You can inspect the house. You can't inspect the block.
Our Neighbors pays verified neighbors to describe daily life on a street — noise, traffic, parking, upkeep — then hands buyers and agents an anonymized report, aggregated across at least three voices so no single person is identifiable. Matchmaking, not surveillance.
We have deliberately written no product code yet. Phase 1 validates demand with a manual pilot first — a failed pilot saves a 12-week build.
The Market Failure
Homebuyers spend 30-45 days vetting the house — inspection, appraisal, comps — but almost no structured insight into what daily life on the block is like. Industry practitioners estimate a bad neighborhood mismatch can cost families $50K-$500K in quality-of-life loss or forced relocations (an estimate the pilot is designed to pressure-test, not a cited figure).
Neighbors know the day-to-day realities of a street — noise, traffic, parking, upkeep, community rhythm — but have no structured way to convey that context to someone considering a move. Agents have no credible way to share it either.
No trusted marketplace exists to fill the gap. If 5-20% of the ~6.5M annual US home sales attach a report at $49-$300 per side, that is a $500M-$2B opportunity — but the attach rate is unproven, and validating it is exactly what the pilot exists to do.
Why This Model Works
✓Matchmaking, not data extraction — neighbors are paid stakeholders, not the product
✓Payment unlocks participation — free surveys ~20% response, paid ~40-60% (a pilot assumption)
✓Payment improves quality — paid respondents are thoughtful; free respondents give fluff
✓Legal clarity — neighbors own their answers; we transact, not extract
✓Privacy is the moat — anonymous by default, N≥3 before any report, never sold
✓Supply-first density, not "network effects" — demand is per-address and time-boxed, so we pre-seed neighbors block-by-block before marketing to buyers there (see fulfillment risk below)
How It Works — One Report, Money in a Circle
A buyer or agent requests a report on a specific address. Verified neighbors on that block are invited and paid to answer a short, objective questionnaire. Responses are anonymized and aggregated — only once at least three neighbors have answered — and the finished report is delivered. The money the buyer/agent pays funds the neighbor payouts; the platform keeps the margin.
1
Request
An agent or buyer requests a report for one listing and pays up front.
2
Neighbors invited & paid
Opted-in neighbors on that block answer 6-8 objective questions for ~$8 each.
3
Anonymized aggregation
Responses are pooled & de-identified. No report is built below N≥3.
4
Report delivered
Buyer/agent receives an aggregated report in 24-48h. No names, ever.
Agents & buyers pay in → neighbors get paid → aggregated report goes out. The platform never sells the underlying data.
The same loop, from each side's chair
STEP 1
Add a listing to your dashboard.
STEP 2
Request a neighborhood report ($150-$300) for that address.
STEP 3
Share a differentiated listing — stronger marketing, fewer surprises post-sale.
STEP 1
Paste the listing you're considering ($49-$99).
STEP 2
We interview verified neighbors on that block, anonymously.
STEP 3
Get an aggregated report in 24-48h (optional moderated Q&A, +$49).
STEP 1
Get invited — opt-in only, never scraped.
STEP 2
Spend ~5 minutes on 6-8 objective questions. Skip any you want.
STEP 3
Get paid (avg $8 gift card). Withdraw your answers anytime.
🏠
Tier 1 · ~67% of revenue
Agents
$150–$300 per listing for a neighborhood report that differentiates a listing. At ~1.6% of a $12.5K commission, it's a rounding error. Hypothesized benefit: faster closings and fewer post-sale disputes — being measured in the pilot, not yet proven.
Go-to-market: Partner with 5-10 local brokerages. Pilot 3 listings at $100, measure time-to-close.
Month 7 target: 10 agents, $12K/month revenue
🔑
Tier 2 · ~33% of revenue
Buyers
$49–$99 on-demand. An aggregated report drawn from a minimum of 3 neighbors — never fewer, so no single voice is identifiable. 24-48h turnaround. Optional live Q&A (+$49) is moderated by our analyst over the aggregated findings — never an unmasked call with an individual neighbor. Under 0.01% of the purchase price.
Go-to-market: Mostly agent-bundled & organic SEO — buyers are one-time, so paid acquisition is a supporting channel, not the engine.
