Raw signal count alone is a misleading metric on its own. An entity that produced ten signals over the past year but has been quiet for the last two months presents a meaningfully different current risk and opportunity profile than one that produced the same ten signals concentrated within the past two weeks — even though both would show an identical total count.
Why this distinction matters for accurate scoring
CIA scoring weights recency explicitly so that stale historical activity contributes less to a current score than fresh, ongoing activity does. This prevents a common false-positive pattern: an entity that was genuinely active months ago but has since gone quiet still showing an artificially elevated current opportunity score purely because of accumulated historical signal count.
How recency interacts with the other two CIA dimensions
- Frequency: how often an entity produces signals, providing the baseline activity level
- Recency: how recently that activity occurred, decaying the weight of older signals over time
- Source diversity: how many independent collection sources are corroborating the current pattern, weighted alongside frequency and recency rather than in isolation
What this means for any entity that's gone quiet
An entity that was highly active six months ago and has since gone quiet will see its CIA score decay accordingly, even if its historical signal count remains technically unchanged. This is intentional — the score is designed to reflect current relevance, not a permanent record of past activity.