Maintenance Mindset: Sample chain-of-custody – why good data still loses arguments

Most reliability failures do not start with bad intent. They start with a quiet assumption. The assumption is this: “If the lab result looks reasonable, the decision must be sound.”

The ASTM D4840 standard exists because that assumption is wrong. I have spent decades watching good engineers lose arguments they should have won – not because their analysis was incorrect, but because they could not prove the integrity of the sample that drove the conclusion. When that happens, the numbers stop mattering. The trend stops mattering. Experience stops mattering and what remains is doubt.

Custody is control, not a signature

ASTM D4840, the Sample Chain-of-Custody Procedures Standard, defines custody as “physical possession or control of a sample from collection through disposal.” That definition is precise on purpose. Custody is not who signed the form. Custody is who can testify, calmly and confidently, that no one could alter the sample without their knowledge. 

That is not a legal trick, but rather an engineering control. If you cannot state that with confidence, the data is already compromised. You may not know how or when, but the system no longer supports the conclusion.

Chain-of-custody failures in plants are rarely dramatic. In fact, most plants break the chain without realizing it. These failures are small, and familiar, and often well-intended.

• A sample bottle set on a bench during shift change.
• A label that almost matches the submission form.
• A cooler left unlocked because “it’s inside the plant.”
• Three people handling a sample because everyone is trying to be helpful.

ASTM D4840 reads like a catalog of these quiet failure modes, not because the standard is theoretical, but because it is written by people who have seen data collapse under scrutiny. Most of the time, nothing happens…until it does.

Chain-of-custody Is a reliability function

Here is where plant people often miss the point. Chain-of-custody is not about compliance. It is about decision survivability. Reliability engineering is not just about making the right call. It is about making a call that can survive challenge. Management challenge. Regulatory challenge. Legal challenge. Warranty challenge.

ASTM D4840 places chain-of-custody inside a larger concept called data defensibility. That matters. Custody alone does not save bad sampling or poor analysis. But without custody, even good work becomes fragile.

• A trend you cannot defend is not a trend.
• A report you cannot defend is not evidence.
• A decision you cannot defend is not control.

One of the most important recommendations in ASTM D4840 is assigning a single custodian at each stage of a sample’s life; 1) Field, 2) Transport, 3) Laboratory, 4) Analysis 5) Diagnosis. That is not paperwork discipline, that is accountability discipline.

Shared responsibility feels collaborative. In practice, it produces gaps. When something goes wrong, no one can say with certainty what happened, only what should have happened. A single custodian closes that gap. Someone who owns the sample. Someone who can say, “It was under my control.” That statement carries more weight than any spreadsheet.

ASTM D4840 also acknowledges electronic custody systems and LIMS platforms. It does not pretend they solve the problem. Digital systems document behavior; they do not correct it. If your process allows casual handoffs, unclear labeling, or unsecured storage, a LIMS system will faithfully record those weaknesses. Digital systems do not replace discipline, and technology does not absolve custody responsibility. It exposes it.

Finally, another quiet strength of ASTM D4840 is its honesty. It explicitly states that custody requirements are situational. Not every grease sample needs legal-grade rigor, but some do. Environmental discharges, warranty disputes, safety incidents, capital equipment failures cost money, reputation, or operating permission. The mistake plants make is not over-controlling everything, it is under-controlling the things that matter most.

Why this matters more than ever

Plants are under more scrutiny than at any point in my career. The data is questioned, decisions are audited, and assumptions are challenged after the fact. In that environment, chain-of-custody becomes invisible insurance. Keep in mind:
•    It does not make machines run better.
•    It does not improve chemistry.
•    It does not fix bearings.

What it does is keep your conclusions upright when someone pushes on them. ASTM D4840 is not a laboratory document hiding in a standards book, it is a reliability control that most plants already own and rarely use well. Like most reliability controls, you only miss it after it fails.