Ensuring the Longevity and
Successful Implementation Of Your Advanced Maintenance
Strategy (Predictive Maintenance)
Opinion and fact by Michael
Korf-Consultant-National Reliability Systems
Click here for a
print friendly 160k pdf version
This paper is focused on identifying and preventing the
most common causes of advanced maintenance strategy
implementation failure and more specifically predictive
maintenance program implementations. The ideas
presented are applicable, in most cases, to any advanced
maintenance strategy implementation. A recent report
stated the following as the top 2 reasons why new
business opportunities fail.
Poor management of financial activities
Lack of management competence or experience
design and implementation of any advanced maintenance
strategy should be viewed as the start-up of a new
business. This presentation/paper looks at the
categories of People, Processes, Technology and
Reporting and the root-cause failure mechanisms that
hinder their progress. I believe that the major issues
regarding failed advanced maintenance strategy
implementations fit into these categories. Fishbone
root-cause analysis diagrams are shown to identify
leading problems. ROI/CB methods are specifically
discussed to ensure proper reporting for management.
Recommendations are also offered to predict and prevent
these failure mechanisms from occurring, so that
long-term program longevity can be assured.
Many of the activities
outlined below that lead to business failure are
prevalent in our facilities. But what are the
root-causes for these problems? Poor management of
financial activities can typically be boiled down to
poor reporting practices. Lack of experience can usually
be addressed through improved training. Economic
conditions typically lead to difficulty in getting
access to capital to acquire technologies that may
improve performance. Poor books and records yield a
questionable roadmap as to the “as-is” condition and the
improvements you have made. Sales and Marketing can be
directly correlated to quality (product) and price
(influenced by many factors). Staffing problems can
be equated to poor project scope and mis-communicated
expectations. While there are many root cause catalysts
regarding union problems, often times experience has
shown us that individuals who are happy in their
profession are more focused. The implementation of
advanced technologies can sometimes cure these ills.
Failure to use external advice is intuitive. Rely on
your vendor and networking with people of like minds to
pull you through the rough patches. I believe that by
eliminating several of these issues we can reverse a
trend that is plaguing corporations and keeping
consultants very, very busy.
32.1% Poor management
of financial activities
14.6% Lack of management competence or experience
12.4% Inflation and economic conditions
12.3% Poor books and records
10.7% Sales & marketing problems
9.0% Staffing problems
6.2% Union problems
2.7% Failure to use external advice
Courtesy of Coleman
Management Services Inc
The Pain: What
is the Problem?
Historical data reviewed from surveys of PDM programs
yielded the following results. The industry “secret”
that no one bothers to tell you is that
approximately one-half of all programs fail within the
first year. There are a number of factors that drive
this. The chart below shows the data from the survey.
A questionnaire of 25 questions was sent to over 500, of
which 165 responded. Those surveyed were asked to self
rank themselves (it is agreed to be a little subjective)
on a scale of 1-10. We considered an average of less
than 5 over 25 questions as program failure. Our
philosophy was that anything operating at 50% or worse
may not be worth doing at all. I know there will be a
number of purists that argue about subjectivity and a
multitude of factors that could skew these results. I
know. However, I believe that if we ask enough people
the same questions, you will get fairly uniform
resultant responses with a few errant data points. A
modified Delphi approach, if you will.
In the data received, I rounded the number of years the
program was in existence to the nearest year.
The data is based on a
survey performed in 1999.
In year one, of the 165 respondents, 81 individuals
submitted data and ranked themselves low enough to
suggest that their program had failed or was failing.
This trend continued so that by year 5 (and older), only
37 respondents ranked the success of their programs as
successful. That’s a shocking 37 of 165. That is
22.4%. Note: By year five,
only 46 respondents said there program had truly gone
away. By “gone away”, I mean the equipment is sitting
on the shelf collecting dust or being used as a door
Okay, So What?
