Research & method · One of three instruments
A homepage is not a strategy, but it is the one part of a company's commercial thinking that anybody can read. So I read 105 of them the way a buyer would, and 98 could be read at all. Each answers six questions or fails to. Cut the field by segment, by country, or leave it alone and look at the shape.
Sites are grouped by how many of the six questions they clear. The top row is a homepage a buyer can do nothing with. The bottom row is one they can read and buy from.
Number of sites in each group. Rust marks the median.
You can cut by segment or by country, but not both at once, so choosing one clears the other. Twenty-one of the country-and-segment combinations contain a single company, and "the one Italian orbital-logistics firm" names it as surely as printing the name would. One axis at a time keeps the smallest group readable at six. Only the six countries with enough companies are listed, which is the same line the audit itself draws. The eleven sites that cleared all six are named further down, because that part is worth saying out loud.
Ranked by the share of companies showing any visible way to buy, from a price list to a quote form to a shop.
Germany reads most like a normal commercial market: twelve of seventeen German sites give a buyer somewhere to go, and six name a private customer, both the highest here. Spain names customers often enough, though four of those seven are agency contracts, and only two of eleven sites offer any way to buy. France names a customer on three of twelve readable sites and shows a buy path on three. The two countries with the largest national space agencies exert the least commercial pull on their own companies' pages, which is a finding I did not expect and cannot yet explain.
Country totals count readable sites only. France loses five sites to unreachable domains and Spain one, so its row is thinner than the country's real sector. Anything outside these six is too small a sample to report.
These are the sites where a buyer can find out what it does, who it is for, who already pays for it, and how to start, without sending an email into the dark. Named here because naming them is praise. Nothing else in the dataset is named.
Version 2, July 2026. The hiring index is recomputed each quarter. The website audit is re-run when the sample changes. Where a count is revised, the correction is published with it rather than folded in quietly. Next revision Q4 2026.
Coded from the homepage, plus one product or pricing page where the homepage pointed at one, on 16 and 22 to 23 July 2026. Sites change. This is what they said on those days.
Answer them about your own homepage and your row lands in the distribution above, next to the other 98. It takes a minute, saves nothing, and asks for no email.