How good is the page your ads land on?
Paste the address and get twenty checks back: can a phone tap your number, is the form short enough to finish, can Google and the AI answer engines read the page, and how fast does it load on a phone. Free, nothing stored, and you can share the result.
What makes a good landing page for a local service business?
A tappable phone number and one short form above the fold, the business name, address or service area, and phone in plain text, real reviews near the call to action, a page that loads in under two seconds on a phone, and structured data that tells search and answer engines who you are, where you work, and when you are open. Unbounce's benchmark data across tens of thousands of pages puts the median landing page around 6.6% conversion, and pages with three form fields convert at roughly twice the rate of pages with seven or more.
How do I get my business into Google AI Overviews and AI Mode?
Google says there are no special requirements: a page has to be indexed and eligible for a snippet, and it recommends the same fundamentals as regular search, including structured data that matches the visible text. Blocking Googlebot removes you from all three; the Google-Extended token only affects Gemini training and grounding, not Search or AI Overviews. In practice, the pages that get cited answer a specific question plainly near the top, carry LocalBusiness and FAQ markup, and load fast.
Which AI crawlers should I allow in robots.txt?
Allow the ones that produce citations: Googlebot, Bingbot (Copilot), OAI-SearchBot (ChatGPT search), PerplexityBot, Claude-SearchBot, and Applebot. Whether to allow training-only crawlers such as GPTBot, ClaudeBot, Google-Extended, and CCBot is a separate business decision that does not change whether ChatGPT or Perplexity can cite you. User-triggered fetchers (ChatGPT-User, Claude-User, Perplexity-User) act when a person asks about your page and may not honor robots.txt at all.
Does my site need an llms.txt file?
Google says no: it does not read machine files made for AI. Some AI tools and retrievers do read llms.txt, and it costs nothing, so the grader treats it as an optional point rather than a requirement. Spend the effort on structured data and a clear answer paragraph first.
How is the score calculated?
The score adds up twenty checks in three groups: conversion is worth 40 points, local trust 30, and search and AI readiness 30. Each check has a fixed weight, shown next to it in the report, so the same page always gets the same score. Mobile speed from Google's PageSpeed Insights is shown separately because a speed test changes from one run to the next.
What does the grader store?
Nothing. It fetches the page you enter once, along with robots.txt, sitemap.xml, and llms.txt, scores them, and returns the report to your browser. The address stays in your browser's URL so you can share the report. If you ask for the fix list by email, that request goes to our form processor with your email and the page address.
How the score is put together.
Conversion, 40 points. Tap-to-call link (10), a call to action in the first screen (8), a lead form (6) and its length (6), a mobile viewport tag (5), and HTTPS (5). The form-length thresholds follow Unbounce’s Conversion Benchmark Report, where conversion falls with every field past three.
Local trust, 30 points. LocalBusiness structured data with address and telephone (12), following Google’s required properties, name, address, and phone in visible text (6), Service markup (4), FAQ markup (4), and review signals (4).
Search and AI readiness, 30 points. Title (4), description (3), one H1 (3), canonical (2), indexability (4), social card (2), image alt text (2), answer-engine crawler access in robots.txt (6), sitemap (2), and llms.txt (2, optional). Crawler roles come from the published documentation of OpenAI, Anthropic, Perplexity, and Google: Google-Extended does not affect Search, AI Overviews, or AI Mode, so blocking it costs no points.
Speed. Reported from Google’s PageSpeed Insights (mobile) alongside the score rather than inside it, because a lab run varies with the moment. Google and Deloitte’s study of 37 sites found a 0.1 second improvement lifted retail conversions 8.4% and travel 10.1%, which is why it is on the report at all.
Limits. The grader reads the HTML the server sends; pages that render everything with JavaScript after load will score low on content checks and should be tested with a rendered-HTML tool as well. It reads one page, not a site.
- Google Search Central, AI features and your website (2026): no additional requirements to appear in AI Overviews or AI Mode; use nosnippet, max-snippet, noindex controls; Google-Extended limits training and grounding only. Source ↗
- Google Search Central, LocalBusiness structured data: required name and address; recommended telephone, url, openingHoursSpecification, geo, priceRange. Source ↗
- OpenAI, Overview of OpenAI crawlers: OAI-SearchBot for ChatGPT search, GPTBot for training, ChatGPT-User for user actions. Source ↗
- Anthropic, Does Anthropic crawl data from the web: ClaudeBot (training), Claude-User, Claude-SearchBot. Source ↗
- Perplexity, Perplexity crawlers: PerplexityBot for search results, Perplexity-User for user requests. Source ↗
- Google Search Central, Google-Extended: does not impact inclusion in Google Search nor act as a ranking signal. Source ↗
- Unbounce, Conversion Benchmark Report (Q4 2024, 41,000 pages, 464 million visits): median landing page conversion 6.6%; conversion falls as form fields increase. Source ↗
- Google and Deloitte, Milliseconds Make Millions (2020, 37 sites, 30 million sessions): a 0.1 second mobile speed improvement lifted retail conversions 8.4% and travel 10.1%. Source ↗
Related: the speed-to-lead calculator prices what happens after the page does its job, and the web design page shows how we build pages that pass every check above.