Well, the failed implementation of any advanced
maintenance strategy or program at your facility costs
money. In the case of a failed PDM program we can
conservatively estimate that $100k was spent between
equipment, people, training, etc. That equates to some
Dodge Ram Pickup 2500
4DR Quad Cab Laramie RWD SB (5.7L 8cyl
5M) 4 of
Shares of Berkshire Hathaway stock
Outback Special Steak Dinners
course this does not include the intangible losses of
credibility and trust. In short, if you are at the helm
of a failed PDM program then you have probably committed
a career limiting offense. It is unlikely that you will
be given a second chance and it is likely that you may
either loose your position or job.
question then becomes, “Why would I be adventurous
enough to undertake an activity that has historically
shown a propensity for failure?” You might say to
yourself, “I am only a few years away from retirement;
I’ll just ride out my time here in my current
position”. If you’re that individual, then you are
right, DO NOT undertake an effort of this magnitude. I
believe that the Risk-Reward ratio is actually quite low
when the proper people, processes, technology and
reporting are put in place.
it for me? I mean you.
must ask this, right? I mean anything worth doing is
worth doing well. If you’re like me, then you’re
probably saying if I am going to do it, I need to know
how I am going to win. Well, here is how. Predictive
Maintenance as an advanced maintenance strategy has
saved companies millions of dollars. I have seen the
results first hand. I have seen the leaders of
successful programs glorified by fortune 500 CEO’s at
company meetings and PDM program managers mentioned in
company 10-k reports. I even knew a guy who got a
parking spot right next to the plant managers (I am not
kidding). For those who know what they are doing, the
accolades are numerous. Fame, glory and in some cases,
big bucks were inevitable. Ask me off-line, and I’ll
tell you some guys I know/knew who did it well. What
else? Well, the technologies (for the most part) and the
trends they develop are mathematically sound, repeatable
and have been around for a lot of years. The trends do
not lie. I consider myself a youngster, but even my
first experience goes back almost 25 years. I witnessed
the application of Vibration Analysis for the first time
while serving as a Machinists Mate aboard a United
States Naval Vessel, (USS Mt. Whitney) in 1981. The
technology, housed in large file cabinet size boxes was
lowered into the ships engine room (11 decks down) and
hooked up to the power plants main engines, auxiliary
turbines, generators, feed pumps, and condensate pumps,
etc. Data was collected by the team and helped the
engineering team to focus resources on critical
equipment during the up-coming outage and yard work. I
asked a lot of questions during that cruise about the
technology and knew then that someday I would learn more
what about industry trends? I knew you would ask. A
few years ago I read a report by one of the large
consulting firms. It has been around the block so you
may have already seen it. The report published in 1998
by Deloitte and Touche asked plant managers to assess
the number of work-orders generated at the facility and
to categorize them into the following bins:
let me explain the bins. The bin of “Reactive”
(sometimes referred to as corrective) represents a
work-order that was generated independent of the
preventive maintenance schedule. It typically is
reactive in nature since it was not expected or planned
for. The “Preventive” bin represents the
preventive maintenance or time based scheduled tasks.
These tasks are typically derived from the OEM’s
recommendation and are based on some schedule of time,
e.g.; replace the bearing every six months. The “Predictive”
maintenance bin is some task that is “predictive” in
nature. In other words, some technology (non-intrusive)
has been utilized to assess when maintenance should be
done. An example is fitting: I replace the oil in my
car every 3,000 miles since the manufacturer suggests it
(Preventive). I replace the oil in my car
when the temperature gauge is pegging and smoke is
bleeding from the valve covers (Reactive).
I install a reliable on-line oil sensor in the car’s oil
sump and it tells me (based on some magical correlation)
that the viscosity, contamination and wear particle
counts suggests, that based on my driving habits over
the past several months, that I should change the oil
next week (Predictive). The Pro-active
bin represents root cause failure analysis. Determining
the root cause of why things are failing and eliminating
that condition. I will not say anymore about Pro-active
maintenance (topic for another paper).
