Editorial policy

How scores are set. Each criterion is scored from 0.0 to 10.0 in one-decimal steps. Nobody gets 10.0. A capability stated on a primary page with names, limits or a sourced number scores 8.0 to 9.5. A capability that is stated but thin, vendor-claim only, or read through a secondary source scores 5.0 to 7.0. A capability the pages read do not mention scores 2.0 to 4.0. Scores between 7.0 and 8.0 are not used. Below 2.0 is reserved for a vendor that documents that it does not do the thing.

The weights are published and sum to 100: playbook authoring model 14, integration count and depth 14, migration tooling and formats 14, approvals and human-in-the-loop controls 12, documented AI features 12, ownership stability 10, case management 8, hosting and data residency 8, pricing transparency 8. A total is the sum of score times weight divided by 100, shown to two decimals. The build checks that the weights sum to 100 and that every published total equals the sum of its parts.

Where two totals are within 0.25 of each other, the page calls them a band, not a ladder. The scores rest on pages read on 2026-11-03, and a score set on a page that was not read in full is flagged for a re-read. We also publish the order with migration and ownership removed, so a reader can see what the weights do.

Vendor claims. A count, a percentage or a time saving that a vendor publishes about itself is labelled a vendor claim and carries the page and the date. An accuracy or closure figure never lifts a score above 8.0 unless the method, the sample and the party who measured it are published. Mate Security is not a SOAR, so each criterion is scored on the answer it documents, and nothing is skipped.

Corrections and vendors. A vendor, a customer or a reader can send a correction to editor@aftersoar.com. Send the page, the claim and a primary source. We check it, change the page if it holds, and add a dated line to the corrections log. A vendor cannot pay to change a score, a verdict, a bout or a position in any ranking.