Confirm the match.
The service looks for one operating Google listing whose website hostname matches the submitted website. A missing or ambiguous match needs resolution before scoring.
THE EVIDENCE BEHIND YOUR REPORT
How your
assessment works.
Understand what was checked before deciding what to change.
Your report combines public listing information, homepage checks and robots policies into a dated assessment. It gives you specific findings to review with your own knowledge of the business.
Rubric 3.0.0. A snapshot of public information. Rankings, traffic and revenue are not measured.
01 / WHAT GOES INTO IT
What the assessment checks.
Read the listed homepage.
The scanner reads the homepage HTML and links it contains. It also reads the relevant robots.txt policies. It does not browse every page or interact with forms.
Keep the result.
The report records its time, rubric version, score, findings and suggested actions. Its private web link and PDF refer to the same stored assessment.
If the requested website cannot be assessed, this service does not generate a score. A blocked scanner is not proof that customers cannot use the website.
02 / HOW THE SCORE WORKS
How the score is calculated.
Each factor scores from 0 to 100. The score is their weighted total, minus 12 points for each failed basic requirement, with a minimum of zero. It is rounded to one decimal place.
| Factor | Weight |
|---|---|
| Reputation | 30% |
| Findability | 25% |
| Content | 20% |
| Trust | 10% |
| AI visibility | 15% |
Reputation: Google rating and review volume, adjusted to reduce the effect of very small review samples.
Findability: structured business identity, phone consistency, city in the title and service-area links.
Content: homepage link count, visible year text, title length, description length and one main heading.
Trust: stated credentials or guarantees and links to social profiles. The scanner does not verify those credentials.
AI visibility: machine-readable identity, services, hours, question-shaped content, a description and robots policies. This measures technical signals, not whether an assistant recommends you.
The 17 website/AI checks listed in the report are not its entire score calculation. Reputation is separate, and basic-requirement deductions also affect the total.
03 / READ THE FINDINGS IN CONTEXT
A detection is a starting point.
What a check actually sees
- “Site depth” counts distinct homepage links, with a threshold of 30. It does not count every page on the site.
- “Freshness” looks for recent year text. A current copyright year does not prove the content is current.
- Mobile readiness checks for a viewport tag. It does not test the whole layout on a phone.
- A booking check finds relevant wording. It does not confirm a form, calendar or phone call works.
- Services, questions and trust checks use visible text and markup patterns. Manual review may find content the scanner missed.
What the report does not establish
- Search position, actual AI citations, traffic, competitor position or revenue.
- A full SEO audit, accessibility audit, security audit or whole-site crawl.
- The quality of the business’s work, or whether every suggested change is worth paying for.
- A guarantee of higher rankings, more customers or a return on spending.
Raw Google rating/review inputs and a complete ledger of basic-requirement verdicts are not retained in the current report response. The published formula explains the scoring, but the report alone cannot reproduce every original input.
04 / PUT IT TO WORK
Check findings before choosing work.
Suggested actions follow an editorial order based on the observed findings. Related identity checks are grouped together. Their position does not predict ranking or revenue gains.
Check a finding against the live site and what you know about the business. You can address it yourself, discuss it with your existing team or agency, or have the team do it. Paid help is priced on the pricing page, and scope is agreed before payment.
CORRECTIONS
Something does not look right.
Public information changes, and automated checks can miss things. Write to team@mainstreetscoreboard.com with the finding and the correct information. Factual corrections are free and never depend on buying anything.
Who makes this and how to request a correctionTHE TECHNICAL REFERENCE
Scoring formulas and thresholds.
Factor checks, deductions and the AI ceiling
Findability, content and trust score the percentage of their own checks that pass: 4, 5 and 2 checks respectively. AI visibility uses 6 checks. Where the primary listed host disallows a checked token, its AI score is capped at 25; this is a ceiling, not an automatic score of 25.
The checked crawler tokens are GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot. Other policy tokens checked are Google-Extended and Applebot-Extended. These tokens have different purposes. A policy restriction is not proof of exclusion from search or every AI answer. Review the owner’s intent before changing any permission.
Basic requirements are HTTPS, a mobile viewport, tap-to-call and booking/quote wording. Each failure costs 12 points; a pass adds no points. The requested-assessment service stops if the website cannot be assessed.
Final score = round(max(0, sum(factor score × factor weight) − deductions), 1).
The reputation formula and its constants
Adjusted rating = (20 × 4.2 + rating × reviews) / (20 + reviews).
Quality = clamp((adjusted rating − 3.5) / 1.5, 0, 1). Volume = min(1, ln(1 + reviews) / ln(1 + 500)).
Reputation = round(100 × (0.70 × quality + 0.30 × volume), 1). No rating or no reviews produces zero. The volume reference is fixed; adding a competitor does not change this business’s score.
Signals that count in more than one factor
LocalBusiness markup is read in both findability and AI identity. Before the AI ceiling or final zero floor constrains the result, a missing value costs 6.25 + 2.5 = 8.75 total points.
A missing meta description affects content and AI description checks: 4 + 2.5 = 6.5 points under the same qualification. An existing description outside the 50-165 character band fails content length only. Factor rounding can affect displayed differences.