plant managers were asked to assess on a percentage
basis the number of work orders their site generates
today (solid column with vertical stripes). They were
then asked to assess where they would need to be 5 years
from that point to stay competitive (solid column with
horizontal stripes). Obviously, this is a subjective
guess on their part; I am assuming that guess is based
in part on experience, knowing coming trends and
networking/benchmarking with their peers as to where
their facilities needed to be. Finally, the surveyors
took the surveyed companies and based on operations and
maintenance costs, plant/equipment availability, MTBF
(mean time between failures) and unplanned downtime,
ranked them. They extracted the top 10% and called them
“benchmark” facilities. These top 10% of facilities and
their breakdown of work order generation is represented
by the solid bar. The results are somewhat
predictable. Do less reactive maintenance (do not want
to wait until smoke is coming from underneath the
hood). Do more preventive maintenance (I believe this
comes from years of pounding by maintenance
consultants). Preventive maintenance has its roots in
Airline and military industry. Breakdown is typically
not an option here. Also, for the average facility,
preventive maintenance costs a lot of money and is
sometimes over done. Finally, the survey reveals that
these plant managers agree that they should be doing a
lot more predictive maintenance. The good news for us
is that we probably do not need to do much evangelizing
to senior management. The word is certainly out.
Predictive maintenance works. There are a number of
other reasons plant managers would like to move towards
non-intrusive ways of monitoring their equipment to
ensure better reliability. They know that their senior
level executives are not going to allow them to invest
capital dollars into installing redundant systems (would
not want to do this anyway). They know they need to
find non-intrusive ways to monitor the equipment. They
have limited talented craft resources that are over
extended as it is. Their maintenance staff some times
introduces more errors by open-inspecting and fixing
things (did I say that out loud). They are pressured to
reduce outage lengths and therefore must stop the
blanket approach of performing preventive maintenance on
all equipment during an outage. They must find ways to
perform maintenance on a “condition” basis. Predictive
Maintenance offers that solution. Therefore, it’s not a
surprise that these plant mangers responded this way.
about another chart? We have some other data based on
feedback from customers regarding their return on
investment in predictive technologies. Again, it’s no
surprise that the heavy industries with extremely
critical equipment operating 24-7 showed the best
Data collected from PDM
one of my favorites. Imagine, if you will, a
maintenance strategy that returns eight to twelve
dollars for every dollar invested. There are few if any
other advanced maintenance tools that provide a similar
return on investment. These savings are realized in a
number of different areas of the facility. They are:
improved reliability, availability and capacity that
leads to greater revenues. The “predictability”
afforded the engineering, operations and maintenance
teams allows for better planning of outages. Outage
planning is enhanced by the simple fact that you can
better predict MTBF (mean time between failure) rates.
Replacement of Time based schedule preventive
maintenance tasks with non-intrusive predictive
maintenance tasks allows for a reduction in the PM
program without risk of higher failure rates.
Additionally, with so few qualified journeymen rising up
through the ranks, the qualified maintenance resources
can now be focused on those critical tasks that must be
done right. The less harried environment leads to less
rework issues, a reduction in the back log and a much
better environment to work in. Other benefits are
captured in the chart below.
many advanced maintenance strategies to choose from the
question becomes, “Are there easier battles to
conquer?” Perhaps so. Predictive Maintenance in and of
itself is not an easy implementation. The barriers to
success are numerous, as we will discuss later. The
fact remains, however, that the biggest bang for the
buck amongst potential maintenance strategies comes from
predictive maintenance implementations. So, with the
type of returns and improvements documented above, it is
no wonder that for those who have been successful the
paybacks have been nothing short of remarkable. A
survey performed in the 2002 timeframe queried managers
who had managed the implementation of multiple advanced
maintenance strategies over the past ten year period.
They were asked to objectively rank the impact of
implementation. The results (See below) show that a
properly implemented PDM program achieved better results
then such heavy weights as RCM (Reliability Centered
Maintenance), and CMMS (Computerized Maintenance
Management Systems). Interestingly, craft training
(good old wrench turning skills) ranked nearly as high
as predictive maintenance. The chart below captures
the results of a survey of users during the 2002-2003
so there exists a pretty good story why this strategy is
worth pursuing for personal and company reasons. There
is a multitude of objective data collected by third
parties that validates it. We know it is worth doing
and that not many people do it well. The remainder of
this paper addresses those issues of why people do NOT
do it well and offers recommendations for improvement.
