When enough is… enough

cho-in-azoneThis is a behavior I’ve noticed quite often in food industry, in chemical or pharmaceutical plants: cleaning and sanitation processes (mainly their duration) are extended beyond the standard procedures at the expense of costs and production capacity.

Fear of harming

In the regulatory-constraint industries like food, chemical or pharma, people on shopfloor are trained and qualified to perform cleaning and sanitation operations. They follow procedures and work instructions, based on standards.

They usually also have frequent training about the importance of sanitation or sterilization and the possible consequences if badly done. Working in food, healthcare or pharma is embracing the sacred mission to bring something good, to cure or relieve customers and/or patients and do everything to prevent hurting them in any way.

They are also reminded what consequences for the organization in case of problem beyond failing to: losing the customers’/patients trust, losing the licence to produce, being sued, being exposed to scandals…scary enough for shopfloor people to take things seriously.

Yet the people on shopfloor seldom have the scientific background to fully understand what is required for good sanitation or sterilization, when doing more is useless or even counterproductive. They also are often left on their own, without expert supervisors to reassure them, answer possible question or take decisions in case of doubt.

Furthermore, the results of sanitation/sterilization is most often only known after a sample of rinsing water or the swabbing of the tool/equipment has been analyzed by some remote lab.

Fearing to harm the organization, or worse the customers / patients or possibly to have to go over the whole lengthy sanitation process again if it is not satisfactory, the sanitation is performed longer than procedures require it. This is base on the belief the more the better.

This seemingly logical and well-intentioned assumption is never challenged, leading to waste detergents, acids, water… and time, simply because over-sanitation is not noticed by management.

Changeovers are even longer

Changeovers in such environments can be long and painstaking due to regulatory constraints and all the paperwork associated. Ignoring the over-sanitation habits can extend the changeover duration even more.

Besides adding costs for no additional value, the additional time spent on sanitation may be needed on critical equipment (bottlenecks) and the time lost will not only never be recovered but the true cost is to be counted in minutes of turnover. And this one can be skyrocketing!


When looking for additional productive capacity or a way to get more out of the current process, check the changeovers’ content and take a closer look on sanitation.

Give the shopfloor personnel clear indication when enough is enough, without risk to harm anybody nor to endanger quality. If necessary, have a real qualified subject matter expert attending these critical phases, ready to support the team and answer any question.

Not only will it take some concerns off the team, but may be a great payback in terms of additional yield.

Feel free to share your thoughts and experience in the comments and share the post if you liked it.

View Christian HOHMANN's profile on LinkedIn

Why SMED is quick win in pharma – Episode III

Improving changeovers in pharma industry is a relatively quick and easy way to… quick wins, faster and easier than usually assumed. This series tells you why.

Episode III: How to and Why it works

In the previous Episode I explained the background of the lag of many traditional pharma makers in regards to Industry Best Practices (mainly Lean) and operational performance. I highlighted faster changeovers as leverage for recovering wasted capacity and improve performances. In Episode II I gave examples of  gains that can be expected and why.
In this Episode, I explain how and why it works.

The first approach is the least “risky” one, which will not change anything, thus not jeopardizing compliance to procedures nor quality. It is based on the fact that changeovers take frequently more time than initially allocated and the assumption that the changeover procedure is sound and the time allocation is reasonable, i.e. changeover doable during the allocated time.

Reminder: pharma is regulation-constrained

As I explained in Episode II, in pharma industry all steps of a changeover are prescribed in procedures and traced. Lots of information and proofs are captured, paperwork filled because a great deal of these procedures are mandatory in order to comply with Good Manufacturing Practices (GMP) and/or local regulation.

This means there is a solid reference base available against which to compare actual way of changing over as well as a reference time allocation (standard time).

In other, lesser regulation-constrained industries, such detailed procedures and capture of data might simply not exist or will be much “lighter”.

The procedure fallacy

Industrial engineering, quality assurance and management assume that when procedures are written and approved, they are the tables of the law that personnel will follow thoroughly.