Month 9 target: $5,900/month revenue
👋
Tier 3 · the paid supply side
Neighbors
$5–$15 per response (avg $8). 5-minute async form, 6-8 objective questions. They earn a gift card and the satisfaction of helping newcomers understand what the neighborhood is really like. You stay in control — skip any question category, withdraw your answers anytime.
"Verified neighbor" badge confirms residency to the platform only — never shown to buyers alongside individual answers.
What a Buyer Actually Receives
This is the product. Every attribute is objective and aggregated across neighbors — no names, no units, no "who said what." Move the slider to see the anonymity threshold in action: below three responses, the report is withheld by design.
123 Maple St · Anonymized Neighborhood Report
Based on 7 verified neighbors
🔒 N = 7 · anonymity threshold met
Noise level2 / 5 · quiet
Walkability4 / 5
Parking ease3 / 5
HOA responsiveness4 / 5
Traffic & safety of streets4 / 5
Would recommend the block86%
"The cul-de-sac gets a little loud on trash-pickup mornings, but otherwise it's peaceful."— Verified neighbor · 4 yrs on the street
"Street parking fills up on weekends when people have guests, but weekdays are fine."— Verified neighbor · 2 yrs on the street
🛡️ No individual neighbor is identifiable. Free-text is reviewed before release. Report requires N ≥ 3 responses.
⛔ Report withheld
Fewer than 3 neighbors have responded. Releasing a report now could make a single person's answers identifiable — so we don't show one at all. The buyer is refunded. This is the rule, not a setting.
Neighbors responded:
7
Drop below 3 to see the report lock.
Privacy Is the Product — The Moat
The defensibility of this business is trust. These six guarantees are hard requirements, engineered in — not marketing promises. The first objection to any neighbor-feedback product is "isn't this creepy surveillance?" Here is why it isn't.
🔢
N ≥ 3 required
No report is generated or shown unless at least three neighbors have answered — so no single response can be traced to a person.
🚫
No reverse lookup
Buyers see aggregated themes only — never a name, a unit, or who-said-what. Triangulation is designed out.
✋
Opt-in only
Neighbors are invited, never scraped. No profile exists for anyone who has not consented.
↩️
Right to withdraw
Neighbors can delete their data or skip any question category, at any time.
🔐
Never sold
Data is used only for the single report it was collected for — never brokered or resold, ever. Explicit in the ToS.
🤝
Paid stakeholders, not product
Neighbors are compensated for their time. The design rule: if a feature could identify one neighbor, it does not ship.
Questionnaire Guardrails — Fair-Housing by Design
Fair-housing / steering is a real legal risk, not a footnote — and we treat it as a potential kill condition (see below). The product describes the place, never the people. The questionnaire is built to make protected-class proxies impossible to ask, and is reviewed by counsel before any VA/FL launch.
✓ What neighbors ARE asked (objective attributes)
Noise levels & general quiet
Traffic and parking
Street upkeep & maintenance
HOA / property-management responsiveness
Lighting & safety of common areas
Nearby amenities & walkability
"Would you recommend the block?"
✕ What we NEVER ask (and can't)
Race, color, or national origin
Religion
Sex or family status (kids, singles)
Disability
"What kind of people live here"
Languages spoken, ages, income, politics
Any proxy for a protected class
By design, and reviewed by counsel before VA/FL launch. If counsel finds the questionnaire cannot reliably avoid protected-class proxies, the product does not launch.
Agent ROI Calculator
Agents are the paying customer. Plug in your numbers — a report is a rounding error against a commission.
Your commission$12,500
Report = share of commission1.6%
Annual spend$4,000
A report is 1.6% of your commission — it pays for itself if it helps the sale close even a few days sooner.
A Rounding Error on the Purchase
A $49-$99 report is under 0.01% of a $500K purchase — the missing line item in due diligence.
$500,000 home purchase
▎ $49-$99 neighborhood report (that thin teal sliver)Downside if you misjudge the block: est. $50K-$500K
You already pay for the rest of due diligence — this is the obvious missing one:
Read this honestly. 77% is the steady-state (month 9+) gross margin above COGS only — Year 1 is a ramp, so its blended margin is lower (~73%). These figures do not yet subtract opex: founder comp, legal/compliance, E&O insurance, Stripe + gift-card fees, and blended CAC. Net profitability depends on those and on CAC, which the pilot must validate. These are pre-revenue projections.
