How to frame a 60-day OEE pilot on one line

Why a pilot, rather than a big project

Faced with a performance-measurement project, many plants hesitate between two extremes: doing nothing, for fear of a heavy undertaking, or launching a full rollout across the entire site, with the risk of spending before having proven anything at all. The 60-day pilot is the third path, and by far the most reasonable: prove the value on one line before committing the plant.

This logic of controlled experimentation is familiar to industry. You do not roll out a new setting on every machine without testing it first. Measuring OEE is no exception: a well-framed pilot lets you answer, facts in hand, the only question that really matters, namely how much OEE this plant can genuinely recover, and at what cost.

Choosing the right pilot line

The choice of line determines the value of the pilot. The classic mistake is to instrument the best line, to reassure yourself, or the worst one, to dramatise the problem. Neither is representative. What you want is an average line with a clear stake: a demanding pace, frequent changeovers, or sensitive quality, so that the result transfers to the rest of the fleet.

A good pilot line also has an available team and engaged supervision. A pilot is not just a matter of a sensor: it is also a human approach, where operators discover their real performance and where supervision learns to steer with data. Choosing a line where these conditions are met maximises the chances of a convincing and reproducible result.

Setting a quantified objective

A pilot with no quantified objective settles nothing. Before you fit the sensor, you have to write down what you are trying to prove: reveal the gap between declared OEE and real OEE, identify the dominant losses, and aim for a measurable gain in OEE points over the duration of the pilot. That objective turns a simple observation into a decisive test.

The objective must stay realistic and verifiable. You do not promise a miracle figure; you commit to measuring reality and acting on the two or three main causes. It is precisely this methodological modesty that makes the pilot credible: it does not sell a dream, it establishes facts that leadership can then extrapolate with full confidence.

Installing without disrupting production

One of the great advantages of the pilot is how light it is to put in place. The sensor goes onto the machine in under an hour, with no PLC project, no production stop and no change to the existing IT. There is no overhaul of the information system, no deep integration: the measurement layer is autonomous and works on old machines as well as new ones.

This absence of disruption is decisive for lifting reservations. A production manager need not fear that the pilot will slow the line or complicate daily work. The first usable data arrives within 48 hours, and production continues exactly as before, except that you finally start to see what is really happening on it.

The milestones of the 60 days

An effective pilot follows a clear calendar. The first two weeks serve to collect the first data and compare real OEE with declared OEE: this is the moment of revelation, often striking. The following weeks are devoted to acting on the dominant losses identified, with concrete, verifiable adjustments.

The last weeks measure the gain obtained and prepare the decision. At the end of the 60 days, you have a quantified before/after on a real line, not a theoretical estimate. This rhythm, neither too short to conclude hastily nor too long to drag on, is calibrated to produce a decision grounded in facts.

Involving the teams from the start

A pilot is never just a matter of technology. Its success owes as much to ownership by the teams as to the quality of the measurement. When operators and supervisors understand that the data is not a surveillance tool but a way of making visible the losses they were suffering without being able to document them, buy-in comes naturally.

This is why it helps to name a champion on the production side, to set up a short data-reading ritual, and to share the first results widely. The pilot then becomes a collective project, whose lessons often go beyond the single instrumented line: it changes the way the whole plant talks about performance.

Concluding and deciding

At the end of the 60 days, two outcomes are possible, and both are methodological successes. Either the pilot has revealed a measurable gain, and the decision to extend is taken on concrete results, line by line. Or nothing actionable came out of it, and you simply stop, without having committed the plant or spent any capex.

This reversibility is what makes the pilot so rational. The risk is minimal, the pilot is free and supported, and the final decision rests on facts specific to the plant rather than on a sales promise. Hutchinson improved its OEE from 42% to 75% with the same headcount and machines, sensor installed in under an hour. That kind of result, measured before any mass rollout, is exactly what a well-framed pilot lets you establish objectively.

From pilot to rollout

When the pilot has proven its value, the extension happens gradually, line by line, building on the lessons of the first. You now know which losses to track, which ritual to install, how to bring the teams on board. The rollout is no longer a leap into the unknown but the replication of a validated model.

This controlled scale-up avoids the pitfall of big projects that bog down. Each newly instrumented line adds recovered capacity and reinforces the culture of steering with data. More than 450 plants across 30+ countries already monitor their OEE to the second with TeepTrak. The pilot is therefore not an end, but the reliable starting point of a lasting improvement approach.

The mistakes that make a pilot fail

A poorly framed pilot can disappoint, not because the measurement fails, but because the approach neglected the human dimension or the method. The classic mistakes are well known: choosing a non-representative line, setting no quantified objective, instrumenting without involving the team, or collecting the data without installing an action ritual. In each of these cases, the sensor works, but the pilot produces no decision.

Avoiding them requires no extra resources, only rigour in the framing. A well-chosen line, a written objective, a named champion, a short data-reading ritual: these few conditions, simple but decisive, make the difference between a pilot that sleeps in a dashboard and a pilot that leads to a rollout. The limiting factor is almost never the technology, it is the method.

What a well-run pilot really delivers

A 60-day pilot proves the value of measurement on one line before committing the whole plant. It assumes a representative line, a quantified objective, an installation in under an hour and a clear calendar of milestones. The teams are involved from the start, and the final decision is taken on facts specific to the plant rather than on a vendor’s claim.

Reversible and with no capex, it is the most rational way to launch an OEE approach. You spend nothing to find out what your real OEE is, you keep production running throughout, and you walk away from the 60 days with a number you can defend in front of leadership. Whatever the outcome, the plant learns something it did not know before about where its capacity is actually going.

FAQ

How long does an OEE pilot last?
60 days is enough to reveal the losses, act on the dominant causes and measure a quantified gain. This rhythm is calibrated to produce a decision grounded in facts, neither too short nor too long.

Do I have to stop production for the pilot?
No. The sensor goes on in under an hour, with no production stop and no change to the existing IT, on old machines as well as new ones. The first data arrives within 48 hours.

Which line should I choose for the pilot?
A representative line, neither the best nor the worst, with a clear stake (pace, frequent changeovers, quality), and an available team. The aim is a result that transfers to the rest of the fleet.

What happens if the pilot yields nothing?
You simply stop. The pilot is free, supported and reversible: no capex committed, no obligation to deploy. That is what makes the approach risk-free.

How do you move from pilot to rollout?
By extending gradually, line by line, replicating the validated model: losses to track, reading ritual, team involvement. The rollout becomes the replication of a success, not a gamble.