This is not always the case. People on shopfloor are pretty much on their own as management is more likely sitting behind a screen, on a desk in a remote location. So there is room for doing things slightly differently than the procedure prescribes. Often it is about swapping some tasks’ order because it is more convenient or the people’s preference.

Procedures are written for the standard (perfect world) case and won’t help if something unexpected happens, e.g. some material is not available or late. As unexpected events, big or small are likely to happen, people in charge of changeover will have to adapt or wait for instructions.

Doing things in a different order and/or in a different way will impact the changeover duration. It might speed up or delay it. The problem is that people on shopfloor may not have sufficient knowledge/insight about the possible overall impact, like negative side effects, of their even so small changes.

Chances are that procedures will be followed globally but variations will be found in the details of execution. That’s one of the reasons for the differences between allocated time and actual changeover duration.

And chances are that actual duration exceeds the usually generous allocated time, reducing the productive capacity, overall effectiveness and efficiency.

The easy way to reduce changeover duration

Let’s be clear: the easiest way to reduce changeover duration in pharma industry will probably not be the most rewarding one in terms of production capacity recovery, but… it’s the easiest one.

It is easy because it is only about sticking to already agreed procedure and standard time, thus no risk assessment nor quality assurance validation required.

This approach is based on following assumptions:

  • Changeover procedure is sound
  • Allocated changeover time(Standard Time)  is reasonable, i.e. changeover can be done within allocated time
  • Excessive duration changeovers outnumber the shorter ones, hence there is a net capacity loss when summed up

Step 1: Gather data about changeover duration. If there was no data capture or what was in place does not serve your purpose, create a form and capture what data is necessary

Step 2: Start analyzing. Look for trends, correlations and if possible causation

Step 3: Display a graph with changeover durations compared to standard time. Update it real time. The simple display of the graph and the information to shopfloor teams that changeovers durations will be monitored is enough to improve the situation, because now there is some management’s attention on it.

Step 4: Go see, ask why and show respect (this is a Lean management mantra). In other words, go and spend some time observing reality on the shopfloor (gemba). Do not hesitate to ask why this or that to people, they are the Subject Matter Experts. While asking, do not lecture but listen truly, without judgement and without disturbing operations. Try to find the root causes of good AND poor performance.

Step 5: suggest or make the necessary changes (without compromising GMP/safety/quality rules) in order to reduce the duration. Chances are the improvement will require someone with the necessary authority and know-how to coordinate the whole changeover, from new material delivery to leftovers sending back to storage place, including paperwork and human resources allocation to roles. Stress the necessity not to exceed allocated time.

Step 6: when necessary, run problem solving kaizen events. Always have at least one operational personnel involved.

Step 7: Keep capturing data, analysing it and understand the causes of longer AND shorter changeovers. Changeovers that will take exactly the allocated time are highly suspect, but will be dealt later. Watch for trend: after a short while, changeovers should seldom exceed the Standard Time, although some accidents may happen.

Step 8: iterate to step 4.


The last time I used this soft approach (2015), we could recover the equivalent of one week of productive capacity within a period of 3 months, roughly 8% and about 170,000 additional units made available for sales. This was done at zero additional costs as all we needed was better organization, committed and refocused people and more management’s attention.

Of course this improvement was sustainable.

The changeovers done way under Standard Time were the proof of the excessive allocation and/or potential better way to proceed. With the relevant data, it was easy to convince management to have the procedure and Standard Time updated, giving the opportunity to improve further using SMED methodology.

If you liked this post, tell it, share it!

Bandeau_CH40_smlView Christian HOHMANN's profile on LinkedIn

Why SMED is quick win in pharma – Episode II

Improving changeovers in pharma industry is a relatively quick and easy way to quick wins, faster and easier than usually assumed. This series tells you why.

Episode II: What Gain? Why?

In the previous Episode I explained the background of the lag of many traditional pharma makers in regards to Industry Best Practices (mainly Lean) and operational performance. I highlighted faster changeovers as leverage for recovering wasted capacity and improve performances.

In this Episode, I explain what gain can be expected and why.

