synthesis-build

name	synthesis-build
description	Synthesis BUILD stage — the FINAL stage of the Diffmode growth-tactics pipeline (fuses white-space ideation + founder-fit adaptation & merge — formerly two separate synthesis passes — into ONE reviewer-gated stage). Reads synthesis-explore.md (the validated emergent mechanisms) + synthesis-constraints.json + growth-factors.json + founder-input + audience-jtbd + competitors-analysis + the 3 think-tank reports + the bundled 2026 channel menu. Phase 1 IDEATES white-space tactics IGNORING constraints; Phase 2 adapts them for founder constraints (preserving the unconventional core) + generates synergy/founder-fit tactics; Phase 3 merges into a final synthesis.md of 7-9 novel demand-gen tactic ideas (target 50%+ unconventional). Uses the lite anti_patterns (NOT a proprietary anti-vectors DB). The free plugin STOPS here.
metadata	{"version":"1.0.0"}

Synthesis — Build (Phase 1 White Space → Phase 2 Founder Fit → Phase 3 Merge) — FINAL stage

This is the last stage of the free pipeline. It fuses three jobs that MUST stay ordered:

1. Phase 1 — IDEATE (white space): write 4-5 genuinely unconventional white-space tactics in full, IGNORING founder constraints (exploration mode).
2. Phase 2 — ADAPT & GENERATE (founder fit): review the Phase-1 tactics and adapt for founder constraints preserving the unconventional core, then generate 4-5 more from the synergy + founder-fit pools.
3. Phase 3 — MERGE: merge into a final synthesis of 7-9 total tactics (target 50%+ unconventional).

It produces a portfolio of novel demand-gen tactic IDEAS and stops. (Prioritization, implementation guides, and the proprietary 576-vector DB are the paid product — do not attempt them here.)

Vector definitions come from growth-factors.json; pools/anti-patterns from synthesis-constraints.json; the validated emergent mechanisms come from synthesis-explore.md. Do NOT read tactics_DB/ (clean-room — no proprietary anti-vectors DB; use the lite anti_patterns field instead).

The three phases are WALLED — do them strictly in order

You MUST write the Phase-1 white-space tactics in FULL before you adapt anything in Phase 2. You have NOT earned the right to reject a tactic for impracticality during Phase 1 — that is Phase 2's job. The output PROVES the ordering held: the ## Pass 1 Disposition table (Phase 2) can only dispose tactics that Phase 1 already wrote in full. If you catch yourself filtering a Phase-1 tactic for "the founder can't do this," STOP — that belongs in Phase 2.

Inputs & Output

• WS/03-think-tanks/demand-generation/synthesis-explore.md (PRIMARY — the validated emergent mechanisms + blind combinations; Phase 1 ideates from these + diverse_white_space)
• WS/03-think-tanks/demand-generation/synthesis-constraints.json (diverse_white_space; pools B_synergy + C_founder_leverage; anti_patterns; category_diversity_requirements)
• WS/03-think-tanks/demand-generation/growth-factors.json (vector definitions)
• WS/01-diagnostics/founder-input.md (Phase 1 READS but does NOT filter by constraints; Phase 2 applies them; goal + deadline)
• WS/02-enrichment/audience-jtbd.md, WS/02-enrichment/competitors-analysis.md
• Think-tank context: competitor-gaps.md, platform-arbitrage.md (channel selection, structural advantage), cross-industry.md (transferable mechanisms / adaptation angles for the new Phase-2 synergy + founder-fit tactics)
• ${CLAUDE_PLUGIN_ROOT}/reference/Marketing-Channel-Menu-2026.md (bundled channel menu)
• OUTPUT: WS/03-think-tanks/demand-generation/synthesis.md

═══════════════════════════════════════════════════════════════════════════

PHASE 1 — IDEATE (white space; IGNORE founder constraints)

═══════════════════════════════════════════════════════════════════════════

Generate 4-5 tactics exclusively from diverse white-space combinations — vector pairs that are novel for this run and exclude over-represented "content-flywheel"-type vectors. This phase prioritizes unconventional exploration over practicality and IGNORES founder constraints (Phase 2 applies them).

Philosophy: Exploration Mode

Mindset: "What interesting mechanisms emerge from these never-combined vectors?" / "What would a founder do with unlimited resources?" — NOT "can this founder do this?" The goal is to discover unconventional approaches Phase 2 can refine or preserve. Write all 4-5 tactics in FULL here, before any adaptation.

Scope Enforcement — DEMAND GENERATION ONLY

Before generating ANY tactic: "Does this bring NEW PEOPLE to the site who've never heard of the product?" YES = include; NO = delete (CRO territory).

Engineering-as-marketing: allowed but scoped. Building things to ATTRACT new users is valid demand gen, but the tactic's PRIMARY purpose must be acquisition, not product improvement. Test: if you removed the distribution/acquisition channel, would the remaining work be on the product roadmap anyway? If YES → not demand gen.