Report: Puppeteer PDF generation from HTML template
No AI yet — structured forms + manual aggregation
Phase 3 — Ongoing
Go-to-Market
Agent acquisition: Direct pitch to top 50 brokers (VA + FL)
Buyer inbound: SEO, content, paid ads ($15-$20 CAC)
Expand 1 new neighborhood every 2 weeks once agent traction exists
Phase 3+: AI summarization of neighbor responses
Key Risks & How We De-Risk
High
Hyper-local liquidity / fulfillment
Demand is per-address and time-boxed — a buyer wants one block inside a 24-48h window, and we need N≥3 opted-in neighbors there already, or we refund. There is no general network effect; every new area is a fresh cold start. Mitigation: pre-seed neighbor supply by subdivision before marketing to buyers there; track fill-rate & time-to-fulfill in the pilot.
High
Report credibility / agent adverse selection
~67% of revenue is agents (usually the seller side), who could try to surface only favorable reports. If buyers learn a report is marketing, trust collapses. Mitigation: reports are released unedited to whoever pays, agents cannot cherry-pick, and provenance ("who commissioned this") is stamped on every report.
High
Fair-housing / steering exposure
Potentially existential — an HUD/DOJ action is a company-ender. The questionnaire covers objective place attributes only, never people or protected-class proxies (see guardrails above). Counsel reviews before VA/FL launch. Kill condition: if counsel can't design proxies out, we don't launch.
High
Buyer CAC exceeds LTV (structural)
Buyers transact once every 7-10 years — LTV ≈ one order, so paid buyer acquisition likely loses money. We treat the business as agent-led: agents are the paying customer and buyers arrive mostly bundled or via organic SEO, not paid ads.
High
Low neighbor participation
Pay neighbors. Pilot reveals rate by week 6. If <30%, rethink incentive structure — hire community managers if needed.
High
Agents don't see ROI
Pilot at $100 to prove faster closes / stronger differentiation. Publish a case study. If velocity doesn't materialize, pivot the pitch to listing differentiation. Offer freemium (3 listings free).
Med
Liability from negative feedback
Anonymity by default. Disclaimer on all reports. Legal review. E&O insurance once first agent customers are live.
Pilot Success Metrics
Neighbor Participation
Target40-60%
Minimum acceptable>30%
Agent Willingness
Target3+ yes at $150-$200/listing
Minimum acceptable2 at $150
Buyer Willingness
Target30%+ say $49-$99 is fair
Minimum acceptable20% at $49
Velocity Signal (directional — n=3)
Target15%+ faster close on pilot listings
Minimum acceptableany measurable lift
Report Fill-Rate
Target% of requests fulfillable at N≥3
Why it mattersmeasures real liquidity
Three pilot listings can only produce a directional signal, not statistical proof — if velocity doesn't show, the agent value prop shifts to differentiation. The 8-week pilot exists to de-risk these assumptions cheaply. If the pilot passes its gates, the path to profitability is credible — month 9 cash-flow positive, Year 1 ~$80K gross profit, Year 2 multi-market scale.
Phase 1 Pilot Kit — Runnable Now, No Code
The whole thesis is validate before building. That only works if the pilot is a real, runnable thing — so here it is. Four instruments run the entire 8-week manual pilot with Google Forms, email, and a spreadsheet. The files below are live in the repo at our-neighbors/pilot/ and map directly to the gates above.
📍
Beachhead: The Villages, FL
We run the pilot where the model is easiest to prove and where we already have ground truth — Jim's own market. The Villages offers dense, near-identical HOA subdivisions (one neighbor pool covers many nearby listings, so supply builds block-by-block fast), high turnover with out-of-state buyers who buy sight-mostly-unseen and crave exactly this "what's the block like" signal, and time-rich retiree neighbors who make the paid-participation thesis easiest to test. It's a controlled first market before expanding across Central Florida.
📋
Neighbor Questionnaire
pilot/neighbor-questionnaire.md
The exact 6–8 objective, fair-housing-safe questions a paid, opted-in neighbor answers in ~5 min (noise, traffic, parking, upkeep, HOA, lighting, walkability, recommend). Includes consent copy and the hard "never ask" protected-class rules. Runs as a Google Form.
Feeds Gate 1 — neighbor participation rate.
🏠
Agent One-Pager
pilot/agent-one-pager.md
The leave-behind and talk-track to pitch listing agents at $100/listing (below the $150–$200 target, to prove value). Villages-specific angle, objection handling, and the explicit pilot ask.