A changeover – as the name tells – is changing over from one finishing series or production batch to the next one. This can mean completely resetting a production line, changing formats or “simply” change the batch and product-related information and paperwork.

Changeover in pharma is more thorough than in lesser regulated activities as most often the lines must be totally cleared from anything related to the finished batch, sometimes cleansed, inspected, approved, reset and restarted.

All steps are traced, lots of information and proofs are captured, paperwork filled. A great deal of these procedures are mandatory in order to comply with Good Manufacturing Practices (GMP) and/or local regulation.

Many (additional or redundant) steps and procedures are the maker’s choice, often as a countermeasure of former problems or deviations. They can also be a consequence of regulation misinterpretation or fear not to be compliant. I will not discuss challenging these self-imposed constraints in this post, but please note these self-inflicted constraints are improvement potentials per se.

The changeover procedure is therefore a relatively lengthy one, closely monitored (mostly afterwards and through paperwork though) by Quality Assurance (QA).

As the big fear is to leave something from previous batch contaminating the new one (it can be a physical part, an information leaflet, etc. in case of packaging) the changeover durations are generously allocated, based on the assumption the more time to perform the changeover, the fewer the risks.

I emphasize; allocated time is much more than strictly necessary to changeover and remaining compliant to procedures and requirements.

On top of that, nobody in management would dare stressing the operators to speed up a changeover in fear of making them forget something or ending up with a quality deviation.

The later leads shop floor operators to extend the changeover duration beyond already generous time allocation, wasting even more productive capacity.

The reasons for the drift may vary:

  • On the positive side, the constant fear to do something wrong or forget something.
  • On the negative side I noticed frequent lack of discipline: operators do not stick to procedures and management is not following-up and monitoring closely enough.

Over time the extended duration tends to be accepted as the new standard, planners including the actual changeover time in their schedule. Nobody questions, nobody challenges.

This is Parkinson’s law: a task will always take all allocated time. Extend the allocation and the task will never again finish before the new allocated time.


During changeover production is stopped, which means changeover is non-productive and reduces the production capacity. Changeovers tend to be more frequent as batch sizes shrink, so the challenge is to change over quick in order to minimize the production stop.

Changeover duration often exceed what is really necessary to perform a changeover in good conditions, without taking chances with safety nor quality nor GMP/regulatory compliance.

The main reasons for this are:

  • Generous time allocation
  • Lack of rigor / discipline
  • Immaturity regarding industrial best practices (Lean, SMED…)

Reducing changeovers durations is a way to recover recover wasted capacity and improve productivity by earning more output with the same resources.

Changeovers waste capacity

How Much?

I just explained why a significant part of wasted capacity can be recovered, but how much is this?

Detailed data are not always available to estimate the recovery potential, therefore a rule-of-thumb can come in handy.

Experience told me the “Lean rule-of-thirds”, which means about 30% (at least) of wastes can be turned into savings. This is relatively scale-invariant but of course much easier to achieve at the beginning of a continuous improvement journey than later, in a more mature state.

Thus, a changeover duration of 2 hours or 120mn could be reduced fairly easily to 80mn or 1hr20, especially if the drift from initial standard time happened.

Example in primary or secondary pharma packaging

Based on the same example as above, with a production rate of 60 units a minute (not very high-speed), forty minute capacity recovered means 40mn x 60 units = 2,400 additional units.

Changeovers may occur several times in a shift or at least in a 24 hour timeframe, multiplying the gain.

Now with highly automated, high-speed equipment, these additional gains may be far higher. Think about how fast tablets are packed into blisters or vials filled with liquid.

Big money

Even after patents have fallen into public domain, some original molecules still sell well, especially when customers are somewhat reluctant to use generics. Therefore if each additional unit yields a net profit of one monetary unit ($, €, £, etc.), which is not uncommon for ex-blockbusters, the additional profit is worth the improvement effort and the Return On Investment fairly quick.

Next Episode: How to and Why it works?

If you liked this post, let it know and share it!


