TIN measures. A study asks a business question a buyer would actually ask, publishes its method before any data is collected, and ships the dataset it was computed from. Anyone can re-run it and get the same answer, or show that it is wrong.
One business question per study, with a number in the answer.
The method is committed to the repository before any data is collected.
The CSV the findings were computed from ships with the study.
Studies
4 studies · no badge, by designAI visibility tool pricing, measured: 36 tools on one unit. A check costs a median $0.0205; 100 daily prompts on 3 engines, $199/mo. Data and CSV.
Read the study →AI citation sources: 2,857 citations from ChatGPT, Gemini, Perplexity, Copilot, Claude and Google AI in 10 countries, coded by site type. Data, method and CSV.
Read the study →AI Overview citations vs Google rankings: in the US, 30.7% of top-10 domains are also cited (69 searches). By position, query type and market, with data.
Read the study →238 AI and automation agencies measured: 67.6% get cited by AI engines, but the top 24 take 78% of citations. Data, method and CSV.
Read the study →The four bands belong to case files, where a claim has been checked against the public record. Here TIN is the instrument, so the thing to trust is the method and the data, both of which are on the page. How that differs, and why, is set out in the dojo.
