Field notes on part proliferation, geometric search, should-cost, and keeping engineering knowledge inside your network — written for the people who actually run CAD libraries.
Engineers spend up to a third of their time searching for information — and re-creating parts that already exist is near the top of the list. Here is what a single duplicate actually costs, why it keeps happening, and why most shape-search tools never solved it in practice.
Read the article →The uncontrolled growth of part numbers is one of the most expensive, least-measured problems in product development — and it compounds. A complete guide to the causes, the full cost, and the five changes that actually reduce it.
Read the article →Shape search matches the geometry of a part instead of its filename or metadata. A step-by-step look at how the technology works — normalization, signatures, indexing, ranking — and why machine learning changed what it can do.
Read the article →Filename search doesn't search your parts — it searches the labels people were too busy to assign consistently. The five ways it breaks, and why shape search answers a fundamentally different and more useful question.
Read the article →A CAD file can be ITAR-controlled technical data, and under export-control rules even showing it to the wrong person can be a violation. Why cloud tools hit a wall in these environments, and what on-premise architecture actually requires.
Read the article →You don't get a reuse culture by telling engineers to value reuse — you get it by making reuse faster than recreation. The four pillars that make it stick, a staged rollout, and the mistakes that sink these programs.
Read the article →PLM-integrated shape search or a standalone CAD-neutral tool? An honest comparison — including why the bundled module so often goes unused, and why deployment architecture, not features, is the deciding factor for controlled libraries.
Read the article →Suspicion that 'we probably have duplicates' doesn't fund anything. A practical step-by-step guide to scoping an audit, defining and clustering duplicates, scoring the savings, and consolidating safely through a governed process.
Read the article →A part's cost drivers — complexity, tolerances, features — are overwhelmingly geometric, which makes the CAD model the natural input for should-cost analysis. How it works, where it pays off, and why it's adjacent to part search.
Read the article →Every other corporate function has a queryable system of record; engineering has network folders. The case for a geometric data warehouse — the layer that makes a company's entire geometry searchable, analyzable, and the foundation for everything from reuse to cost to AI.
Read the article →