A competency matrix for dosing operators

“He’s done four batches with me. He’s fine.” That was the entire sign-off for a new operator at a site handling 60 m³ a day. Three weeks later he dosed a batch in grams per litre while the sheet was in kilograms per cubic metre, and the error went unnoticed because the number looked plausible on both scales.

Competency is not a feeling about a person. It is a set of tasks, done to a standard, evidenced, and rechecked on a schedule. A matrix is just the paper that makes that visible.

Columns that do actual work

Five columns are enough: task, standard, evidence, assessor, recheck date. The task is specific — “prepare a 10 m³ batch at a specified dose”, not “operate the mixer”. The standard is what good looks like, written so two people would judge it the same way. The evidence is what remains afterwards: a signed batch sheet, a jar test photo, a witnessed run. The assessor is a named person. The recheck date is a date, not “annually”.

The evidence column is where most matrices collapse. If the only evidence is a signature on a training record, the matrix records attendance rather than ability. A batch sheet with the operator’s initials, the calculated dose and the test result is evidence. A certificate is not.

Four levels, not two

Binary trained or untrained forces managers to overstate. Four levels work better: assisted, where the person performs the task with someone standing by; supervised, where someone checks the result; independent, where they run it and report; and trainer, where they can teach it and assess others.

Most sites need independent for routine dosing and trainer for at least two people per shift pattern. Without a trainer on each pattern, the night shift cannot qualify anyone, and the gap is filled by whoever is willing.

Assess without an exam

The most reliable assessment is a witnessed batch with a deliberate wrinkle. Ask the operator to dose a batch of known volume, then ask what they would change if the conductivity were 30 percent higher. Watch whether they reach for the test kit or for a rule of thumb. Then break something small — a blocked eductor, a wrong lot number on a bag — and see whether they notice before the batch is committed.

That takes forty minutes and tells you more than a multiple-choice paper. Record what you saw in one sentence. “Dosed 10 m³ at 2.4 kg/m³, checked conductivity first, spotted lot mismatch before opening” is a usable record.

The unit error that passed every check

The operator in the opening example had been trained by someone who used a different convention. The batch sheet said kg/m³, the verbal instruction had been “about two and a half grams a litre”, and both are roughly the same number at that dose. The mistake only surfaced when a batch was scaled up to 25 m³ and the arithmetic did not match the store issue.

The fix was not retraining alone. The site rewrote the batch sheet to show the calculation in a single line with units spelled out, added a store issue cross-check, and made the first ten independent batches of any new operator subject to supervisor sign-off. The error has not recurred, and the matrix now shows why: the evidence for independent level includes ten signed batches, not one conversation.

Refresher triggers

Time is the weakest trigger. Better ones: a change of polymer grade, a change in leachate source, a near miss, a return from long leave, or a new piece of equipment. Any of those should send the operator back to supervised level for one batch, which is a small cost and a real check.

Review the matrix at the same meeting where you review the dose records. The two documents describe the same thing from different angles, and when they disagree — good records, unqualified operators — one of them is wrong. The leachate solidification user guide sets out the batch steps worth assessing against.