Latest Content
Latest Community Postings
Recent Blog Entries
Community Downloads
ChExpress Blog
Ankur's Tech Blog
Community Admin Blog
Energy Efficient Hot and Cold Water
Electrical Process Tomography
Biodiesel: The Road Ahead
Methanol Plant Capacity Enhancement
Plate and Frame Heat Exchangers: Preliminary Design
Compressor Surging Under Control
Plant and Equipment Wellness, Part 1: Observing Variability

Share this topic:

Mike Sondalini: Enterprise Asset Management Best-Practices Powered by Lifetime Reliability Solutions.
B Eng (Hons), MBA, CP Eng.  In an engineering and management career spanning 25 years he has held project engineering and maintenance management positions at the Swan Brewery and at Coogee Chemicals, a national Australian industrial and mining chemical manufacturer.  He is also a qualified mechanical tradesman.   Along with authoring numerous maintenance and industrial asset management publications sold on the Internet, he developed the UPTIME training series for chemical and process plant operators and maintainers.  His consultancy 'Lifetime Reliability Solutions' ( specialises in identifying manufacturing and production wastes and losses and solving them using proprietary optimization solutions known as ‘ACE’ (Accuracy Controlled Enterprise), 'DOCTOR' (Design Options and Costs Total Optimization Review) and ‘DAFT Costing’ (Defect and Failure True Costing).  He is a past Chairman of the WA Chapter of the Maintenance Engineering Society of Australia.  Mike is based in Perth, Western Australia. You can contact Mike by email, phone or fax using the details on his website contact page
Published January 8, 2008

Plant and Equipment Wellness:
Part 1 - Observing Variability

Example 1.1:  Inventory Replenishment Mayhem

The stock replenishment process involved the ocean shipment of raw material from a manufacturer to the company.  For some months prior the investigation the company had been running out of stock across a range of
products.  The impact on the company’s business was the inability to supply products on-time to their clients.  Their warehouse replenishment process was not able to maintain adequate stocks of product.  They were using their safety stock and not getting resupply quickly enough to reliably meet their client’s orders.  This was relayed back to them by annoyed clients through particularly strong correspondence and telephone calls.  The company had no appreciation of what was causing the stock-outs and requested that the situation be investigated.

asset5_2.gif (23511 bytes)
Figure 1.2: Frequency Plot of Product Stock-Out

The investigation began by collecting data on products stocked-out over the previous two years.  With that information a frequency plot of the products that had suffered stock-outs was developed on a spreadsheet.  Figure 1.2 shows the frequency plot.  From it can be identified time periods in the prior two years where the frequency of stock-outs had intensified.   The company was currently suffering increased number of stock-out over an increasing number of product ranges.  The frequency plot proved and confirmed the seriousness of the situation.

The next step was to determine what was causing the lack of supply.  For this it was necessary to look at the history of deliveries from the manufacturer.  Historical records of delivery dates were sourced and trended.  Figure 1.3 is a graph of a run chart for the delivery dates.  It shows a great deal of variability in the deliveries for the most recent months.  Basically the deliveries were not as regular as they historically were.  In recent months they were up to two weeks late when they should have been arriving weekly.

Further information on the situation was identified in Figure 1.4, which is a graph of the numbers of orders in each delivery.  This graph indicated that there was also variability in the amount of product being provided on each shipment.  Instead of have regular shipments of ten to eleven containers each delivery.  The ships were varying from four to twenty-seven containers per delivery.

asset5_3.gif (32151 bytes)
Figure 1.3: Ship Departure Dates
asset5_4.gif (28323 bytes)
Figure 1.4: Number of Containers on Each Ship

When inquiries were made it was found that the regular shipping line had one of its two ships in for a two month maintenance outage.  Where once there was regular weekly shipment, the only ship left on the run was now fortnightly.  To get product to the customer during the maintenance outage the manufacturer had started booking transport with various international shipping companies.  These ships had irregular departure schedules and only took numbers of sea containers they needed to fill the empty bays left after prior commitments were filled.  Sometime they took few containers and other times they took many.  The consequence of the irregular departure of the international carriers with either small or large amounts of product was the stock-outs suffered by the company.

The company suffered because of the irregular supply of goods from the manufacturer.  The irregularity was due to the high variability of international ocean shipping, further complicated by the feast-or-famine quantities of product on each ship.  Variability in the replenishment process had caused major disruption to the customer’s business.

In response to the temporary shipping problems the customer increased the amount of stock in-transit, which effectively increased their inventory levels until the second ship was repaired and returned to the weekly run.

To prevent stock-outs in future it would be useful to install pre-emptive monitoring of the manufacturer’s shipping arrangements to identify when a sea shipment did not leave on-time so a rail delivery could be booked instead.

