A single similarity percentage is useful, but it flattens a genuinely richer underlying comparison into one number. Looking more closely at the six validated cases reveals that they don't all share the same kind of signal shape — they cluster into recognisable categories, or archetypes, each with somewhat different characteristic timelines and signal compositions.

The archetypes visible across the current library

SpaceX and Tesla both show what might be called an infrastructure-building archetype — sustained capital deployment, physical manufacturing or engineering milestones, and notably longer lead times (9.4 and 9.7 years respectively) than the rest of the library. Shopify and OpenAI cluster into more of a platform-growth archetype — steady developer or merchant ecosystem expansion, shorter lead times (8.1 and 6.8 years), and signal profiles built more on cross-platform digital activity than physical infrastructure milestones.

Bitcoin and NVIDIA sit somewhat distinctly from both clusters — Bitcoin as a protocol-adoption archetype with the longest lead time in the library at 11.3 years, and NVIDIA showing characteristics of both infrastructure and platform patterns simultaneously, which may help explain why it carries the library's highest peak opportunity score at 0.91.

Why archetype matters alongside similarity score

  • An entity matching SpaceX's signal shape specifically suggests an infrastructure-building pattern, with the longer maturity timeline that implies
  • An entity matching Shopify or OpenAI's shape suggests a platform-growth pattern, historically associated with a somewhat shorter path to recognition
  • Knowing which archetype an entity resembles, not just its raw similarity percentage, gives a more complete basis for understanding what kind of trajectory the comparison actually implies

Where this goes as the library grows

With only six cases, these archetype groupings are observational rather than statistically rigorous — a genuine pattern worth naming, not a definitively proven taxonomy. As the precedent library grows, archetype classification is expected to become a more formal, more reliable layer of the comparison output.