← Back home

Methodology

Scoring: simulation vs live questions

The free score is a heuristic estimate. Paid / trial runs ask ChatGPT and Gemini real buyer-style questions. Scores can diverge — that gap is the diagnostic value, not a product flaw.

Simulated recommendation rate (free)

We score structure, semantic clarity, schema, FAQ coverage, and more, then map that to “chance AI would recommend you.” Fast and low-cost for triage.

Live question checks (paid / trial)

We ask ChatGPT and Gemini in a buyer voice and count brand/product mentions. Models, timing, and competition can make live rates lower than the sim score.

Default prompt count and engines

A paid/trial run defaults to about 5 buyer-style prompts (configurable, max 10), usually on ChatGPT and Gemini. Checkout can change engines; prompt count stays a product default to control cost and expectations. You pay per delivery — not an unlimited probe subscription.

Recommended loop: rewrite → paste → re-check

Simulate → prescribe rewrite → paste live → re-check with live probes. Your workspace compares rates for the same URL over time.

Monitoring is step two

A single rewrite answers “can I get recommended now?” Monitoring fits after you ship — checkout leads with the single pack; monitoring stays optional.

Want anonymized narrative examples? See Cases