Bandeau_CH36View Christian HOHMANN's profile on LinkedIn

You are probably wrong when identifying your bottleneck

Things moved on since Eli Goldratt’s revealed Theory of Constraints through his business novel and bestseller “the Goal”. Most of today’s processes are more complex than 30 years ago: supply chains are stretched over the globe, new products are launched more frequently and batches are changed more often, among others. Thus identifying the bottleneck or capacity constraint is more difficult and must also be redone more often.

In this video, Philip Marris shares some “lessons from the road”, real case lessons learned from more than 30 cases in the last 10 years. Surprisingly, if companies were rather good identifying their bottleneck, it turned out that now, in 80% of cases companies are wrong about where their capacity constraint is.

Main learnings are:

  • companies are often confusing where the constraint should be with where it really is
  • ERP data is not a reliable way to identify constraints
  • companies tend to have an outdated / obsolete analysis of the situation
  • new quality requirements often create new capacity constraints
  • (bad) cost cutting decisions create new bad constraints

As a conclusion, he points out that being wrong about where a company’s constraints are is good news since it implies that there are significant opportunities remaining to improve performance drastically and rapidly.

If you liked this post, let it know and share it!

About the author

View Christian HOHMANN's profile on LinkedIn

Why SMED is quick win in pharma – Episode I

Improving changeovers in pharma industry is a relatively quick and easy way to quick wins, faster and easier than usually assumed. This series tells you why.

Episode I: The Background

There are roughly two cases to consider when addressing operational performance in pharma industry :

  1. the traditional (big) ones
  2. the tollers and generic makers.

Traditional pharma makers used* to be protected by their patents, granting them several years (about 20) of monopoly in order to payoff the huge initial investments in R&D.

*Many blockbusters drugs have lost patent protection, allowing generic makers to produce and sell them much cheaper. The dramatic drop in revenue for the original makers brought up the term “patent cliff”, a metaphor for the sudden fall of incomes. For manufacturers without other blockbusters to compensate, it is a sudden exposure to leaner and meaner competitors.

Tollers (or subcontractors) and generic makers are not investing in hazardous R&D and will not have patent protection in return. They manufacture for others or “copy” drugs after they fell into public domain and sell at much lower price.

On the operational side, during the blockbusters years – when huge incomes were secured – the (big) pharmas did not care much about capacity exploitation and efficiency. When more capacity was required, new equipment, lines or even whole factories were bought/built.

The payoff was such that it was faster to setup a new facility and run it at relatively low productivity level than to try to improve already installed capacities.

Falling off the cliff

The consequences years after, once most of the blockbusters fell off the patent cliff into public domain and related revenues plummeted, are:

  • huge overcapacities, often whole plants,
  • low productivity* compared to other industries,
  • lower maturity regarding industrial best practices (e.g. Lean Manufacturing, Lean Management),
  • no real sense of urgency** to improve in operations,
  • lack of agility,
  • a “sudden” and unprepared facing of leaner and meaner competitors, meaning ordinary competition,

*OEE (Overall Equipment Effectiveness), is often in the range 15-35% when in other industries it is rather in the 50-65% range.

**this lax posture of well paid pharma workers, even when “the platform is finally burning”, make them the “spoiled children” from the perspective of others, struggling in harsher competition with lesser compensation.

Tollers and generic makers must be lean and efficient at once because of their business model. They don’t have secured incomes nor a patent shielded-off competition. They compete with makers in low wage countries, with lower sales prices, hence lower profit per unit.

They nevertheless have to invest and carry costs related to regulatory compliance.

Given the circumstances these makers understood much sooner the importance to close the gap with Best Practices in more mature industries. This does not mean that generic makers are all best in class, but they had powerful and early drivers to push them up the performance ladder.

Now that the scene is set, what is the link to quick changeovers?

For traditional pharma experiencing normal competition, investments are no more that easy they once were. Given the many remaining overcapacities, when delivery is poor, the solution is no more installing additional capacity but make better use of the one installed.

When looking closer how the installed capacity is wasted, changeover duration most often pops up as a major cause. And as changeovers tend to multiply with the shortening of runs and smaller, more frequent batches, they become good candidates for capacity / performance improvement.

In the next Episode, we’ll see how much can be gained and why.

