What's new in Discoverability Engineering

The book now has a chapter, “A Discoverability Score Measures Whether Finders Who Don’t Know Your Name Meet You,” after “The Same Loop Qualifies a Federal Supplier and a Health-Tech Partner.” It separates recognition, whether an answer engine describes a subject correctly when it’s named, from discoverability, whether the engine introduces the subject when a finder describes a problem without its name. It gives a nine-part Discoverability Score with its weights, ten question classes, a benchmark that’s frozen before the first run, a seven-attribute evidence rubric, the Competitive Inclusion Set, the Evidence Gap Register, and a rule that every finding carries an observation, a hypothesis, and an experiment. Three worked examples follow: the author scored from public evidence, an anonymized engineering firm whose public evidence is thin, and an anonymized consulting firm with hundreds of pages and no inclusion. A fourth run, on invented data, shows a retest with a regression and an intervention that did nothing. Three prompts close the chapter, and the glossary gained seven terms. The online companion carries the rest: the formula for every dimension, the question classes in full, the evidence rubric, the retest procedure, the full worked examples and the invented run, and a toolkit you can download that validates your inputs and computes the score. The chapter after it moved up one number, so the book now has eighteen chapters. A printed copy you already own doesn’t have it.

Merriden’s Actionboard figure in “Controlled Tests and the Actionboard Show Whether Discovery Is Getting More Correct” now has a fourteenth tile, prior-art retrievability, added on September 28. It is the row a subject adds once it publishes prior art, and the thirteen questions in the table are unchanged. A printed copy you already own has the earlier figure.

The book now has a chapter, “Patents and Prior Art Are Public Evidence That Needs a Date Anyone Can Prove,” after “Government Records, Supplier Portals, and Directories Must Agree.” It shows how to publish a disclosure that an examiner can find and date, how to make your own patents readable as capability evidence, and how to check a supplier’s or partner’s patents before you qualify them. It adds three tools (the Prior Art Record, the Examiner Retrieval Test, and the Publish-or-Protect decision), a patent node in the Public Evidence Graph, and short additions to the chapters on finders, names, government records, and failure modes. Later chapters, figures, and tables moved up one number, so the book now has seventeen chapters. The dated facts behind it, such as venues, search tools, and grace periods, are in the online appendix. A printed copy you already own doesn’t have it.

A prime contractor can’t team with a supplier it can’t find, and more of the finding is now done by AI. Picture a capture team that needs three qualified small businesses in 45 days. Its AI sourcing agent searches the web, SAM.gov, supplier portals and answer engines, and a supplier that fits perfectly doesn’t make the shortlist because its certifications are in a photo, its past performance is in an emailed PDF, and its government record is under an old name. Discoverability Engineering shows how to build the public evidence that lets a careful human or machine find you, understand you, check your claims, decide you qualify, and reach the right person. You write a Finder Specification, replace keyword lists with the questions finders actually ask, tie every public claim to evidence, keep one consistent story across your website, SAM.gov and directories, and test answer engines on a schedule. It is for small businesses and the teams that source from them.