Tactic type	Verdict
Build-to-attract (80% distribution / 20% build)	✅ Demand gen
Product improvement with marketing framing	❌ Product dev
Improve onboarding / re-engage churned / upsell existing	❌ CRO / Retention

Process (Phase 1)

1. Load diverse white space from synthesis-constraints.json (diverse_white_space). These are pre-filtered to exclude content-flywheel-type vectors and limit repetition (max 2 per anchor). Display the loaded pairs in a table. Category awareness: aim for ≥2 different category prefixes across your 4-5 tactics (soft target here; Phase 2 applies the per-prefix minimums from category_diversity_requirements, which are built to be satisfiable within 7-9 tactics of 2-3 vectors — if a minimum is still unreachable, say so explicitly in the disposition rather than silently missing it).
2. For each combination, derive the emergent mechanism (reuse/extend the matching synthesis-explore.md mechanism where one exists): "What UNUSUAL capability emerges?" It must differ from what either vector produces alone and feel uncomfortable/counterintuitive. Do NOT consider whether the founder can execute it.
3. Design a complete tactic from each mechanism. Name it per the naming rule in ${CLAUDE_PLUGIN_ROOT}/reference/writing-style.md: plain words that tell a smart friend what you'd actually do — what you do + the twist — with the unconventional mechanism visible in the name (not generic marketing, and not an invented codename either). No practicality filtering — include tactics even if they need skills/budget the founder lacks or feel risky.

Kill List (still applies in exploration mode)

Skip combinations whose core action is: start a newsletter · write SEO content · build community (join Discord/Slack) · demo the product · interview customers · improve the core product · re-engage churned users · upgrade existing customers · improve onboarding · build a tool/dashboard/calculator (engineering, distribution as afterthought) · create a feature and announce it · develop an integration. If the white-space pair only produces a kill-list action, it didn't yield anything novel — skip it.

Self-Audit (per Phase-1 tactic)

1. Remove all vector references — does the tactic still sound unusual? NO → vectors are labels on a conventional action (regenerate from a different pair).
2. Day-1 action on the Kill List? YES → kill & regenerate.
3. Marketer Test — would a B2B marketer recommend this unprompted? YES → too conventional, reconsider.
4. Cannibalization Test — does it teach customers to DIY what the product does? YES → reframe (emphasize the PAIN of manual methods, don't teach the method).

Phase-1 gate: 4-5 white-space tactics, each from a diverse_white_space pair (IDs exist in growth-factors.json), each passing all 4 self-audit tests, 100% unconventional target, demand-gen only. Carry a one-line Practicality Assessment per tactic (Difficulty / Resource intensity / Risk / Phase-2 action: Preserve / Adapt / Consider dropping) so Phase 2 can act.

═══════════════════════════════════════════════════════════════════════════

PHASE 2 — ADAPT & GENERATE (founder leverage; ≥60% of Phase 1 PRESERVED)

═══════════════════════════════════════════════════════════════════════════

Philosophy: Preserve the Unconventional Core (CRITICAL)

The #1 goal of Phase 2 is to PRESERVE Phase-1's unconventional mechanisms. What makes a tactic unconventional: the positioning (contrarian / anti-establishment), the structural advantage (competitor lockout, platform arbitrage), and the uncomfortable edge that makes traditional marketers hesitate.

When adapting for constraints: DO reduce budget/time, simplify execution, scale down scope — while keeping the uncomfortable positioning, the structural advantage, and the contrarian angle. DON'T soften the move itself to make it "safer," replace a structural advantage with an execution advantage, sand down the edge that creates the lockout, or drop a tactic because the founder hasn't done it before — unfamiliarity is not a constraint. The unconventional POSITIONING must survive adaptation — and per the naming rule (Output language below), it survives in PLAIN WORDS: the uncomfortable move stays visible in the name; the edge lives in the mechanism, never in invented vocabulary.

• BAD: "Call Out Enterprise Tools by Name in Your Docs" → "Technical Citation Authority" ❌ (the confrontation vanished → lost the lockout)
• BAD: "Anti-Enterprise Citation Rebellion" ❌ (edge hidden behind a codename — a smart friend can't tell what you'd do)
• GOOD: "Call Out Enterprise Tools by Name in Your Docs (Lean Version)" ✅ (the uncomfortable move is right in the name; only budget/time reduced)

Aim for ≥60% Phase-1 preservation with unconventional edges intact.

