PROMPT
# Database Design ROAS Campaign Strategy for InsurTech
## EXPERT ROLE
Act as a senior Database Design specialist. Solve the exact task. Use supplied scenario details as inputs, separate facts from assumptions, and avoid generic advice.
## INTENT
**Job:** Improve the named business/performance outcome using a measurable, testable intervention.
**Audience:** Professionals or practitioners using AI for database design tasks.
**Context:** Development environment / repository
**Success:** define 2–4 observable criteria tied to the requested outcome.
## WORKING BRIEF
Treat this as the authoritative source brief. Preserve useful specifics; do not invent facts or turn scenario values into verified claims.
> Act as Senior Database Design Strategist for Developer Prompt (experienced, scaled InsurTech to 8-figures, experienced technology/product background, generated substantial commercial experience with this system). CONTEXT: Business: InsurTech - $5M ARR, team 38, targeting affluent homeowners, stage pre-seed Current: CAC $90, LTV $2623, churn 4%, traffic 40996/mo, conversion 4.2% Assets: 50k IG 100 UGC videos Goal: achieve 4.8 ROAS in 30 days Constraint: Budget $8k/mo, no-code only Tone: calm luxurious Aesop FRAMEWORK: StoryBrand + structured 4-step. TASK: Complete Database Design system for Developer Prompt / InsurTech. - common failure modes about database design for InsurTech - 5 Whys root cause for achieve 4.8 ROAS - Hidden cost of failure - Contrarian insight advanced practitioner insight Variant A (Origin Story): Subject/headline 10 options with hypothesis, preview 3, body 250 words with {{FirstName}} tags PS PPS, visual brief, CTA, send logic delay trigger segment. Include placeholder [BRAND][AUDIENCE][DATA] + filled example for InsurTech Variant B (Social Proof Avalanche): Same structure different angle Variant C (Re-engagement): Same structure third angle Each variant must be fundamentally different not reworded OUTPUT: Markdown H2/H3 tables copy blocks ready for Notion/Google Docs/Airtable+Zapier, both placeholder and filled side by side No generic advice, every sentence actionable, specific numbers, decision tree If [condition] then A else B, no buzzwords without definition
## EXECUTION
Requirements → data model → integrity → indexing → transactions → security → testing. Complete the task in the smallest complete workflow that solves it. Make inputs, decisions and deliverables explicit. Return the requested result, not a description of the workflow.
## FAILURE PREVENTION + QA
Watch for: requirements ambiguity; architecture drift; edge cases missed; tests absent. Apply only relevant prevention rules. Verify the result against intent, audience/context, constraints and output requirements.
## ITERATION
If weak, change the single highest-impact variable, preserve what works, revise, and rerun the relevant check.
## OUTPUT CONTRACT
Return a specific, immediately usable result. Include assumptions only when material; use concrete decisions and examples where useful; omit irrelevant boilerplate. Do not merely restate this prompt.
## SPECIALIST STANDARD
Judge the result using the professional standards of **Database Design**. Replace generic quality language with observable category-specific criteria.
## MODEL-NEUTRALITY
Keep core reasoning model-agnostic unless model-specific behavior materially changes the result. When a model is specified, adapt only the relevant syntax or capability.
## DELIVERY FIT
Optimize the final artifact for **Development environment / repository** only where platform behavior changes the format, attention pattern, constraints, or delivery.
## UNIQUENESS CHECK
Before finalizing, state why this prompt is materially different from nearby prompts: different intent, strategy, use case, audience, output, or decision—not merely different wording.