◯ LEARNING MODEL

ALL PROJECTS · HYPOTHESIS REGISTRY
FILTER: ALL LEAD GENERATION HUDSON VOICE RLOOKUP
Assumptions blind. Hypotheses guide. Every unknown is tracked here until proven or disproven by data.
30
Total
0
Confirmed
0
In Progress
30
Untested
CONVERSION 4
h-001 UNTESTED
What is the close rate on warm leads from public records triggers?
Manually work 20+ public records leads. Log every contact attempt and outcome. Calculate: reached / quoted / bound rates.
Sample progress0/20
Grok flagged this as the #1 thing to prove before spending any ad budget.
h-015 UNTESTED
What is the optimal follow-up timing after initial contact — same day, 24h, 48h, 72h?
Log timestamp of every contact attempt and response. After 50 leads, analyze which follow-up interval produces the highest response rate.
Sample progress0/50
Industry data suggests contact rate drops 80% after 5 minutes for initial contact. Follow-up timing is a separate question.
h-024 UNTESTED
What is the cross-sell rate from homeowners to auto on the same household within 90 days of binding the home policy?
Track every bound homeowners policy. Log whether same household auto was quoted and bound within 90 days. Calculate cross-sell rate. Note the conversation that triggered it.
Sample progress0/20
Home-to-auto bundle is the most natural cross-sell in P&C. Hypothesis: offering the bundle quote at bind increases auto attachment rate significantly vs. waiting 90 days.
h-025 UNTESTED
At what point in the relationship is the cross-sell conversation most effective — at first quote, at bind, or at 30-day follow-up?
Split bound clients into 3 groups. Group A: cross-sell offered at bind. Group B: cross-sell offered at 30-day follow-up call. Group C: cross-sell offered only when client calls in. Track cross-sell conversion rate per group.
Sample progress0/30
Timing of cross-sell is likely as important as the offer itself. Too early = feels like a pitch. Too late = they already went elsewhere.
PRODUCT 5
h-002 UNTESTED
Which product line (auto) has the fastest path from warm lead to bound policy?
Ask User based on existing book experience. Then validate against first 30 days of pipeline outcomes.
Sample progress0/10
Determines first campaign product line. Cannot start ads without this answer.
h-019 UNTESTED
What is the best lead trigger for auto insurance — new vehicle registration, license reinstatement, or young driver (18-24 aging onto own policy)?
Run three separate ad sets targeting each trigger segment. Track CPL and close rate per segment over 30 days. Florida DMV new registration data available via public records.
Sample progress0/30
Auto closes fastest of all product lines. Need to identify which entry trigger produces the warmest lead, not just the cheapest.
h-021 UNTESTED
Which life insurance trigger converts best — new homeowner (mortgage protection), new parent (birth record), or newly married (marriage record)?
Pull marriage records and birth records from Miami-Dade Clerk. Compare close rate for life leads from each trigger event against new homeowner life cross-sell. Minimum 15 per segment.
Sample progress0/45
Mortgage protection is the easiest conversation opener for life — the need is concrete and quantifiable. But birth records may produce higher urgency. Only data will tell.
h-022 UNTESTED
Does ACA open enrollment window (Nov–Jan) produce meaningfully higher health insurance lead volume than year-round outreach?
Run health insurance campaigns in two windows: open enrollment (Nov–Jan) and a comparable non-enrollment period. Track CPL, response rate, and bind rate. Note: SEP (Special Enrollment Period) leads exist year-round for qualifying events.
Sample progress0/50
ACA open enrollment is high competition. SEP leads (job loss, marriage, new baby, turning 26) are year-round and less contested. Hypothesis: SEP leads have better CPL outside enrollment window.
h-029 UNTESTED
Is home insurance the right product to lead with, or does auto have a shorter path to first bound policy given current FL carrier appointment constraints?
Requires User's input: which product line can he bind fastest and most reliably in current FL market? Compare time-to-bind for auto vs. home leads in first 30 days of manual test.