Feeds Gate 2 — agent willingness to pay.
🔑
Buyer WTP Survey
pilot/buyer-wtp-survey.md
A short survey to ~20 buyers with a price ladder ($0→$150+) to test whether $49–$99 clears. Asks about value and willingness to pay only — never demographics.
Feeds Gate 3 — buyer willingness to pay.
📊
Results Tracker
pilot/results-tracker.md
The single scoreboard: the 8-week gate table (🟢/🟡/🔴), recruitment/pitch/survey logs, velocity & fill-rate tables, COGS actuals, and a Go/No-Go decision block. A red on any gate stops Phase 2.
Rolls up all five gates + the counsel review into one Go/No-Go.
Recruit 30–50 Villages neighbors (HOA newsletter, Facebook, NextDoor · $10 incentive) → run the questionnaire form.
Pitch 10 agents with the one-pager; take 1–3 pilot listings each at $100.
Survey ~20 buyers on WTP; hand-build PDF reports for fulfilled requests.
Score every result in the tracker weekly; read the gates at week 8.
Is my block covered?
We seed neighbors subdivision-by-subdivision, starting in The Villages. Drop your block and we'll note the interest — no report is generated, nothing is scraped, and this only records that someone asked.
We store your entry only in this browser (localStorage) for the pilot waitlist — no server, no account, no email required. Clear your browser data and it's gone.
✓ Noted — thank you
We've recorded interest for your block. When we seed neighbors in your area during the pilot, this is exactly the demand signal we look for.
Nothing was sent anywhere — your entry lives only in this browser.
What Would Make Us Kill This
A plan you trust names its own exits. Every number on this page is a pre-validation projection. The 8-week manual pilot gates everything — and a red on any of these stops Phase 2 (no code gets written):
Neighbor participation below 30% — the paid-supply thesis fails.
Fewer than 2 agents willing to pay at $150 — no paying customer.
Under 20% of buyers say $49 is fair — pricing doesn't clear.
No measurable velocity or differentiation lift for agents — the pitch is unfounded.
Counsel can't design the questionnaire to avoid protected-class proxies — fair-housing risk is unacceptable.
Fill-rate too low to serve buyers inside their decision window — the hyper-local liquidity problem is unsolved.
Load-bearing assumptions the pilot must validate (not yet facts)
Neighbors respond at 40-60% when paid
vs. ~20% unpaid — untested at our incentive.
Reports make listings close ~15-20% faster
Plausibly zero effect — closings are gated by financing/inspection.
5-20% of home sales attach a report
The entire TAM rests on this. Unproven.
A bad-block mismatch costs $50K-$500K
Practitioner estimate, not a cited figure.
"Isn't This Creepy?" — Straight Answers
The gut reaction to a neighbor-feedback product is suspicion. Good — here are the direct answers.
Is this spying on neighbors?
No. Participation is opt-in and paid. Neighbors are invited, choose every answer, describe the place (not people), and can skip anything. Nothing is scraped and no profile exists for anyone who hasn't consented.
Can a buyer figure out who said what?
No. Reports require at least three responses (N≥3), buyers only ever see aggregated themes — never a name, a unit, or a single quote tied to a person — and free-text is reviewed before release. There is no reverse lookup.
Do you sell this data?
Never. Data is used only for the one report it was collected for, then it's done. No brokering, no resale — it's explicit in the Terms of Service, and it covers site analytics too.
Is this legal — what about fair housing?
The questionnaire asks only about objective neighborhood attributes — noise, traffic, parking, HOA responsiveness, upkeep — never about who lives there or any protected class. Counsel reviews the design before any VA/FL launch, and if proxies can't be engineered out, we don't launch.
Can a neighbor change their mind?
Yes — anytime. Neighbors can decline any question category up front and withdraw their responses later. They're paid stakeholders, not the product.
Doesn't the "live call" break anonymity?
No. The optional +$49 upgrade is a moderated Q&A with our analyst over the aggregated findings — not an unmasked call with an individual neighbor. Nothing that could identify a specific person is exposed.
A concept in active design — not yet live
The house is inspected. The block never is.
The whole thesis is validate-before-build. If your street is in the Villages pilot area, tell us — it's the exact demand signal the pilot is built to read. Stored only in your browser; nothing is sent anywhere.