Context

A mid-size analytical lab serving regulated industrial customers.

On paper: solid KPIs, passing audits, acceptable turnaround times.

In reality: constant pressure, late nights, and a quiet loss of credibility with senior customers.

The LIMS showed no red flags. The floor told a different story.

The Blind Spot

Management relied on:

• Sample counts
• Pass/fail run status
• Average turnaround time

What they didn’t see:

• Partial runs quietly restarted
• Analysts compensating for flaky instruments
• “Successful” results produced under stress
• A permanent yellow state mistaken for normal

Nothing was broken enough to fail. Everything was degraded enough to hurt.

What They Started Measuring (Outside LIMS)

1. Rework Ratio (Expanded Definition)

Not just formal reruns.

Included:

• Manual data cleanup
• Analyst-triggered retries without logged failure
• Calibration nudges mid-run

Finding:
For every 1 logged rerun, 3.4 invisible rework events occurred.

2. Abort-with-Value Events

Runs that:

• Were stopped early
• Produced some usable data
• Required human judgment to salvage

Finding:
18% of weekly runs fell into this gray zone — none visible to management.

3. Perceived Reliability Score (Technician-Led)

Technicians rated instruments weekly:

• “Would you trust this system on a critical sample today?”

Finding:
Several instruments rated “unreliable” for months while uptime stayed above 90%.

Availability ≠ trust.

4. Yellow State Persistence

They tracked how often systems sat in:

• Warning thresholds
• Soft QC alerts
• Drift-but-not-failure conditions

Finding:
The lab operated in a constant yellow state 62% of the time.

No alarms. Just erosion.

5. Cognitive Load Signals

Measured indirectly via:

• Override frequency
• Manual intervention density
• End-of-shift correction spikes

Finding:
Senior analysts were acting as living control systems.

Burnout was structural, not personal.

The Inflection Point

A major customer asked a simple question during renewal:

“Why do your results take longer to explain than to generate?”

Management couldn’t answer with data. That was the moment.

What Changed After Visibility

Within 90 days:

• Rework dropped 27% without new equipment
• One instrument was retired early — saving months of downstream cost
• Method updates prioritized fragility, not age
• Staffing conversations became evidence-based, not emotional

Most importantly:

• Yellow stopped being normal
• Trust was discussed explicitly
• Reliability became a shared language

The Real Lesson

LIMS told them what completed.

These metrics showed what it cost to complete.

Invisible work is still work.
Unlogged risk is still risk.
And a lab that’s always yellow is already late.

If your dashboards are calm but your people aren’t — measure that gap.

‍