Phase 2a: Review Phase-1 Tactics (PRESERVE / ADAPT / DROP)

Decision	Criteria	Action
PRESERVE	founder can execute with minor adaptation	keep, small resource adjustments
ADAPT	mechanism valuable, execution needs rework	redesign execution, keep mechanism
DROP	requires more time or money than the founder has	remove (counts toward the ≤40% drop allowance)

A skill the founder would need to LEARN is NOT a reason to drop. State the learning cost honestly (time to competence, quality of first attempts) and keep the tactic. Founders will learn any channel that brings customers — unfamiliarity is not a constraint.

Timeline gate: if a structural-advantage timeline (6-24 mo) exceeds the founder's deadline by >3x AND the founder has a ≤90-day deadline → ADAPT so the tactic produces observable acquisition signal (traffic/signups/trials) within the deadline, even while the moat builds over time. If it fundamentally can't signal within 3x the deadline → DROP.

Adaptation rules: budget too high → reduce scope, not approach; skill missing → state the learning cost honestly (time to competence, quality of first attempts) and KEEP the tactic — a founder will learn any channel that brings customers; time-intensive → phase it; positioning feels risky → KEEP IT (the risk IS the advantage).

DROP threshold: at most 2 Phase-1 tactics dropped (preserve ≥60%).

Channel-justification rule. For each tactic, the Primary Channel choice must be justified against the mechanism — explain WHY this channel amplifies THIS mechanism (not just "the founder can do content"). If the mechanism would work equally on any channel, the channel choice is arbitrary — pick the one with the lowest competitive adoption from the matrix. The channel shapes the execution steps; it is a design input, not a label.

Deception veto (NEVER allowed — binary, non-scored)

Reject any tactic relying on: fabricated scarcity/urgency · multi-account astroturfing / fake social proof · fake-door pages for nonexistent features · manipulating AI search via seeded fabricated discussions · consensus loops of fake accounts · any tactic whose core mechanism is deception (even if effective). Also reject: targeting existing users for upsells (retention, not acquisition). Journalist test: "If a journalist investigated how we execute this, would it be embarrassing?" YES → reframe or DROP. Cross-check every tactic against synthesis-constraints.json anti_patterns (the clean-room replacement for the proprietary anti-vectors DB) — any HIGH-severity match is blocking.

Phase 2b: Generate Synergy & Founder-Leverage Tactics

From synthesis-constraints.json: Pool B (synergy pairs) + Pool C (founder-leverage). You MUST generate at least 4 new tactics here so the final set reaches 7-9 (the founder gets a portfolio of ideas; this is where free stops, so depth-of-options matters). If you can't find 4 quality combinations, explain why, still produce your best 4 with caveats, and flag low-confidence ones [LOW CONFIDENCE].

must_include enforcement (Pool B). Every Pool-B must_include pair must be validly used or validly substituted across the FINAL tactic set. Validly used = both of the pair's vector IDs appear together within a single tactic's **Source:** … Vectors list (block-level co-occurrence — not the two IDs scattered across different tactics). Validly substituted = an explicit one-line note that names the pair AND names a replacement vector that exists in growth-factors.json (e.g. "B3 lever-NNN+struct-NNN substituted — neither vector survived the founder-fit check; replaced with resource-NNN which is in growth-factors.json"). A silently dropped Pool-B pair, or a substitution whose replacement isn't in growth-factors.json, is a defect — list each Pool-B pair and mark it used (which tactic #) or validly substituted (which replacement). This is in addition to the "≥4 new tactics" floor, not a replacement for it.

These MUST pass founder-constraint checks: budget in range; time realistic; leverages an asset the founder possesses (or states honestly what they'd need to learn). Apply the Complexity Budget (inlined — the paid pipeline kept this in a monolith):

Team capacity	Max vectors/tactic	Max monthly $ on external help
Solo, <10 hrs/wk	2	$0-100
Solo, 10-25 hrs/wk	2-3	$100-300
Small team	3	$300+

Tactics exceeding budget are flagged [HIGH COMPLEXITY]. Apply the Kill List (inlined from the tactic-design step): start a newsletter · write SEO content · build community · demo the product · interview customers · improve the core product · re-engage churned · upgrade existing · improve onboarding · build a tool/dashboard · create-and-announce a feature · develop an integration. A Day-1 action matching these = not novel demand gen.

═══════════════════════════════════════════════════════════════════════════

PHASE 3 — MERGE, DEDUP, AND THE FINAL MIX

═══════════════════════════════════════════════════════════════════════════

Combine preserved/adapted Phase-1 + new Phase-2 tactics. Dedup: if two tactics share the same Day-1 action, keep the one with the stronger structural advantage.

Target final mix (ENFORCED): 3-4 from Phase 1 (white space) + 4-5 from Phase 2 (synergy/founder-fit) = 7-9 total (minimum 7); unconventional ≥50%. Respect category_diversity_requirements (no single category prefix > 60% of vectors used).

