steps in reality

Language Optional 결과물 작성 언어 (default: English)

If brand-memory/ exists, auto-extract business context — no need to ask.

If this is a Refresh, show which research files changed since last synthesis and ask: "Want me to update the affected sections?"

Process

Step 1: Load & Validate Research Data

Goal: Load all research-memory/ files and confirm sufficient data for cross-analysis.

• Read all .md files in research-memory/

• Verify 3 required files have content (not just scaffold headers)

• From each file, extract key data points needed for cross-analysis:

• market-landscape.md → Macro Trends (opportunity/threat tags), Market Structure Map, Seasonality

• competitive-intel.md → Competitive Set, Positioning Matrix, Channel Activity Matrix, Gaps & Opportunities

• customer-insight.md → Audience Segments (with priority), Pain Points & Unmet Needs, Media Consumption Map

• customer-language.md (if exists) → Pain Expressions, Desire Expressions, Trigger Phrases

• Tag each data area: Rich / Adequate / Thin

• If brand-memory/ exists → load positioning, target audience, brand voice for alignment check

• For Refresh mode → compare research-log.md timestamps to identify what changed

Output: Data inventory with richness assessment. Proceed to Step 2 only if all 3 required files pass.

Step 2: Cross-Analysis (3 Matrices)

This is the core of the skill. Each cross-analysis combines TWO OR MORE data sources to reveal insights that no single source shows alone.

Cross-Analysis 1: Market Trends × Competitive Gaps → Opportunities to Seize Now

Connect:

• Macro Trends tagged as "Opportunity" (from market-landscape.md)

• Gaps & Opportunities + Channel Activity gaps (from competitive-intel.md)

Analysis framework:

For each Opportunity trend: → Is there a competitive gap aligned with this trend? → If YES: This is a "seize now" opportunity → Rate: Urgency (High/Med/Low) based on trend timeframe + gap openness → Rate: Attractiveness (High/Med/Low) based on market size of trend + depth of gap

Output format:

Trend Gap Opportunity Urgency Attractiveness

1 [from market-landscape] [from competitive-intel] [synthesized insight] H/M/L H/M/L

Cross-Analysis 2: Customer Pain × Competitor Weakness → Messaging We Can Own

Connect:

• Pain Points & Unmet Needs (from customer-insight.md)

• Positioning Matrix weaknesses + Messaging gaps (from competitive-intel.md)

• Customer Language (from customer-language.md, if available)

Analysis framework:

For each High-severity Pain Point: → Which competitors address this? Which don't? → If UNDERSERVED: This is a messaging opportunity → Find matching customer language (exact phrases from customer-language.md) → Draft a messaging direction that speaks to the pain in customer's own words

Output format:

Customer Pain Competitor Weakness Messaging Direction Customer Language

1 [from customer-insight] [from competitive-intel] [synthesized messaging angle] "[exact phrase]" or N/A

Cross-Analysis 3: Market Structure × Audience Segments → Best Entry Point

Connect:

• Market Structure Map — price tiers, channels, sub-categories (from market-landscape.md)

• Audience Segments with priority ranking (from customer-insight.md)

Analysis framework:

For the Primary Segment: → Which price tier do they occupy? Is this tier crowded or open? → Which channels are they on? (cross-ref with Media Consumption Map) → Which sub-category aligns best with their needs? → Rate entry feasibility: Easy / Moderate / Hard

Repeat for Segment 2 if data is sufficient.

Output format:

Segment Price Tier Channel Sub-Category Entry Feasibility Priority

1 [from customer-insight] [from market-landscape] [cross-ref] [fit] E/M/H 1st

Step 3: Strategic Recommendations (3-5)

Goal: Distill cross-analyses into 3-5 actionable strategic recommendations.

For each recommendation, provide:

Element Description

What Specific action to take

Why Which cross-analysis (CA1/CA2/CA3) + specific insight supports this

Priority High / Medium / Low — based on urgency × impact

Effort Quick Win (1-2 weeks) / Mid-term (1-3 months) / Long-term (3-6 months)

Prioritization logic:

• High Priority + Quick Win → Do first

• High Priority + Long-term → Plan now, start building

• Medium Priority + Quick Win → Easy wins to stack

• Low Priority → Park for later

If user provided business goals → weight recommendations toward that goal. If brand-memory/ loaded → check each recommendation against brand positioning for consistency.

Step 4: Immediate Next Steps (3-5)

Goal: Turn the highest-priority Quick Win recommendations into concrete action items linked to execution skills.

