Chapter 77: After Every Complaint
Two blank control records sat beside four complainants whose files gained the same reference hold within twenty-four hours.
On Friday, Marcus began the isolated demonstration with the selection rule, not the result.
The HRIS and IT custodians had exported L09 directly to KRR in read-only native format.
The package included event IDs, server times, case keys, and source receipts.
Names, addresses, reference text, and outside organizations remained redacted from my view.
Marcus explained how the systems generated and retained the fields.
His existing anti-retaliation protection remained active while he performed the demonstration.
He did not choose which cases helped KRR's theory.
The four test records came from the locked cross-department protocol approved the week before.
The two controls came from the same periods and systems.
Each control had a comparable employment status but no recorded complaint in the tested window. The protocol matched time band, record type, and eligible reference activity without pretending two people were perfect statistical controls.
Samuel required the control rule to be fixed before the outcome fields opened. KRR stored the query, selection timestamp, and hash so no one could quietly redraw the comparison after seeing the holds.
Marcus exposed the first complainant timeline.
The complaint receipt appeared at 9:14.
The risk designation followed before the same time the next day.
A reference hold then entered the workflow under a separate event ID.
The second case followed the same order in seventeen hours.
The third took just under twelve.
The fourth took twenty-three hours and forty-one minutes.
Every interval fell inside one day.
The logs proved the order and system activity. They also showed that ordinary corrections left distinct events, allowing Marcus to explain why none had been silently backdated.
They did not prove that anyone acted because of hostility toward a complaint.
They did not show what a recipient later believed or decided.
Grant's method response raised selection bias and requested noncomplainant comparisons.
Samuel pointed to the two locked controls already on the screen.
Marcus opened the first control timeline.
Its ordinary reference field changed twice during the period.
No reliability label or strategy-office hold appeared.
The second control included a manager correction and a routine employment verification.
It also carried no comparable hold.
Two controls could not prove that no untested employee ever received one.
They did test Grant's claim that the hold appeared routinely without complaints.
In these selected records, it did not.
Samuel recorded the result with the sample size beside it.
The result narrowed one defense while leaving prevalence, intent, and external harm outside the log's reach.
I asked whether twenty-four hours could be presented as motive.
“Only as timing,” he said.
Motive still required documents and witnesses capable of explaining the instruction.
The L09 export would support access and sequence, not a person's private purpose.
It could not identify the author of a label merely because an administrator account processed the field.
KRR sealed the native package and generated separate redacted demonstration receipts.
The source custodians retained their original systems and verification authority. Their declarations identified ordinary corrections and administrator actions, allowing challenges without exposing the protected personnel content.
Marcus left with no copy of witness statements or outside-contact reports.
I left with no copy of L09.
The pressure on Marcus was now easier to predict and harder to dismiss. His role in producing the export could make him a target even though he had not selected its conclusion.
Dana requested written confirmation that his protection covered this expanded demonstration.
Marcus confirmed future questions would return through his attorney and protected custodian channel, not through his WMG supervisors.
KRR logged the request before closing the room.
Samuel then opened the identity custodian's mapping result for the third case.
The project-management ID mapped to Malik Thompson.
KRR scheduled an independent-reference and address-protection meeting for Monday.
Until Malik consented, his name would remain outside any public description.
Until that meeting, KRR would not contact his prospective employer.