Five things, not two hundred rows

A crawl of a mid-sized business site will find somewhere between eighty and four hundred issues. Almost none of them are worth a conversation this month.

The skill is not finding problems. Any crawler finds problems. The skill is ranking them.

Severity is not impact

Crawlers sort by severity: how broken a thing is in the abstract. That is the wrong axis. What you want is impact: what this specific problem costs this specific business.

An accessibility warning on an image in a 2019 blog post is technically a real issue. A noindex on the services page is a business emergency. A severity sort can easily put them near each other.

Here is the order I use, top to bottom:

1. Can Google reach and index the pages that make money? Nothing else matters if the answer is no.
2. Is the journey broken? Broken links, server errors and redirect loops on the path to a booking or a checkout.
3. Is there a trust problem? Mixed content, certificate warnings, anything that makes a browser complain.
4. Is it slow where it counts? Mobile speed on the pages people actually land on.
5. Everything else.

If you cannot say what a fix is worth, it does not go in the top five.

Effort belongs in the ranking too

A high-impact fix that takes a developer three days competes badly against three medium-impact fixes that take an afternoon between them. Agencies that ignore effort end up recommending things that never get done, which is the same as recommending nothing.

Tag every item with impact and effort. Sort by impact, then break ties by effort ascending. The list rearranges itself into something a client will actually complete.

What to do with the other 195

Keep them. Show them in an appendix, in a table, grouped by category. The client will not read it, but its presence does two things: it proves you looked properly, and it means that when someone eventually asks "what about alt text", the answer is already written down.

Just do not put it on page one.