FL homeowners market is severely constrained — admitted carriers are limited, rates are high, Citizens non-renewals are creating a hard market. Auto is simpler, faster, and less carrier-dependent. Home may not be the right lead product even though it has the strongest public records trigger. This is an existential question for month one.
PERSONALIZATION 3
h-003 UNTESTED
Does referencing a specific property address in an SMS increase or decrease response rate?
Split 50 leads: 25 receive address-specific SMS, 25 receive generic SMS. Compare response rate and opt-out rate.
Sample progress0/50
Some markets find property-specific messaging impressive. Others find it intrusive. Only data will tell.
h-008 UNTESTED
Does Cuban Spanish outperform general Spanish in 33125/33126/33135 zip codes?
Run two Spanish ad sets: one using Cuban Spanish cultural references and register, one using neutral Spanish. Compare CTR and lead quality over 30 days.
Sample progress0/100
Requires Spanish copywriter or User's review to validate cultural accuracy.
h-023 UNTESTED
Does targeting self-employed and gig workers outperform general demographic targeting for health insurance leads in Miami-Dade?
Run two Meta ad sets: one targeting self-employed/freelancer/gig worker interests, one targeting broad 25-54 demographic. Compare CPL and quote rate. Self-employed = no employer coverage = active need.
Sample progress0/100
Miami has a large gig economy and self-employed population, especially in the Hispanic community. These people have no HR department offering coverage — they're actively uninsured or underinsured.
MARKET 1
h-004 UNTESTED
Is the Citizens Insurance non-renewal angle already saturated in South Florida?
Check Meta Ad Library for active insurance ads using Citizens non-renewal messaging. Check Google Keyword Planner for CPC trends on 'Citizens Insurance alternative'. Assess saturation level.
Claude synthesis flagged this as a blind spot. Every marketer in FL knows about depopulation.
OPERATIONS 2
h-005 UNTESTED
What carriers can User currently bind with for homeowners in the Florida admitted market?
User provides list of active carrier appointments. Cross-reference against which products public records leads most need.
Existential constraint. If carrier appointments are thin, lead quality is irrelevant — can't bind the policy.
h-016 UNTESTED
Does the agency write non-standard auto (high-risk drivers, no prior coverage)?
User confirms yes/no. Determines whether affordability segment is viable at lower end.
Determines scope of affordability segment campaign.
FINANCIALS 3
h-006 UNTESTED
What are the actual commission rates by product line from User's carrier contracts?
User provides commission schedule from each carrier for auto, home, life, health.
Every ROI model built so far is a hypothesis. Real numbers required before any spend target is set.
h-007 UNTESTED
What is the average first-year premium for a typical policy in the target zip codes?
User provides average from existing book by product line.
Varies significantly by zip code, property type, and risk profile. Cannot assume from industry averages.
h-011 UNTESTED
What is the actual cost per postcard via Lob or PostGrid at current rates?
Create account on Lob.com and PostGrid. Get actual per-piece pricing for 4x6 postcard with variable data printing. Compare.
Budget allocation for direct mail cannot be finalized without real pricing.
TECHNOLOGY 4
h-009 UNTESTED
Will Meta flag or restrict a new 2-agent insurance ad account with no prior ad history?
Launch Meta Business Manager under DBA. Run first campaign. Monitor for account restrictions or reduced delivery within first 7 days.
Grok flagged this. Insurance is a restricted category. New accounts with no history get scrutinized.
h-012 UNTESTED
What is the minimum audio recording needed for a quality ElevenLabs voice clone?
Review ElevenLabs documentation. Test with a 5-minute recording vs. 15-minute recording. Evaluate voice match quality.
Required before asking User to record. Don't waste his time if requirements are unclear.
h-013 UNTESTED
Do the Miami-Dade Socrata endpoint IDs in caganut_pipeline_v1.py match live production data?
Visit data.miamidade.gov. Search for deed transfer and building permit datasets. Verify exact endpoint IDs. Update script if different.
