Configuration
The lines every judgement in Odau is made against. Changing one changes what the lenses say about the portfolio, without touching a single record.
Scoring Thresholds
These are Odau's defaults, and every one of them is yours to change. Odau reports what it measured — where it calls something healthy or worth reviewing, it is measuring against the line set here, not against an opinion of its own.
Applicable Utilization
Odau defaultSeats where an application is seat-based, allocation otherwise. This is the figure waste and Portfolio Performance both read.
Usage Signal Score
Odau defaultThe composite of licence utilization, engagement and consistency.
Licence Utilization
Odau defaultActive people against purchased seats.
Engagement Rate
Odau defaultActive people against the seats actually handed out.
Consistency
Odau defaultHow much of the period an active person shows up in, rather than whether they showed up at all.
System Health Score
Odau defaultWhether a record is complete enough to govern — not whether the tool earns its money.
Adoption Rate
No thresholds, deliberately. Low adoption is context rather than a finding — a tool bought for four specialists is not failing because 2,796 people do not use it — so Odau does not band it and there is no line to set.
Approval Workflows
A request copies the chain at the moment it is submitted and keeps that copy for the rest of its life. Editing a workflow changes what the next request gets — never what somebody is already waiting on, and never what a closed request was decided against.
How These Are Applied
A score at or above its healthy line reads as healthy. Below the review line it reads as worth reviewing. Everything between the two is benchmark — neither a finding nor a clean bill. The review line can never sit above the healthy line, because the band between them would have nothing in it.
A score with no value is not banded at all. An application with no usage source is reported as unmeasured rather than placed below a line it was never assessed against — an absence of data is not a poor result.