Phase 3.5: Demand-Gen Purity Pre-Check (BLOCKING)

For EACH final tactic answer: (1) Day-1 action in one sentence; (2) time split BUILDING vs DISTRIBUTING — if BUILDING > 40%, it's product dev, not demand gen; (3) target = NEW people (demand gen) or EXISTING users (retention/CRO). Any tactic failing any question → replace. Max purity failures in final output: 0.

Phase 3.6: Competitor Self-Harm Check

For each tactic vs the top competitors in competitors-analysis.md: does it require a competitor's platform? send traffic to a competitor without capture? educate customers in a way that benefits competitors equally? → reframe or remove.

═══════════════════════════════════════════════════════════════════════════

Output Format

═══════════════════════════════════════════════════════════════════════════

Output language (REQUIRED)

synthesis.md is the deliverable the founder actually reads. Its body copy follows ${CLAUDE_PLUGIN_ROOT}/reference/writing-style.md: plain English at grade 6–8, ≤~20 words per sentence, the banned-jargon table respected, vector IDs ONLY in the **Source:**/Traceability lines (never in prose), tactic names that pass the smart-friend test, and an **In plain English:** line under every tactic name. The template's section headings and field labels stay exactly as written below.

Save to WS/03-think-tanks/demand-generation/synthesis.md. In the per-tactic origin tags, Pass 1 White Space = a Phase-1 tactic that was preserved/adapted; Pass 2 Synergy/Founder-Fit = a Phase-2 generated tactic.

# Demand Generation Tactics Synthesis
## Synthesis Overview
- Method: Two-Pass Vector Synthesis over a per-run LIGHT growth-factors DB · Total Tactics: N
- Pass summary table (Pass 1 white space → N; Pass 2 synergy/founder-fit → N)
- White Space Retention Rate: X% (target ≥60%)
- Unconventional Ratio: UNCONVENTIONAL / HYBRID / CONVENTIONAL (target unconv ≥50%, conv ≤30%)
- Constraint Summary (budget / team / skills / stage / goal)
- Channel Selection Summary (Saturated to AVOID / Moderate / Open Opportunity — from the competitive channel matrix)

## Pass 1 Disposition
| Pass 1 Tactic | PRESERVE/ADAPT/DROP | Reason |

## Generated Tactics
### Tactic #N: <Name> — [Pass 1 White Space | Pass 2 Synergy/Founder-Fit]
**In plain English:** <one ≤20-word sentence — what you actually do>
**Source:** Origin · Vectors `id-1` + `id-2` (exist in growth-factors.json)
**Primary Channel(s):** from the channel menu · Competitive adoption (0-2 opportunity / 3-5 moderate / 6+ saturated) · why this channel
**Emergent Mechanism:** what emerges from the combination
**Structural Advantage:** Type / Duration / Why competitors can't copy
**Synthesis Logic:** what each vector contributes + how they interact (2-3 sentences)
**Adapted to Your Constraints:** (if ADAPTED) what changed; how it fits the founder
**How to Execute:** 1. … 2. … 3. … (specific tools/platforms, adapted for constraints)
**Required Resources:** Time / Budget (in range) / Tools
**Skills Required:** table — skill / founder has? / status
**Expected Timeline & Success Metrics:** early signal (Wk 1-2) / key metric (Month 1) / time-to-signal / decision point
 checked vs synthesis-constraints.json anti

Core Action Strip Test (for the unconventional ratio): write each tactic's Day-1 action in one sentence, remove ALL adjectives, ask "would a traditional marketing consultant recommend this exact action?" YES → CONVENTIONAL; NO → UNCONVENTIONAL. Target ≥50% unconventional; <40% = FAIL (replace conventional tactics).

Validation Checklist (final)

• 7-9 tactics (minimum 7); 3-4 from Pass 1 + 4-5 from Pass 2.
• ## Pass 1 Disposition present (the structural proof that Phase-1 ideation preceded Phase-2 adaptation) — every Phase-1 tactic marked PRESERVE/ADAPT/DROP with a reason.
• White-space retention ≥60%; unconventional ≥50% (Core Action Strip Test).
• Phase 3.5 purity: 0 failures (every tactic brings NEW people; building ≤40%).
• Deception veto: 0 tactics fail the journalist test / match a HIGH anti_pattern.
• Category diversity respected (no prefix > 60% of vectors used).
• Every vector ID exists in growth-factors.json; every tactic traces to a combination.
• Every Pool-B must_include pair is validly used (both IDs co-located in one tactic) or validly substituted (replacement named and present in growth-factors.json).
• Every tactic carries an **In plain English:** line (≤20 words, plain words); every tactic name passes the smart-friend test; body copy follows reference/writing-style.md (no banned jargon, no vector IDs in prose).
• ## Post-Synthesis Self-Review present with retention %, unconventional %, constraint compliance, no-duplicates, Final Grade.
• STOP here — no prioritization, no implementation guides (paid).