Research
Research from
real websites.
We read public files on real company websites and report what we find. Every study shows its method, the date we checked and a data file you can download.
Kookies research studies use only data we collect ourselves from public web pages and files, such as robots.txt, llms.txt and home page HTML. Each study names its sample, the date it was checked and its limits, and offers the full data as a CSV. We do not run surveys or estimate numbers we did not measure.
Studies
What we found.
Do top cloud companies block AI crawlers?
Almost none of them block AI crawlers.
How many top cloud companies have an llms.txt?
59 of the 100 companies on the Forbes Cloud 100 (2025) list have an llms.txt file, and 13 also publish llms-full.txt.
How ready are top cloud companies for AI search?
Most score well on the basics.
Do top cloud companies show their prices?
Most do, at least in part.
How fast are top cloud company home pages?
Most are slow on mobile.
Do pricing pages and articles score like the home page?
Close, but a little lower.
How long are the demo forms of top cloud companies?
The median demo or contact sales form asks for 6 things on its first step.
Do top cloud companies send security headers?
92% of the 99 home pages we could read force HTTPS with Strict-Transport-Security, and 73% stop other sites from framing the page.
How do top cloud companies look when shared?
99% of the 90 home pages we could read set a share image, and every one of those images loaded.
What do top cloud sites load before you click?
80% of the 99 home pages loaded at least one tracking tool before any click, and 72% loaded an ad or B2B intent pixel.
How accessible are the home pages of top cloud companies?
The median home page scores 93 out of 100 on Lighthouse accessibility, and 74% score 90 or more.
How we research
Real data,
shown in full.
Six rules every study follows, from the list of sites to the file you download.
- 01A named sample
Each study says exactly which sites it covers and where that list comes from.
- 02A declared bot
Our requests say who we are, with a link back to this page. We follow robots.txt.
- 03Public data only
We read what any visitor or crawler can read: public pages and files. No logins, nothing private, no form is ever sent.
- 04Same checks, same day
Every site gets the same rule based checks, run by code, in one pass.
- 05Not read, never guessed
Sites that refuse us are listed as not read. We never fill in a number.
- 06Data you can check
Every study has a CSV with one row per site, so you can check any number on the page.
Next studies
Get the next
study first.
We publish a new study when the data is in. Leave your email and we send each one when it is out. Nothing else.
We send the next study the day it is out.