Bandeau_CH40_smlView Christian HOHMANN's profile on LinkedIn

Your next bottleneck is elsewhere (and in the future)

Theory of Constraints provides the five focusing steps, an iterative improvement process which aims to focus efforts on the sole system constraint (often a bottleneck).

These five steps are:

  1. Identify the constraint (bottleneck)
  2. Exploit the constraint; improve capacity utilization
  3. Subordinate all non-constraint resources to the constraint
  4. Increase the capacity of constraint if relevant
  5. Repeat step 1 if the constraint has changed

The final step is an invitation to continuous improvement, but also a warning: do not let inertia, passivity and acceptance of the status quo become the constraint.

Yet one other aspect of this warning remains mostly unknown.

While teams work hard to exploit the bottleneck resource and recover some wasted capacity, they do not anticipate that the next bottleneck is located elsewhere in the future.

Most teams working to elevate a capacity constraint do not imagine that the additional capacity that will be recovered requires immediate anticipated loading.

Indeed, most of the time, once the goal is reached and the bottleneck is no longer the constraint, they “expect” to see another bottleneck emerge in their area, as if they were playing whack-a-mole; hit one, wait for the next to pop-up.

Chance are that the next bottleneck will probably not be found within their perimeter. The next bottleneck can be upstreams, in another department or with some supplier.

The next bottleneck will instead most likely be found either in development, engineering, marketing or sales. And it will come as a surprise due to lack of anticipation.

The next bottleneck may be the order book, as sales nor marketing did not anticipate the loading of the recovered capacity. It may be development, unable to bring forward the launch of the next product.

It lies in the future is a warning about the necessity to anticipate it and the probable time lag before the anticipated efforts pay off.

Bandeau_CH38View Christian HOHMANN's profile on LinkedIn

Recover wasted capacity with SMED

SMED is a systematic approach to quick changeovers in order to minimize machine downtime. It is welcome to recover wasted capacity on a bottleneck resource.

  • A bottleneck is a resource with a capacity, in average, lower or equal to customer’s demand. A true bottleneck runs 24 hours 7 days a week and still cannot supply what is required.
  • A bottleneck is usually very expensive and/or difficult to get. If this is not the case, the solution is obvious: buy additional capacity!

More about bottlenecks.

As a bottleneck is the limiting factor hindering getting more out of the process, hence making money in case of for-profit organizations, and as the bottleneck resource is not easy to duplicate, the best option is to get more through it.

Analyzing the capacity

Almost every resource has a maximum capacity that can seldom be utilized, because of some reasons that waste some of it. The resource’s capacity can be can be depicted by a lab beaker as shown below, with a maximum capacity of 100 (%).

Several events during production will hinder the utilization of the full capacity:

  • Machine downtime
  • Lack of supplies
  • Human rest time
  • Etc.

These events are wasting the precious capacity, just as if our beaker was leaking. Changeovers usually are an important cause of capacity loss.

Reducing drastically the changeover duration with SMED is like fixing the leak and recovering a part of the wasted capacity.

In case of a bottleneck this is very important because the recovered capacity is converted in additional sales without (or delaying) investment in scarce or highly expensive additional resource.


View Christian HOHMANN's profile on LinkedIn

SMED explained while doing laundry – part two

In part one I explained SMED is a systematic approach to quick changeovers in order to minimize machine downtime.

In order to explain what SMED is good for to non-specialists, I used the laundry example, in which the mundane washing machine is a very precious resource and it should wash (add value) as much as possible in 24 hours a day. Conversely, such a precious resource should be protected from any downtime, especially when changing (laundry) batch.

After describing a typical lengthy and far from efficient changeover, this second episode will show how to drastically improve the performance by reducing the changeover duration.

Distinguish internal and external setup

The very first thing to understand when striving to reduce changeover duration is to distinguish internal and external setup operations.

Internal Setup (IS) operations are done within the machine or so close to it that safety requires to have it stopped to achieve them. Typically changing a tool fixed on a machine or adjustment inside the machine.