The disruption of regular delivery to the company in Example 1.1 was caused by a ‘special cause’ event – the ship repairs.  A ‘special cause’ event is an extraordinary occurrence in a process that cannot be attributed to the process.  Had there been no ship repairs the customer would have been supplied normally each week via the usual process.  The ship failure was outside of the control of the replenishment process but it impacted badly on it.

Fluctuation that is due to the natural variability of a process is called ‘common cause’ variation.  The cross hair game was an example of the effects of common cause variation.  Where the pen landed depended on the behaviour of the process variables affecting the drop – steadiness of hand, accuracy over target, evenness of release, etc.  The spread of hit locations is normal for the cross hair game process.  To have the pen fall into the circle when dropped by hand has more to do with luck than with skill.  To always hit within the circle needs a change of process that has no element of luck, not an increase in the skills of the person doing the job.  Dropping a pen by human hand from a height of 300mm and expecting it always hit inside a 2mm circle is impossible, the common cause variability of that process is too great for the accuracy required.

There are many organisations trying to achieve impossible results using business and operating processes with ‘common cause’ variation that cannot reliably produce the performance they want.  Such businesses employ processes containing inherent volatility that naturally produce outcomes outside the business requirements.  Trying to manage an organisation with systems and processes that cannot achieve its business aims because they produce highly variable results is an exercise in futility that will cause great waste, distress for all involved and emotional burn-out for its managers.

Business process ‘common cause’ variability cannot be controlled unless changes are made in how the process operates.  In contrast, ‘special cause’ variability can be controlled by stopping the influence of the extraordinary event.  The effect of the ship repair in Example 1.1 could have been prevented by introducing other modes of transport, such as rail or road to replace the failed ship, if it was known that a delivery could not be made on-time.  ‘Special cause’ issues can be addressed simply by stopping them from happening.  But with ‘common cause’ issues nothing can be done to prevent them because they are inherent in the process.

It is the nature of every process to produce variation.  The challenge for business and operations processes is two-fold.   One is to create processes with only ‘natural’ variation and no ‘special cause’ variation.  Second is to select or develop processes with ‘natural’ variation well within the required performance.  This allows the organisation to focus mainly on stopping ‘special cause’ problems sure in the knowledge that the process itself is inherently stable and produces good product.

When a business or operating process no longer performs within its normal limits first look for a ‘special cause’ of the change.  Only after all ‘special causes’ are eliminated can you be sure that just natural ‘common cause’ variation remains.  If the ‘common cause’ variations are still too volatile you have justification for improving or changing the process.  By following that sequence you confirm if any special cause variations are masking the natural process variability and are producing effects to confuse the analysis.  If a ‘special cause’ is mistaken for a ‘common cause’ the wrong decisions will be made to address the problem.

So far we have seen examples of variability in a game and variability in the supply chain of an organisation.  Being able to get a picture of the variability brought a clearer appreciation of what was happening within the process.  It allowed powerful, relevant questions to be asked that led to a more profound understanding of the situation’s causes and their resolution.   There is great value to be gained when an organisation observes the variability of its business processes.  Once a ‘picture’ is available of how a process behaves, focused effort can be brought to bear on controlling its variability.  Example 1.2 is of a mining operation where the consensus was to invest a quarter of a billion dollars to expand production 50% when in fact it may have been unnecessary if production variability had first been addresses.


< Observing Variability

Example 2: The Hidden Factory >

By: Mike Sondalini, Enterprise Asset Management Columnist for


Subscribe to RSS headline updates from:
Powered by FeedBurner

Send this Page to a Friend

smalllogo.gif (4001 bytes)

  • Stay up to date on new content
  • Post questions and answers in our forums
  • Access downloads and attachments
  • Read member blogs and start your own blog
  • Connect with members via our friends feature
  • Receive and post status updates
Main Page

Terms of Use

Premium Content

Login  |  Signup   |  Learn More

Online Store

Engineering Software  |   Training Material  |  Continuing Education  | 
Maint Best Practices  |  Book Reviews

ChE Jobs

Free for Job Seekers

Free Articles

Newly Released  |  Popular Titles  |  Process Engineering  |  Process Design  | 
Process Safety | CPI Business  |  Engineered Writing   |  As I See It  |  Students


Sign up for the ChE Plus Newsletter

Message Boards

Interact with other professionals and students on our online message baords

Other Resources

Free Software  |  Physical Properties  |  Submit an Aritlce  |  Linking to Us  | 
Career Research  |  Why ChE?   |  Staff  |  Q & A's

Link Directories

ChE Links Technical Directory  |  Cheresources Partner Directory

Free Publications

Free Trade Publications from Cheresources

Advertise with Us

Advertise Your Company, Products, or Services

Contact Us

Send Us Your Feedback

The Chemical Engineers' Resource Page, 1442 Goswick Ridge Road, Midlothian, VA 23114, Email
All material is Copyright 2008,