Defect Elimination

Reliability Operating Centre

Stop paying for the same failure twice. We find what keeps breaking, work out why, and stay on it until the failure mode is gone rather than just repaired faster.

Where defect elimination usually breaks down

Almost every operation has a defect elimination process. Very few have one that is changing the failure rate.

Only the big events get investigated

An incident triggers an investigation. A twenty minute stoppage does not. But the stoppage that happens fourteen times a year usually costs more than the event that got the attention, and nobody has ever added it up.

Actions are raised, then quietly age

The investigation produces recommendations. They go into a register. The register grows faster than it closes, people stop expecting anything to happen, and attendance at the meeting drops.

Nobody checks whether it worked

The action is marked complete and everyone moves on. Six months later the same failure appears and is treated as a new event, because no one went back to the data to see whether the fix actually changed anything.

A defect register that grows faster than it closes is worse than no register. It turns a reliability problem into a credibility problem.

What we do

A closed loop, run every month, whether or not anything has gone wrong lately.

Rank the bad actors by cost, not by countWe pull work order and downtime history and rank by what each failure mode actually costs you in lost production and maintenance spend. The list this produces is rarely the list the site expected.

Investigate with a structured methodApollo Root Cause Analysis, facilitated by a certified trainer. Structured cause and effect rather than a workshop that arrives at whatever the loudest person in the room believed at the start.

Turn causes into owned, dated actionsEvery cause gets a specific action, a named owner and a date. Actions that cannot be owned do not go on the register, because an unowned action is a decision not to act, recorded as though it were a plan.

Chase closure and report both numbersActions raised and actions closed, reported separately every month. We are accountable for the first. You are accountable for the second. Neither of us gets to hide inside a single blended figure.

Verify the failure mode actually stoppedMOST OFTEN SKIPPEDThree and six months after closure we go back to the data and test whether the failure rate changed. You get told which fixes worked, which did nothing, and which need another pass. This is the only step that proves the programme is worth running. What the removed failure modes were costing carries into loss accounting.

What you get

Deliverable	What it contains	Cadence
Bad actor register	Failure modes ranked by annualised cost, with the data behind the ranking	Quarterly refresh
Facilitated investigations	Apollo RCA on nominated trigger events, with cause and effect chart and evidence	Within 5 business days of trigger
Defect register	Causes, actions, owners, dates and status, maintained by us not by you	Live
Closure and verification report	Raised versus closed, plus 3 and 6 month verification on prior fixes	Monthly
Elimination value	What the removed failure modes were costing, carried into the loss model	Monthly

Frequently asked questions

What is defect elimination?

A closed-loop process that finds the failures costing you the most, works out why they happen, fixes the cause, and then verifies against your own data that the failure mode actually stopped. It differs from repair in that the goal is that the failure never returns.

How is defect elimination different from root cause analysis?

Root cause analysis is one step inside it — the investigation. Defect elimination is the loop around that step: ranking bad actors by cost, turning causes into owned and dated actions, chasing closure, and going back to the data at three and six months to verify the failure rate changed.

Which failures are worth investigating?

The ones that cost the most per year, which is rarely the list a site expects. A twenty-minute stoppage that happens fourteen times a year usually costs more than the single big event that got the investigation. We rank by annualised cost from your work order and downtime history, not by event size.

How do you know the programme is working?

Two numbers reported separately every month — actions raised and actions closed — plus verification at three and six months that closed actions actually changed the failure rate. A register that grows faster than it closes is the failure mode of defect elimination itself.

Find out what your recurring failures actually cost

The bad actor ranking comes out of the Foundation phase. Most sites are surprised by what is at the top.

Book a fit and data readiness check