Output types, missing data, and what never becomes a score.

How to read evidence

Every number on screen has a type. Use the type, source, vintage, and geography beside the figure before you treat it as a reason to commit capital.

Missing stays missing. Decised does not invent a midpoint, zero, or average, and it does not turn crime, news, or venue maps into a decision score.

Output types

User fact: observed input you supplied.
User assumption: planning figure; treat as editable.
Unverified evidence: community or keyword context; descriptive only.
Observed benchmark: published official rate or count for a geography and industry.
Model estimate: statistical output from a versioned engine.
Derived calculation: arithmetic on other typed values (for example cash runway).

What never becomes a decision score

Recorded crime in the map radius
Local news and headlines matching investment-related keywords
OpenStreetMap venue pins
Companies House register-survival probabilities (research-only until gated)

Same-industry concentration

This is registered-business presence in a SIC section compared with other UK local authorities. It does not measure demand, footfall, spending, capacity, rent, or true competitive saturation.

Business language for ONS activity

Office for National Statistics (ONS) “birth” and “death” rates appear as business openings and business closures. The comparison of openings with closures is for the same published period. Higher or lower is not automatically good or bad. These figures come from Inter-Departmental Business Register (IDBR) activity aggregates, not a Decised success score.

Evidence priority

The evidence-priority table on a decision case is an experimental heuristic. It has not been validated against real decisions and outcomes.