Public buyer-material coverage at 48 technical software companies

What was measured

We selected 48 active-company records from the research corpus: the first 12 alphabetically in each of four categories after excluding the historical Airplane record. We extracted every distinct first-party URL cited throughout each saved research record, then matched page purpose and evidence notes against a disclosed checklist.

The company records were reviewed from 3 to 9 July 2026. Individual cited sources can be older, and the record review date is not the page publication date. Current checks on Baseten, Qdrant, Clerk, WorkOS, Promptfoo, and Cribl were completed on 8 September 2026 for the related pricing and source-matching reports.

Coverage in the selected records

Pricing or packaging pages were matched for 36 companies. Customer or case-study pages were matched for 30, technical documentation for 28, provider security or trust material for 20, and deployment material for 19.

The indicators overlap. Baseten had separate documentation, trust, and customer sources, while its pricing and dedicated-inference pages also described deployment choices. A company contributes once to each indicator, and the CSV retains one representative matched URL per indicator.

How to use the result

• For a buyer research pack, check deployment and operating responsibility deliberately. Less than half of the reviewed records had a matched first-party deployment source.
• Keep pricing, technical documentation, security evidence, customer proof, and deployment material as separate checklist items. Finding one does not answer the others.
• Record the page date and the unanswered question. A link without a scope or observation date is difficult to review later.
• Open the matched URL before using a claim. A page category tells you where to look; it does not prove that the page answers a buyer's exact question.

Limits

The sample is a convenience sample from one research corpus. Alphabetical selection is reproducible, but it is not random. Saved research records contain different numbers of first-party citations because research depth varied by company. Domain migrations were mapped for known cases such as Port, Cortex, ConductorOne, Coval, CNTXT AI, Netris, and Abnormal Security.

A blank CSV cell means no qualifying URL was found in the reviewed saved research record. It does not prove the material is absent from the live site. No association with growth, conversion, sales cycle, or company performance was tested.