Flagged in pipeline script comments. Must verify before first production run.
h-017 UNTESTED
What is the correct API model ID for GPT-5.5 in the OpenClaw model tier?
Check OpenAI API documentation for current model IDs. Test the correct identifier in job_models.py.
Returned invalid model ID error on first review run. GPT-5.5 not available or wrong identifier used.
COMPLIANCE 1
h-014 UNTESTED
Does Florida insurance advertising law restrict property-specific references in unsolicited outreach?
Legal review: Florida Statutes §626 + Florida DOI advertising guidelines. Specific question: can we reference property address or purchase price in outbound SMS and direct mail?
High priority before sending any personalized outreach. TCPA + state law intersection.
CHANNEL 7
h-010 UNTESTED
What is the response rate for triggered direct mail (deed transfer) vs. generic insurance mailers?
Send 50 personalized triggered postcards. Track QR scan rate and callback rate. Compare to industry benchmarks for generic insurance mail (approx 1-3%).
Sample progress0/50
Industry benchmark for generic insurance mail is 1-3% response. Triggered mail hypothesis: significantly higher. Must be validated.
h-018 UNTESTED
Does a referral program for existing clients generate leads that convert at a higher rate than paid leads?
Design a simple referral incentive. Track referred leads vs. paid leads over 60 days. Compare contact rate, quote rate, bind rate.
Sample progress0/20
Hypothesis: word-of-mouth referrals reach underinsured community members who are invisible to digital targeting.
h-020 UNTESTED
Do car dealership partnerships generate auto leads that close at a higher rate than cold digital leads?
Approach 3 independent car dealerships in target zip codes with a referral arrangement. Track leads sourced via dealership vs. digital over 60 days. Compare close rate and time-to-bind.
Sample progress0/20
Dealerships need the buyer insured before they drive off the lot. Timing is perfect. Hypothesis: close rate is significantly higher because the trigger is immediate and unavoidable.
h-026 UNTESTED
Do real estate agent referral partnerships generate leads that close at a higher rate than public records cold outreach?
Establish referral arrangements with 3-5 independent real estate agents in target zip codes. Track leads sourced from agent referrals vs. public records. Compare close rate, time-to-bind, and average premium.
Sample progress0/20
Real estate agents need their buyers insured to close. A warm referral from a trusted agent carries implicit credibility. Hypothesis: close rate 2x+ vs. cold public records outreach.
h-027 UNTESTED
Do mortgage broker partnerships generate higher-quality home insurance leads than public records deed transfer triggers?
Establish referral arrangements with 3-5 mortgage brokers. Track lead quality vs. deed transfer pipeline leads. Mortgage broker lead = pre-closing, higher intent, confirmed financing. Deed transfer lead = post-closing, may already have coverage.
Sample progress0/20
Mortgage broker referral hits the buyer before closing — they need insurance to close the loan. This is earlier in the funnel than deed transfer and may produce higher urgency.
h-028 UNTESTED
Does in-person presence at community events (health fairs, church events, neighborhood associations) in target zip codes produce bindable leads?
Attend 3-5 community events in 33125/33126/33135. Collect contact info with consent. Track contact rate, quote rate, bind rate vs. digital channels. Log which event type produced the best leads.
Sample progress0/30
For health and life insurance in underserved Hispanic communities, in-person trust-building is potentially the most effective channel. People who distrust digital ads or don't respond to mail still respond to a face they recognize.
h-030 UNTESTED
Do notary publics or other trusted community figures in Hispanic neighborhoods generate insurance referrals when approached with a partnership arrangement?
Identify 3-5 notary publics or community connectors in 33125/33126/33135. Propose a simple referral arrangement. Track referral volume and lead quality over 60 days.
Sample progress0/10
In Latin American culture, the notary public is a trusted legal and financial figure. Notaries in Miami often serve as community hubs — people come to them for document notarization, translations, and advice. This is an untapped referral channel that competitors are not working.