Read what your page actually sends
Why the word count is at the top of the list
Metadata is easy to grade and rarely the problem. The problem we see over and over on real client sites is subtler and far more expensive: the page is a shell. The head is complete, the title is well written, and the body that arrives over the wire contains forty words of boilerplate because everything else is rendered in the browser.
To a person it looks perfect. To anything reading the response — and that includes several of the crawlers that feed answer engines — there is nothing on the page to quote. This is the default state of a site built with an app framework and no server rendering, which increasingly means a site built by an AI app builder.
You can check it without any tool: curl -s https://yoursite.com | wc -w. If that number is small, it is the first thing to fix and everything else can wait.
Metadata is necessary and not sufficient
Fix what is listed here, then find out whether it changed anything. Real buyer questions across ChatGPT, Perplexity and Gemini will tell you who is being recommended instead of your client, and which questions the site has no answer for at all.
Questions
See who the engines recommend instead
12 buyer questions across ChatGPT, Perplexity and Gemini. You get the questions your client loses, who won them, and what is missing from the site. No score, no signup wall.