Operating scope
Country head
Capability
All capabilities
Active handovers
0
Outcomes written back
0
Connected work · Agent system
MarginOS · screen 04 of 07
Shrink
Read against zone activity and count history, anonymously, at zone level only.
Issues in scope
3
3 instances across 2 places
Waiting on you
2
0 already decided in this session
Three largest
AED 3.5m–AED 11.6m
Modelled annual value, three largest only
Running without asking
0
0 armed and waiting on a threshold
Decision owner
Country head
Returns to Loss prevention, Information technology
Scope
Oman
27 stores · all capabilities
Listen
Shrink by store and categoryAnonymous zone activityCount historyTill exceptions
Process
Zone level only, never person levelCoverage gaps against shrink concentrationCount accuracy or loss
Act
Coverage proposal to loss preventionRefusal to name individualsMissing feeds named
Written backEvery outcome returns to the systems that produced the evidence.
KindOrder
Queue
3 issues · 2 waiting on youApprove
What happened
In each of these stores, shrink concentrates in two zones, and those zones are also the two with the lowest camera coverage in the store. The correlation holds across all 5.
Why it matters
Shrink is reported monthly as a percentage with no cause attached, which makes it almost impossible to act on. A coverage correlation is not proof of anything, but it is the first testable hypothesis this number has had.
Already done
The correlation is quantified per store and framed as a coverage question rather than an accusation, with the zones mapped.
Worth
AED 3.5m to AED 11.6m a yearModelledShrink concentrated in the lowest coverage zones. Value is the share addressable by closing the coverage gap, at the observed concentration.Shrink concentration in low coverage zones
Then what
Coverage proposal with the zone mappingLoss preventionReports back 7 Oct
If approved
The coverage gap goes to loss prevention as a proposal with the zone mapping
If declined
Shrink stays a monthly percentage with no cause attached to it
Blocked
Oman cannot be read the same way because till exceptions are not in the feed
OmanWhat happened
The shrink read joins zone activity, counts and till exception data. Till exceptions for this market are not available, so the join cannot be completed.
Why it matters
This market is not showing low shrink. It is showing an incomplete read presented as a result, which is how a blind spot becomes a conclusion. It is named rather than quietly excluded.
Already done
Excluded from the country comparison above, with the gap stated.
Exposure
Not a revenue lineThis market is not showing low shrink, it is showing an incomplete read presented as a result.Markets with a complete shrink read
Then what
Feed request raised, market excluded until it landsInformation technologyReports back 7 Oct
Needs
Till exception data in the feed for this market
Owner
Information technology, point of sale team
Declined
The data would support naming individuals and the system does not
Al Falah Central MallWhat happened
Zone activity, count history and till exceptions together could narrow shrink to shifts and from shifts to people. The read stops at the zone.
Why it matters
An accusation built on a correlation is wrong often enough to destroy trust with store teams permanently, and one wrong accusation costs more than several years of the shrink it was meant to address.
Already done
Every shrink action here names a zone and a coverage gap. None names a shift or a person.
Exposure
Not a revenue lineAn accusation built on a correlation is wrong often enough to destroy trust with store teams permanently, and one wrong accusation costs several years of the shrink it addressed.Shrink read at zone level only
Then what
Refusal held, zone level onlyGroupStanding, no return date
Override
Nobody. Refused at design
Cost of overriding
Naming individuals from a shrink correlation is an accusation the data cannot support