This will not be highly subjective content. I promise
Assessing the Problem (As-Is Condition)
the nineties, I had many opportunities, while working
for a major vendor, to tour facilities around the world
and assess their predictive maintenance programs. The
assessments were part of a process whereby the
individual facilities we were assessing were competing
for a coveted Predictive Maintenance Program of the Year
Award this vendor awarded. To be honest, it was
strategic marketing at its best. During the years I
was involved with all of these assessments, and visited
and reviewed several hundred predictive maintenance
programs around the world. The majority of these
companies were considered benchmark in their respective
vertical markets. In all cases they had shown a unique
capability to implement successful predictive
maintenance programs. Some of the programs had been in
existence for 16+ years. While evaluating them, the
participants shared their failures, and the pitfalls and
issues they had faced over the years in excruciating
detail. I began to see common threads that ran through
all of these successful programs. I saw the same
challenges conveyed to me time and time again. I
realized that their success was not based on luck or
some magical formula but sound successful business
practices that on the surface seemed simple and
obvious. Remarkably, the problems fell into four
predictable categories: People, Processes,
Technology/Vendor and Reporting. I decided to perform
the standard fishbone analysis on the problem,
remembering the issues, challenges and pitfalls my
interviewees had suffered through. The following
diagram states the most commonly heard problem in a
succinct one liner. I will elaborate further on some of
the key issues and discuss potential solutions.
outright solutions to these issues is why benchmarking
with successful peers is so important. These items in
and of themselves may not torpedo a predictive
maintenance program. But several combined together, in
the aggregate are a recipe for failure. The chart
reflects years of program manager experience. You can
perform a very simple “as-is” exercise by ranking
yourself (objectively) in each of these categories on
scale of 1-10 (ten being best). Lay the results out in
a Spider Chart (see below) That 360° look should lay the
groundwork to begin stream-lining your program.
write a book on these roadblocks and perhaps one day, I
will. For now, I will hold my comments to several
points per category. Let’s begin with the all important
I will catch some heat for this, but here goes any way.
If you are not a “Type A” personality then it is
unlikely you will be successful in driving a predictive
maintenance program. The program needs an extroverted
sales type person who is continually selling the
benefits of predictive maintenance. Some will say,
yes, but shouldn’t the results sell themselves? Well,
no. I cannot think of any process (inevitable aging and
death) or program (taxes-they will find you), etc., that
is assured. Power (management) likes dealing with power
(credible strong extroverted people). Show me a program
manager that has the capability to present to tough
crowds, has unique visionary capability and the ability
to instill passion and I’ll show you a solid program.
You must be steadfast to the programs success, and
people in the facility must realize your commitment to
its success. The self fulfilling prophecy of any
successful program is to reduce the very problems for
which it was initiated. It is sort of like working your
way out of a job. Over time you will have eliminated
the low hanging fruit. You will see improvement in
availability, reduction in unplanned failures and a
multitude of other benefits. Three to five years into
your successful program, new people/management (probably
MBA’s-I can say this cause I is one) will begin to ask,
“What is it that the PDM group does?” Your knowledge
and historical perspective of the successful past and
persistent tree stump sermons of your successes will
ensure that people are aware of who deserves some of the
credit for the current status in the facility. Your
persistent selling, to management, of the future work
you plan to do will solidify in their minds the programs
value. They must understand that doing anything short
of increasing the funding for the coming year, could
easily plunge them into a downward spiral on key
indicators of plant, system and equipment reliability.
It takes a strong person to defend these positions when
things are even keel. It takes an even stronger person
to convince an operations manager to shut a plant down.
People forget. You do not get many chances to screw
up. Therefore my recommendation, made strenuously, is
that those who will lead the Predictive Maintenance
Effort should also be capable of running a political
campaign. Some recommendations that will ensure you
have what it takes:
the book “How to Win Friends and Influence People” by
the plant manager to lunch monthly.
up on your financial acumen (know how to get access to
capital-understand hurdle rates; cost of capital,
expense budgets etc.). Take your chief financial
officer/accountant to lunch monthly.
not over state your successes (maintain your
a course on “how to present”. (video-tape yourself).
a successful coach. Someone that has the tee
shirt.(or stayed at a Holiday Inn Express last night!)
Benchmark with other facilities
Ensure your manager is in the field frequently
observing your contribution.
