The fourth showed whether witness information had been preserved.
The fifth showed whether the complaint appeared in the monthly performance report.
Evelyn explained that the investigation had now expanded beyond the original sample.
The team had reviewed several hundred historical complaints and identified a substantial number requiring further examination.
Some involved incomplete documentation.
Others had been closed before requested witness interviews were conducted.
A smaller group contained evidence suggesting that information had been removed or reclassified in ways that materially affected the final review.
The investigators were still distinguishing between technical errors, negligent decisions, and intentional interference.
But the pattern was clear enough to require immediate action.
"How many cases are we talking about?" one director asked.
Evelyn gave the current verified figures and emphasized that they remained preliminary.
"The total may change as we examine archived records. We are not presenting every case as misconduct. We are identifying cases where the process did not reliably support the conclusions recorded in the final report."
The director frowned.
"So the company may have been reporting successful complaint resolution without knowing whether the complaints were actually resolved."
"That is a fair description of the risk."
Daniel shifted in his chair.
"The system was never intended to suppress complaints."
Evelyn turned toward him.
"I am not suggesting that the system was originally designed for that purpose. I am explaining how its incentives and controls interacted."
Daniel leaned forward.
"The closure targets were introduced to improve response times. We had thousands of complaints sitting unanswered. Customers were waiting weeks for basic responses."
"That is correct," Evelyn said. "The original objective was legitimate."
"Then we need to be careful not to condemn the entire process because some people failed to follow it."
"We are being careful. That is why we are separating the failures."
Vanessa watched Daniel.
He was not entirely wrong.
The closure targets had been created to address a genuine problem.
Sterling had once struggled to answer customers promptly. Passengers had waited weeks for reimbursement decisions. Employees had submitted reports that remained unanswered for months.
The original reform had been sensible.
But a sensible measure could become dangerous when people began treating it as the goal rather than a tool.
Vanessa had seen that happen in other areas of the business.
A punctuality target could encourage managers to delay boarding reports until a flight departed.
A cost target could discourage necessary maintenance spending.
A customer satisfaction target could pressure employees to offer compensation without addressing the underlying problem.
Numbers were useful.
They were also capable of concealing reality.
"Daniel," Vanessa said, "I agree that the original objective was legitimate. But once the closure rate became a measure of success, did anyone verify that the complaints had been investigated properly?"
"We had quality assurance reviews."
"How often?"
"Quarterly."
"Did those reviews include complaints that had been reassigned?"
"Some did."
"Did they examine whether witness information had been carried forward?"
Daniel hesitated.
"I would need to check."
"Did they compare the original complaint against the final closure summary?"
"I don't know."
Vanessa nodded.
"Then we have identified the problem. We measured the outcome without adequately verifying the process that produced it."
Daniel looked down.
Evelyn advanced to another slide.
This one contained an email exchange dated fourteen months earlier.
The sender was a regional customer experience manager.
The recipients included Martin Keller, the customer experience director, and several senior operational managers.
The subject line read:
Quarterly Closure Performance โ Required Improvements
The email stated that the division was below its target for complaint resolution and needed to reduce the number of cases remaining open at the end of each reporting period.
A reply from the customer experience director said:
Please ensure that all cases are reviewed promptly and that the appropriate closure codes are applied. We cannot continue carrying unresolved matters into the next quarter.
The wording was not inherently improper.
But a later message, sent privately to several team leaders, was more troubling.
Where evidence cannot be independently verified, please use the standard closure template and avoid extended correspondence. We need consistency across the division.
The instruction did not explicitly say that legitimate complaints should be dismissed.
However, in cases where witnesses had been identified but never contacted, the language created a mechanism for closure without meaningful verification.
Evelyn displayed another message.
This one came from a team leader who had questioned the practice.
If a passenger provides the names of witnesses, should we not contact them before closing the case?
The response was brief.
Use judgment. We cannot allow every allegation to create an open-ended investigation.
Vanessa read the message again.
It was a familiar argument.
There were practical limits to every investigation. Not every complaint could be pursued indefinitely. Some witnesses would be unavailable. Some allegations would lack sufficient evidence. Some disputes would remain unresolved despite a reasonable inquiry.
But the question was not whether every allegation deserved unlimited investigation.
The question was whether the company had established a fair and consistent standard for deciding what evidence to seek.
The records suggested that it had not.
Worse, the closure targets had made the decision financially and professionally significant for the managers involved.
Evelyn returned to the main presentation.