Right First Time: measuring RFT when rework goes unrecorded

Walk through most assembly areas on a Tuesday afternoon and you will see it: a fitter quietly redoing a sub-assembly that came back from the previous shift. Nobody fills in a form. Nobody books a job number against it. It just gets sorted, and the line moves on. Ask the quality manager what the right first time rate is and they will quote you a figure that has never once included that fitter’s afternoon.

That gap, between what the number says and what the floor knows, is where most RFT programmes quietly fall apart.

Why the figure is almost always too high

Right first time should be straightforward: how many units came off the line without needing any further work? In practice, the calculation depends entirely on what gets recorded. If your shop floor culture treats informal rework as “just getting on with it,” the jobs never hit the system, and the RFT figure drifts upward, disconnected from reality.

This matters because a flattering RFT number kills the urgency to investigate root causes. If quality looks fine on paper, there is no pressure to find out why the third shift keeps sending parts back, or why one cell produces twice the rework of the one next to it.

What counts as rework, and what does not

Before you can measure RFT properly, you need a working definition that everyone agrees on. Rework is any operation performed on a part or assembly that would not have been needed if the job had been done correctly the first time. That includes:

• Trimming a weld that came back from inspection
• Reapplying a torque that was not achieved first time
• Correcting a label or documentation error before the unit can ship
• Any touch that re-enters a completed operation

It does not include planned finishing operations, design-specified secondary processes, or customer-requested modifications. The distinction matters. If you lump planned finishing into rework, you inflate the problem. If you exclude genuine defect corrections, you hide it.

Getting the number without adding paperwork

The resistance to capturing rework is almost always about paperwork. Operators on a fast line are not going to stop and raise an NCR every time they spend ten minutes sorting out a minor fault. If you make it that hard, they will not do it.

A few approaches that have worked on real production floors:

Tally boards at the cell. A simple laminated sheet at each station with columns for the main defect types. The operator marks a tally. The team leader totals it at shift end and enters one number. No NCR, no long description, no supervisor signature. You lose traceability, but you gain honest data.

Exit gate checks. A quality checkpoint at the end of each cell or process step, separate from the operator. The checker logs pass or fail. Fails go back. The log gives you a daily RFT figure per cell without any reliance on operators self-reporting. The downside is it costs a person, or at least part of a person’s time.

Cycle time variance. If your line is balanced, an operation that regularly runs twenty per cent over standard cycle time is a signal that something is being reworked or repeated. You can pull this from most MES systems without adding any new data entry. It will not tell you what the problem is, but it will tell you where to look.

The honest advice is to use two of these together. Tally boards give you a quick, cheap baseline. Cycle time variance gives you independent corroboration. If both point at the same cell, you have something worth investigating.

Turning a better number into a better process

Capturing rework is only useful if you do something with it. Once you have a more honest RFT figure, the next step is a regular review, ideally weekly, where the team looks at where rework is concentrated. Not to blame anyone, but to ask whether the issue is a tool, a fixture, a material, an instruction, or a training gap.

The pattern usually becomes obvious quickly. Three of the last five weeks, rework has clustered on the same operation. That is not bad luck. That is a process that needs attention.

If your losses are spread across too many categories to prioritise, the losses calculator can help you put rough values against each one and decide where to start. And if the rework problem turns out to be rooted in how the work area is set up and maintained, the 5S audit tool gives you a structured way to assess and track that side of things.

RFT is one of the simplest quality metrics there is. The difficulty is not the maths. It is creating a floor culture where the real number gets recorded, not the number that makes the weekly report look acceptable.