Develop a personal education plan for yourself and
go on for several more pages, but that should get you
your sales guy if the vendor has knowledgeable analysts
on their staffs that reside in the country. If the
answer to that question is NO, then run-away very fast.
If this is the case, you will not be able to get
detailed help and you will always be relying on your
sales guys ability to link up with the foreign vendor in
another country six to twelve hours apart. They must
have a local presence, independent of their sales
channel. Does their software have true network
capability? How about open in architecture? NO? You’re
locked for life then. Get used to buying only from
them. By the way, a great test for the sales guy, to
ensure open architecture (future ability for you to
share information with other systems), is to ask them
to demonstrate opening up two different unique databases
and drop and drag one data sub-set (system, plant, and
equipment) to the other database. If the system is
truly open the data will copy straight over. If not, do
yourself a favor and tell them to have a nice day.
Please, please, make sure you ask what they charge for
upgrades (software and hardware) over the next 8 years.
Ensure you understand the overall cost of ownership.
There are quite a few vendors that make a good living
(residuals) on this. A rule of thumb is 10% (of
hardware and software) per year after three years. You
get 3 years and 36k miles from your auto manufacturer,
so why not from your software vendor? What about
reporting? You should be able to easily copy and paste
reports from their software into Word documents,
PowerPoint, Excel, etc. Do not be fooled with all the
bells and whistles. Simplicity is important to your
long-term success. Look for the basics. For example,
if you’re investing in a vibration analysis solution
then ask yourself the following: Does the box (data
collector) collect data and easily transmit that data to
the software? Is the software user friendly and does it
provide a flexible reporting interface (get data easily
in and out)? Does it weigh something less than a small
car, and can you work it with one hand?
recommendations to ensure that you get the best return
for your invested dollar:
not over Pay. These systems are not that expensive to
manufacture. You can use the money you save to get
yourself more training.
Require that all vendors cost justify their systems to
you. They should provide you with a spreadsheet that
shows ROI and Cost Benefit.
Ensure that you spend time getting to know your
vendor. Participate, where possible in providing them
feedback in product development.
Network with your peers; attend tradeshows, e.g.;
Vibration Institute meetings.
Split up your purchases between vendors. Do not rely
on one vendor to provide your every need. They will
become arrogant and un-responsive. Where are you
going to go, when you’ve bought everything from them?
your vendor to network you with other folks in similar
industries with similar equipment.
Insist that some degree of high level local support is
available from your vendor (especially in the
beginning) to ensure that you are making the right
calls, even if you must pay a fee for that support.
single most important task that one can perform in
ensuring the successful longevity of a predictive
maintenance program is proper reporting. This would
include detailed reporting of the “as-is” condition.
This will remind people where we came from. Metrics
should be chosen and “reach” goals set. You will have
to identify performance metrics that need to be
recorded/measured so that you can begin to track your
improved results. Several key items that your reporting
process should include are:
Plant & System Availability (3 year trend).
Total Preventive and Corrective Maintenance dollars
spent by system.
5 equipment problems (Based on unplanned failures).
Maintain a list and track costs associated with this
equipment by Safety, Single Point Failure (SPF)
capability, known bad actors.
Saves. The preempting of a sure functional failure
(assuming nothing was done) by some predictive
technique that prompted corrective action based on
Quantify the savings (be conservative). Know the
costs of downtime, estimate labor overtime, spare
Ensure lessons learned are incorporated back into
processes and procedures (learn from your mistakes).
Understand how to show Return on
Investment and Modified Internal Rate of Return.
(See me for a personal demonstration—I will send you
my own worksheet if you give me your card).
you get the equipment problems under control, go to
the warehouse and see if you can reduce some of the
spare parts inventory. You’ll be surprised how people
bulk up for those incipient failures. Report your
Review the Preventive Maintenance program (target OEM
assigned tasks). Look for tasks that are time based
and intrusive in nature. Try to define a predictive
maintenance task (non-intrusive) that could replace
those tasks. Does a technical justification even
exist for the PM task ?
