---
type: "Article"
title: "We built a score to catch generic content. Its first casualty was our own spec."
description: "The Non-Commodity Score is a pre-publish gate for generic, commoditised content. Its v0 spec failed its own test. Here is the teardown and the rebuild."
resource: "https://www.laurelinlabs.com/blog/non-commodity-score-teardown"
tags: ["non-commodity-score", "topical-authority-content"]
generated: { by: "laurelin-site-build/2026-08-25", at: "2026-08-25T12:00:00+00:00" }
status: "stable"
stale_after: "2027-02-25T00:00:00+00:00"
---
# We built a score to catch generic content. Its first casualty was our own spec.

Published 2026-07-07, updated 2026-08-25. Author: [Piers Butler](/entities/piers-butler.md).

The Non-Commodity Score is a pre-publish gate that scores a page 0 to 100 on how far it sits from generic, commoditised content, across four vectors: information gain, experiential evidence, empirical telemetry and entity connectivity. The v0 specification failed its own test: its flagship metric measured distance from the topic centroid, which rewards off-topic gibberish, and the embedding model it named cannot produce document vectors. This teardown records what was wrong, what we changed (a relevance floor, additive on-topic gain, a calibrated model and quoted evidence for every grade) and why the tool now marks its own homework honestly.

## FAQ

### How is information gain measured in the Non-Commodity Score?

As additive, on-topic novelty: claims, data and examples the page contains that the top-ranking set for the query does not, measured after a relevance floor so drifting off topic cannot score. Every gain is quoted back as evidence.

### Can I run the Non-Commodity Score on my own pages?

Yes. The free tool at laurelinlabs.com/tools/non-commodity-score scores any URL, two runs without signup, with the quoted evidence behind each grade.

Hub: [Topical Content](/topics/topical-content.md)