This page is the technical reference for how the Neur Score™ is produced. It is kept current with the production engine; the date at the bottom is the last time it was checked against the code.
Scoring model
- Five pillars (Market, Competition, Location, Financial, Personal), each scored, then weighted by industry category and summed to a 0–100 overall score.
- Thresholds for market inputs are state-relative where possible, so a city is judged against its own state's medians rather than only national ones.
- A demand signal from the industry's local employment trend feeds the Market pillar.
- Household income is used once, in the Location fit, and is not double-counted in Market.
- The Competition pillar combines a distance-weighted count of competitors within the trade radius, their average rating, the share of unrated listings, and saturation per resident.
- The Financial pillar compares the stated budget with the industry's startup cost planning range (table below).
- The Personal pillar is built from the questionnaire: industry experience, prior ownership, partners, commitment and timeline.
Pillar weights
The Neur Score™ weights the five pillars differently by industry category. Location carries the most weight for walk-in businesses (food and beverage, retail, beauty, hospitality, entertainment) and almost none for online and professional services, where Personal fit and Market demand dominate. Financial readiness is weighted most heavily in capital-intensive categories (construction, automotive, transportation, agriculture, manufacturing). The exact weights and their calibration are proprietary and are re-tuned against completed reports; the report itself shows each pillar's weighted contribution.
Verdict thresholds
80 and above: Strong Opportunity. 65–79: Promising. 50–64: Buildable with adjustments. 35–49: Rethink the plan. Below 35: Not the right time. Every report lists the changes that would raise the score and by how many points.
Startup cost planning ranges (USD)
These are the industry-level ranges the Financial pillar compares against your budget. They are planning figures, not quotes.
| Category | Low | High |
|---|---|---|
| Food & Beverage | 50,000 | 300,000 |
| Retail | 25,000 | 150,000 |
| E-commerce | 2,000 | 25,000 |
| Technology | 5,000 | 50,000 |
| Professional Services | 5,000 | 30,000 |
| Fitness | 30,000 | 200,000 |
| Health & Wellness | 15,000 | 100,000 |
| Childcare | 30,000 | 150,000 |
| Construction | 25,000 | 200,000 |
| Home Services | 10,000 | 50,000 |
| Beauty & Personal Care | 20,000 | 100,000 |
| Education & Tutoring | 5,000 | 50,000 |
| Real Estate | 10,000 | 100,000 |
| Hospitality | 50,000 | 500,000 |
| Entertainment | 20,000 | 200,000 |
| Automotive | 50,000 | 300,000 |
| Finance | 15,000 | 75,000 |
| Transportation | 10,000 | 100,000 |
| Agriculture | 20,000 | 200,000 |
| Manufacturing | 50,000 | 500,000 |
Seed industries carry a more specific three-tier estimate (derived from typical ticket size and volume) that the report shows in its financial section.
Data vintages
- Census ACS: 5-year estimates via the Census API. National reference medians in the engine are from ACS 2022.
- BLS QCEW: latest available quarters, county level.
- BLS CES: latest monthly national series.
- Google Places: fetched at report time; competitor results are cached per business keyword and location.
Known limitations
- ACS 5-year estimates smooth over recent change in fast-growing areas.
- QCEW suppresses some small-county industry cells for confidentiality; where a cell is missing the engine falls back to broader geography and says so.
- Google Places may omit very new businesses or include closed ones.
- Online businesses skip the Competition and Location pillars in favour of a search-market analysis.
- Scores are estimates. See Why Neur for the full list of caveats.
Infrastructure
- Application: Next.js on Vercel. Database and authentication: Supabase. Payments: Stripe. Transactional email: Resend.
- Sign-up and sign-in are protected by Cloudflare Turnstile.
- Public endpoints for machines:
/llms.txt,/llms-full.txt,/docs/neur.json,/sitemap.xml,/robots.txt.
Last checked against production code: 2026-09-07.