For each Next Step:

Element Description

Action Specific, concrete task ("Write landing page targeting [segment] with [messaging angle]")

Execution Skill Which marketing skill to use (e.g., 06-direct-response-copy , 05-lead-magnet )

Input from Research What research data feeds this action (specific files + sections)

Timeline Estimated time to complete

Success Metric How to measure if it worked

Skill connection map: 03-positioning-angles (positioning) · 06-direct-response-copy (landing pages/ads) · 05-lead-magnet (free offers) · 09-email-sequences (email flows) · 07-seo-content (SEO articles) · 08-newsletter (newsletter) · 10-content-atomizer (repurposing) · 04-keyword-research (keywords)

Step 5: Save & Log

Goal: Write all findings to research-memory/strategy-brief.md and log the execution.

5a. Write strategy-brief.md

Language rule: 섹션 헤더와 테이블 컬럼명은 영어로 유지합니다. 본문, 셀 값, 설명, 분석 텍스트는 사용자가 지정한 언어로 작성합니다. 언어가 지정되지 않으면 English로 작성합니다.

Use the exact schema from references/strategy-brief-schema.md . Key rules:

• Tag every authored section with [research-synthesizer]

• Leave [expert-validator] sections (Expert Consensus, Expert Divergence) as empty scaffold

• Executive Summary: 5-7 findings, each with source file tags like [market + competitive]

• Cross-Analysis tables: Use Step 2 output formats

• Strategic Recommendations: Use Step 3 format (What / Why / Priority / Effort)

• Immediate Next Steps: Use Step 4 format (Action / Skill / Input / Timeline / Metric)

For Refresh mode: Do NOT overwrite the entire file. Update only sections affected by changed research data. Preserve all [expert-validator] sections untouched. Append > Updated: [date] below changed section headers.

5b. Update research-log.md

Append one row to the log:

| [YYYY-MM-DD] | research-synthesizer | Full Synthesis / Refresh | [key insights summary] | None (internal analysis) |

Analysis Quality Standards

Good synthesis = connects 2+ sources → points to specific action → uses concrete data → acknowledges gaps.

Bad synthesis (avoid) = restates single-source findings as "insights" → makes unsourced claims → generic recommendations like "improve marketing."

Data gaps: Tag thin cells with ⚠️ Limited data — run [skill] for deeper insight . Never fabricate connections. 2 strong cross-analyses + 1 flagged > 3 weak ones.

Quality Checklist

Before saving, verify:

• All 3 cross-analyses completed (CA1: Trends×Gaps, CA2: Pain×Weakness, CA3: Structure×Segments)

• Each cross-analysis connects 2+ data sources (not single-source summaries)

• Every insight traces back to specific files and sections in research-memory/

• Executive Summary has 5-7 findings with source file tags

• 3-5 strategic recommendations, each with What/Why/Priority/Effort

• 3-5 next steps linked to specific execution skills

• Data gaps flagged honestly (not papered over)

• [research-synthesizer] tag on all authored sections

• [expert-validator] sections left as empty scaffold

• research-log.md updated with execution record

Example (Abbreviated)

Context: research-memory/ contains data from market-scanner, competitor-finder, competitor-analyzer, audience-profiler, and voice-of-customer — all about "Marketing skill packs for solo marketers."

Executive Summary:

• AI marketing education market growing 12-15% CAGR, but "ready-to-use skill packs" is an empty niche [market + competitive]

• Top competitor weakness: all teach theory, none provide plug-and-play execution templates [competitive + customer]

• Primary segment (solo marketers, 25-40) converts via newsletter → free resource → purchase [customer]

• Customer language centers on "just tell me what to do" — execution anxiety is the #1 pain [customer-language + customer]

• Twitter/X and newsletter are the highest-ROI channels; competitors underinvest in email [competitive + customer]

CA1 — Opportunity: AI democratization trend (🟢) × No "AI + templates" competitor = "The AI Marketing Execution Pack" positioning

CA2 — Messaging: "I learn but can't apply" pain × Competitors only teach theory = "Stop learning. Start doing." (customer phrase: "just give me something I can copy-paste")

CA3 — Entry Point: Solo marketers (Primary) × $99-$299 tier × Newsletter channel = DTC newsletter funnel as entry

Next Steps:

• Write landing page with "Stop learning, start doing" angle → 06-direct-response-copy

• Create free "5 AI Marketing Templates" lead magnet → 05-lead-magnet

• Build welcome email sequence (newsletter → free → paid) → 09-email-sequences

What This Skill Does NOT Do

• Collect new data → This skill reads existing research-memory/ ONLY. For new data, run the appropriate research skill.

• Expert validation → Use expert-validator (adds multi-agent expert review)

• Market research → Use market-scanner , competitor-finder , audience-profiler , voice-of-customer

• Execute marketing → Use execution skills (copy, SEO, email, etc.) — this skill tells you WHAT to execute and WHY

Research Synthesizer stays focused on connecting dots — finding the strategic meaning where different research streams intersect.