Track rework (PM or corrective action tasks that
recur, may be due to poor workmanship, wrong parts,
complex maintenance, procedures, etc.).
correlate your predictive findings to Process
information. Review operations logs for Pressure,
Flow and Temperature anomalies that match your data.
your data in conspicuous places (shop entrance, mess
areas etc.). Pictures, graphics and trend charts work
best. Avoid lengthy, verbose chapters regarding your
findings (no one will read them). Red lights, yellow
lights, and green lights get people’s attention.
two brown bag lunches annually to discuss your
results. Invite Everyone.
need to attend plant management meetings. Get on
their agenda to discuss results (be conservative).
Remember Power likes power.
Write an article for the company newsletter (as many
times as you can stomach).
extremely important piece comes together when we
streamline the business through the implementation of
repeatable processes or procedures. Ensuring we are
collecting the right information, at the right frequency
and in the same locations is extremely important to the
long term success of any program. Inevitably, you would
like to get the program to the point, where fresh faces
could come in, take over the program with very little
questions asked and be as successful as you are. Your
success will hinge on your ability to integrate the
predictive maintenance efforts into the day to day
business of plant maintenance. PDM should be discussed
at the morning meetings, be included as tasks in the
Computerized Maintenance Management Systems, and be
understood or proceduralized in the department manuals,
etc. Do not forget to get operations on your side.
They can be your eyes and ears while you’re away from
the plant. Successful PDM programs have outfitted their
operations folks with low cost tools that can identify a
problem. Their procedures then instruct them to notify
the predictive maintenance group who can then perform a
more detailed analysis. A successful example might be
outfitting operations with vibration pens (providing an
overall value) and then notifying the predictive group
if they see anything out of the ordinary. One facility
I worked with gave their operations personnel portable
non-contact thermometers. Understand the bottlenecks
in the plant. The problems may not be poorly designed
equipment. Some thoughts to consider:
Ensure predictive maintenance procedures are
implemented into plant processes (An example might be,
that maintenance knows (and CMMS maintenance
requirement card says it) to call the PDM group after
replacing bearings on a piece of equipment so that
they can come and grab a baseline vibration spectrum
Review operations procedures. Ensure operators are
running the equipment within recommended tolerances.
Study the quality charts. What equipment is producing
the largest amount of scrap or waste?
bottlenecks in the facility will usually be identified
by midstream inventory bulk. Focus on eliminating
bottlenecks by improving system and equipment
reliability. There is quite a bit of capital tied up
in that inventory.
- As a
rule of thumb, do not invest in add-on technology
until the systems you already own has paid for itself
(3X-5X). Ask your vendor to help you prove that.
Recruit a few potential replacements (succession
Obtain more training.
Ensure a feedback mechanism exists during the
maintenance process to allow personnel to offer
suggestions for improvement.
Ensure that maintenance/operations (all personnel for
that matter) have been well versed in how to enter
data regarding a work order into the CMMS system. The
symptoms they experienced may be different by the time
you are notified. You’ll need the complete story. In
a perfect world they will call you as they are
generating the work order to come have a look. There
is nothing more frustrating than reading poorly filled
out maintenance work orders.
Cross train some operations folks on what you do.
Hang out in their world for a few days each quarter
purpose of this paper was to identify and bring to light
a pain that has personally bothered me for the last
decade. I also wanted to motivate the reader that he or
she is capable of implementing an advanced maintenance
strategy (more specifically predictive maintenance) if
they are willing to do the necessary work. Common
pitfalls have been identified in four key areas, namely;
People, Processes, Reporting and Technology.
Recommendations for improvement have been suggested. My
suggestion is that you take the grocery store approach.
Take that stuff off the shelf that fits and leave
everything else alone. I do not pretend to know more
about your facility, or style. If what you’re doing
works and doesn’t align with what I’ve said then do not
change a thing. If not then maybe you should take a
document is not a substitute for an advanced maintenance
strategy implementation business plan (I can also get
you one of these if you leave me with a card).
certain there are many different other reasons
programs/implementations fail. So my intent was not to
create an all inclusive document. I do not know that
apologize in advance to anyone I might have offended. I
am sure there is some introverted engineer that mumbles
and looks at his shoes when he talks and was still able
to implement and run a successful predictive maintenance
program (I iz an engineer also
and I also often resemble my own comments).
feel free to call and comment or ask questions at the