SMED in Pharmaceutical Manufacturing: Where the Changeover Clock Actually Starts

The last tablet leaves the press at 14:47. The operator logs the batch end time. The changeover clock starts. Except it does not, not really. The cleaning crew will not arrive for another twenty minutes. The next batch paperwork is still waiting for a second signature. The new tooling is sitting in a cage three bays away because nobody confirmed the schedule change from the morning meeting. By the time the first bolt comes off, it is 15:31.

This is the gap that most pharmaceutical manufacturers never measure. They track the changeover. They do not track everything that happens before the changeover begins. In our experience working across the UK manufacturing community, this distinction is where SMED either delivers genuine cycle time reduction or quietly fails while producing a convincing-looking project report.

The SMED Clock Starts Before the Last Unit Leaves the Line

Shigeo Shingo’s original framework for Single Minute Exchange of Die separates internal activities, those that can only happen with the machine stopped, from external activities, those that can happen while the machine is still running. The entire discipline of SMED is about converting as much internal time as possible into external time.

In pharmaceutical manufacturing, the problem is that teams routinely misclassify a third category as internal time: waiting time that nobody owns. The machine is stopped. The operator is present. But the activity has not started because something upstream has not been resolved.

For example, in a solid dose facility running multiple SKUs, the bill of materials for the next batch may not be confirmed until QA signs off on the previous batch record. If that sign-off happens after the line stops, every minute of that wait appears inside the changeover window. It is not a changeover problem. It is a scheduling and documentation problem. SMED cannot fix it, but SMED gets blamed for the number.

The clock, properly understood, starts at the last good unit of the outgoing batch. Everything from that moment until the first good unit of the incoming batch is changeover time. That includes the wait, the walk, the hunt for paperwork, and the conversation with the shift supervisor about which product is actually running next.

Why Pharmaceutical Changeovers Are Structurally Harder Than Other Sectors

A press shop in the automotive supplier network can often run a changeover with two people and a trolley of pre-staged tooling. A pharmaceutical tablet compression or filling line carries regulatory obligations that create mandatory internal time you cannot eliminate, only compress.

GMP cleaning validation, for instance, requires documented evidence that residues from the previous product have been removed to an acceptable level before the next product runs. In a multi-product facility, that cleaning process may be the single largest block of time in the changeover window. It cannot be converted to external time because the machine must be stopped and accessible for cleaning to occur.

This is where lean manufacturing UK practitioners sometimes apply automotive SMED thinking without adjusting for the regulatory environment, and produce targets that are physically impossible to hit without a regulatory submission. The result is a demoralised team and a project that quietly gets shelved.

The more productive approach is to separate the changeover window into three distinct zones:

• Pre-stop preparation: everything that can be staged, confirmed, and positioned before the line stops
• Regulatory mandatory time: cleaning, verification, documentation that cannot be compressed without a change control
• Mechanical and setup time: tooling changes, format parts, line clearance checks that respond directly to SMED methodology

Most facilities have done reasonable work on the third zone. The first zone is where significant time remains. The second zone is where the conversation with QA and regulatory affairs needs to happen before any SMED project is scoped.

What Pre-Stop Preparation Actually Looks Like in Practice

In a packaging line changeover, pre-stop preparation might include confirming the next batch number and reconciling it against the printed batch record before the current run ends. It might mean staging all format parts on a dedicated shadow board adjacent to the line, verified against a checklist, so the operator does not leave the machine area during the changeover. It might mean ensuring the cleaning materials, in the correct concentrations and volumes for the specific product transition, are physically present at the point of use before the last unit is packed.

None of these activities require the machine to be stopped. All of them, if left undone, add minutes to the internal window.

Our team has observed that the most common failure mode is not that operators are slow or poorly trained. It is that the information they need to begin pre-stop preparation arrives too late, or not at all. The production schedule changes at 13:00 for a line that finishes at 14:45. The operator finds out at 14:50. That sequence makes pre-stop preparation structurally impossible.

In the UK manufacturing community, the facilities that have made the most consistent progress on changeover time are those that treated scheduling visibility as a lean tool, not a planning department problem. When operators can see what is running next with enough lead time to prepare, the changeover clock starts much earlier in a productive sense, even though the machine is still running.

The Measurement Problem That Keeps Changeover Times Artificially High

Many pharmaceutical plants measure OEE using data from their manufacturing execution system. The MES records line state: running, stopped, changeover. But the changeover state is often triggered manually, by an operator pressing a button or logging an activity. That moment of manual logging rarely coincides with the actual last good unit.

The result is a measured changeover time that is shorter than the real changeover time, because the pre-changeover waiting period sits in an ambiguous state, sometimes logged as downtime, sometimes as changeover, sometimes not captured at all.

This matters enormously for SMED projects because you cannot reduce time you have not measured. If your MES shows a mean changeover of, say, ninety minutes, and your actual elapsed time from last good unit to first good unit is consistently longer, your improvement target is wrong. You will optimise the visible portion and leave the largest opportunity untouched.

The practical fix is a paper or digital observation exercise, not an MES report. Someone with a stopwatch and a structured observation sheet stands at the line and records every activity from the moment the last unit is produced. They note what happens, who does it, when it starts, when it ends, and what caused any waiting. This is not sophisticated. It is the original time and motion study that underpins all of lean manufacturing, and it remains the most reliable way to see what is actually happening rather than what the system believes is happening.

In a manufacturing forum UK context, this kind of peer observation, where engineers from different facilities share what they have found in their own time studies, consistently surfaces patterns that no single site would identify alone. One site’s anomaly is another site’s standard practice, and that comparison is where improvement ideas come from.

Practical Takeaway: Map the Full Changeover Window This Week

Choose one line that has a changeover scheduled in the next five working days. Assign one person to observe the full event, starting from the moment the outgoing batch begins its final units. Ask them to record every activity, every wait, and every movement using a simple table: activity, start time, end time, person responsible, and one column for whether this activity could have started earlier.

Do not filter the observation through your MES data first. Let the observation stand on its own. When it is complete, map the activities on a timeline and draw a vertical line at the point where your MES says the changeover began. Everything to the left of that line is invisible time. That is where your first conversation needs to happen.

Bring the observation data to your next shift review. Ask the operators present whether the pattern is typical. In our experience, they will tell you it is, and they will also tell you exactly why the pre-stop preparation did not happen. The reasons are almost always systemic: late schedule confirmation, missing materials, documentation delays. None of them require a capital investment to address.

If you want to compare your findings against what other UK pharmaceutical and industrial operators are seeing in their facilities, that conversation is exactly what we built leanIq to support. Post your observation data, ask what others have tried, and find out whether the pattern you are seeing is specific to your site or common across the sector.

Start the observation before the changeover begins. That is where the time is.