External Setup (ES) operations either have nothing to do with the machine itself, like filling a tracking record form or bringing raw material from warehouse near to the machine, either can be achieved without endangering oneself. ES operation can be achieved while the machine keeps running.

Many times, some operations and task are first seen as External Setups, but after closer analysis are totally unnecessary. This happens often when countermeasures to some old problem have not been suppressed after the problem was settled.

Internal and external setup with laundry

Back to our laundry problem; what are internal and external setups?

Considering the chart of a typical changeover as displayed hereunder,

The time wasted until “operator” noticed the end of washing cycle is no Internal Setup nor is it External, it’s just plain waste and must not happen.

“Searching for empty clean basket” to receive the washed laundry and “searching for detergent” and “sorting laundry” should not happen at the expense of machine time. These operations should be included in preparation prior to changeover, which is External Setup.

If work environment, here laundry room, is well organized and tidy, searching for items like basket and detergent should be unnecessary. Even so in preparation – External Setup – this time is not impeding the washing machine performance, it can be saved for more valuable occupation.

The changeover duration after separating ES and IS should be drastically reduced to almost bare minimum. Almost, because the remaining IS operation are likely to be optimized (reduced).

The changeover process after improvement should look like this:

This simplified laundry example is very similar to the changeover duration optimization done in production lines, machining cells, etc.

Important notice

When working to reduce the machine downtime during changeovers, most of people try to improve some technical aspect like fittings, fastenings, jigs and so on. As this fictional example shows, most of the improvement potential is found in organization.

>Read more about SMED

About the authorView Christian HOHMANN's profile on LinkedIn

SMED explained while doing laundry – Part one

SMED is a systematic approach to quick changeovers in order to minimize machine downtime.

>Lisez-moi en français

To explain what SMED is good for to non-specialists discovering it, I use the laundry example.

Let’s imagine the mundane washing machine is a very precious resource and it should wash (add value) as much as possible in 24 hours a day.

The machine stops at the end of a washing cycle and there’s no choice but swapping washed laundry with a new batch to-be-washed laundry. This exchange is equivalent to production changeover. In order to minimize the washing machine stoppage, the laundry swap has to be as swift as possible.

But before designing a laundry cannon to shoot the clothes high velocity into machine’s drum, let us observe how a laundry swap is done (go see).

The usual changeover cycle

When the machine stops at the end of a cycle, the “operator” has to notice it and be ready for action. If the machine does not signal itself and/or the operator is not waiting for this moment, the reaction time until noticing and acting is wasted.

The first thing to do is to remove the washed laundry from the drum. This clean laundry should be protected against stains and put in a clean empty basket for example.

Ah! Empty basket? Where is the empty basket? The time searching for the empty basket with arms full of clean laundry is lost for washing as the machine sits idle waiting to be fed another batch.

When the clean laundry is eventually dropped into a clean empty basket, and assuming the “operator” does not waste additional washing machine time to further take care of clean laundry, he’ll has to come back to the machine to load it with the next batch of laundry to be washed.

At this point boys have to understand to their amazement that delicate, white, dark and light laundry have to be sorted and washed separately, except for those who like their white shirt turn light pink thanks their daughter/sister/wife’s blouse. Girls usually know this.

If the to-be-washed laundry was not sorted beforehand, the time to sort it lets the machine idle and wasting more time.

Finally, one part of the sorted laundry is thrown into the drum, a gesture that could have be done in a softer manner, as the violent acceleration of clothes does not compensate much of the time lost until that point.

The changeover is not finished as long as the machine is not restarted. And before restarting, it should be refilled with detergent.

Detergent? Where’s the detergent?

While searching for detergent, the machine… sits idle.

When detergent is filled and machine started, the changeover is done.

This imaginary scenario is not that uncommon in households nor is the equivalent not uncommon in factories.

Instead of washing machine we’ll find machine tools or any production or test equipment. Instead of laundry it’s raw material or parts and instead of detergent, operators may look for tools, templates, jigs, plans or whatever should be handy and is not.

In the next post of this series, we’ll see how the washing machine downtime can be drastically reduced!

About the authorView Christian HOHMANN's profile on LinkedIn