{{ headline }}
{{ sp.p }}
{{ flagMsg }}
i
The numbers behind that
i
Where the weakness sits
i
How to read the rest of this
Where the property stands today.
{{ infoState }}
Reads
{{ r.id }} {{ r.n }}{{ r.m }} {{ r.scText }}
{{ r.na }}
Reads showing insufficient evidence are not counted as weaknessesi
The order things happened in.
{{ infoSequence }}
What changed, and when
i
iWhy these are reported together
{{ ev.det.hDate }} Compound breakhow far the evidence goes →
{{ ev.d }} {{ ev.x }}{{ ev.t }} →
{{ sigName }}
i
{{ sigWhy }}
iAbout these predictions
Performance either side of each change
The three causes we can defend.
What we are attributing this to
i
View basis
i
How we are allowed to word it
i
Three possible causes. Each under test.
The compression is real and measured. We have not yet determined which of three causes is producing it. Each candidate below has a test under way, with results due 15 November.
Worth acting on now
The testing record
i
{{ econHeadPlain }} {{ econHeadEm }}
The figure below is what the current trend produces if it continues and nothing is done. It is not a loss, it is not carried as one, and it will not appear as attributed value at risk unless a cause is confirmed.
These are costs incurred during a defined interval that has closed. They are not annualised and not capitalised. Capitalising a cost that has already stopped would overstate it.
i
iHolds on a hold or a sale
{{ econSecH }}
i
{{ econRowsH }}
Scope of the finding
{{ portfolioNote }}
What we had, and what we didn't.
{{ infoEvidence }}
Coverage
Where the evidence fell short
i
{{ nr.na }}
Where your systems disagree
i
How a conclusion gets made here.
31 of 38 reads are derived from defined data or collection protocols. Senior operating judgment is reserved for the limited questions that cannot be reduced without destroying their meaning.
Where every number came from
Each score carries its source document, the location within it, and when it was recorded. Where two of your systems disagree, both figures are shown and neither is used to support a finding until the disagreement is resolved.
Dates taken from correspondence rather than a system log are marked as reconstructed. A finding built on a remembered date is weaker than one built on a logged date, and you should be able to see which you have.
What we are allowed to claim
Findings are placed on a scale according to how far the evidence goes. Timing alone permits one kind of statement. Timing plus tested alternatives permits another. A measured reversal permits the strongest. The wording in your report is controlled by where the finding sits, not by how confident anyone feels.
We name at most three causes. Where nothing reaches confirmation, we report hypotheses and say so rather than promoting the strongest guess.
What happens to your information
Your documents and exports remain yours and can be deleted when the engagement closes. Your diagnosis is yours to use internally, permanently.
We do not ingest resident information. Employee records are reduced to role and seat before analysis. Nothing that leaves here identifies you, your asset, your people, or your residents.
No outside AI provider trains on your data. That is a different thing from us improving our own method, and we do not permit it.
What we keep, and why it helps you
What we retain is the record of whether our approach worked. When we predict something and get it right, or get it wrong, we keep that. Every client benefits from the record built before them, and the only way that works is if it keeps being built.
Methodology version 1.0. Thresholds in early calibration are marked as such in your report rather than presented as settled.
Secure intake for your engagement.
What the engagement requires, what has arrived, and what is outstanding. Documents are held for the engagement team and can be deleted at any time. Nothing is analysed automatically.