The latest wave
in the chemical industry is to implement Enterprise Resource Planning
(ERP) systems. The goal is worthy having a cohesive information system
to cover all business needs from financial controls and reporting, the
management of our relationships with and sales to our customers, long-range
planning of capacity requirements and short-range production scheduling,
supply chain management, inventory management and cost controls. ERP,
in short, should provide all employees in the corporation with information
to support them in performing their jobs.
Since everyone will be dependent upon the data from this pervasive information
system, everyone is dependent upon everyone else in the organization
to input accurate and timely data. The tool is really everyone's-ERP
serves many 'masters' in an organization. The chemical industry has
followed the historic pattern of other industries in choosing information
systems such as ERP. The initial packages were selected by the finance
department of these organizations. And the financial groups within the
chemical companies have done an excellent job of choosing their tool.
They have typically chosen the best financial tool available today.
But chemical production is a very complicated business. The complications
of balancing preferred sequencing in production against the customer
demands (which never matches that sequence!) against the need to keep
this high-fixed cost facility running to drive unit costs down is just
one example of pressures that will be placed upon an ERP system. Balancing
fluctuating quality parameters and controlling processes with strict
set points, producing end products to specifications using different
technologies, understanding the cost ramifications of new product introductions
or product abandonment, these are just some of the complications of
managing the chemical production site.
Why point out the obvious? After all, you live with and manage these
complications everyday. The whole point here is that your tools for
managing this complicated business-your ERP system-must have enough
robust functions to mirror your information requirements, however complicated
they might be. Without these functions you will have to either change
the way you do business, use another tool to manage that part of the
business, or change your ERP system. While no packaged ERP solution
will meet anyone's requirements perfectly, finding a closest fit to
your requirements means that the ERP tool will leverage everyone's productivity
and improve the company's overall performance.
Looks good to me
Selecting software from the various choices available in the marketplace
is a bewildering experience. So many details, so many presentations,
so much the same. If you have never selected and implemented an ERP
package before, there are some lessons others have learned that you
need to watch out for.
The first step is to determine what you really need in a system in order
to provide the best fitting tool possible. The definition must have
the same scope as your ERP implementation-and everyone who will be using
the ERP tool should be represented in defining the total requirements.
This is a tedious task, but there are some tools which can help you
in this effort. Remember that you need to identify those information
requirements which are unique to your particular business. For example,
you might need to report inventory cycle counts by height of the liquid
in a tank and have the system convert this to the inventory unit of
measure (UOM). This is a relatively unique requirement as for example,
an electronic manufacturer does not need this function. On the other
hand, making sure that the system will balance both the debits and the
credits when posting to the general ledger is a universal requirement.
For the chemical producer, missing either of those functions used as
examples (tank measurements and balanced entries) would mean that the
ERP system was not a proper tool for their needs. The problem, of course,
is to find out if there are any of these holes in the system before
buying. And that's the rub. One can never know everything there is to
know about the complex details of an ERP system before purchasing that
system. So one must know what questions to ask the vendors. Sometimes
we don't ask a question because it seems too detailed. But sometimes
we don't ask a question because it seems too obvious and we end up sadly
surprised when learning to use the ERP tool. For example, every business
in the world requires that a general ledger has an ability to check
for a balance between the debits and the credits before posting the
entry. This is clearly an obvious question which does not need to be
asked. So what are the obvious questions which should be asked?
Far beneath the surface...
You need to ask the questions about parts of the ERP system which are
critical to your success as a chemical producer, but which are not universally
required around the world by every industry. Remember that the majority
of ERP suppliers are selling their systems to many industries-automotive,
aerospace and defense, electronics, banking, insurance, health care
and chemical companies. The general ledger requirements are all the
same, but there are significant differences in other segments of the
ERP offering.
Here is a real conundrum. Most of you have not worked in these industries.
Therefore, it is not possible for you to know the differences in how
they manage their businesses and what unique requirements they would
have from an ERP system. So how do you know which requirements you have
as a chemical producer which are different than the requirements of,
say, a machine tool maker?
The danger is, of course, that in looking at the surface of ERP systems
they all seem to be doing a good job. And sometimes the fatal flaw is
not sitting on the surface of the system waiting to be easily seen.
The flaws are typically down deep in the system, far beneath the surface,
to be found only when you actually go to use the tool.
Fatal flaw # 1: UOM conversions
At the Plant Centric Conference held in April 1998, a representative
of Witco was describing his company's ERP implementation. His stunning
news was that Witco in the US will convert all of its plant operations
to the metric system. The immediate question should be Why such drastic
action?
"Going Metric" has drastic and wide-reaching ramifications.
You see such a seemingly little, bitsy detail as a UOM is all pervasive.
Just as you need units of measure to describe your formulations, to
record your purchases and their receipts, to measure volume of production,
to meter out shipments to your customers, to calculate performance measurements,
the UOM is all pervasive in an ERP system. It is so pervasive that,
like the air, you don't think about it. It is so critical you might
not even think to ask an ERP vendor if they can do UOM conversions and
how-always ask them how-they do it.
So, if your ERP system does not provide you with a sufficiently robust
tool in the area of UOM conversion, and you want to have a universally
consistent information system as Witco does, then you will be forced
into a difficult decision. Witco chose to change the way it does business
in a significant part of the world. And one can only imagine the additional
burden this will place upon the business. They believe that linking
up MES level systems will assist their operators in verifying their
inputs, but what about other areas of the business?
Will they ask their vendors to label their products in metric units?
What about their customers? Will they change pricing to metric units
and ask their customers to work in metric? What about historic data?
Will it all be converted? Even the general ledger-the universal tool
for everyone-could present them with a problem because some accounts
carry not only dollar values but quantities. And those quantities have
an implied UOM, not a recorded UOM. One can only guess that how many
mistakes will be innocently made over the time. UOM is a single field
on any screen in an ERP system, but not having robust enough function
to support UOM and its conversion could be a fatal flaw.
Fatal Flaw # 2: Quality is just for inspectors
While the UOM is pervasive and so common as to be almost taken for granted,
quality is another critical factor for most chemical producers, and
it is completely invisible. What the chemical ERP system must do is
to make the quality parameters visible to those who need to know. Since
the 'need to know' personnel are throughout the whole business, the
ERP system must treat quality an intrinsic part of the whole, not just
an 'add on' module used only by laboratory personnel or quality inspectors.
There are many places where quality is critical to the management of
the chemical operations. If mistakes are made in these areas, the results
could range from a small cost problem to a lost customer to real also
safety hazards. Quality information is needed by the scheduling personnel
who may be able to blend off some slightly off-spec materials in the
next production run of that product. And quality information is required
by our customers. They need to know the quality parameters of what they
are buying. Indeed, the chemical producer may be offering a 'sell-to-specification'
service, wherein the producer is supplying a product which meets his
customer's specifications for the customer's own end use needs. In this
case, if the quality information is not made available to the customer
service representative taking the customer order, how will they be able
to provide the customer with assurances about delivery?
Fatal Flaw # 3: Lot number=serial number
Tracking various quantities of products with different qualities requires
some reference number in the ERP system. For some suppliers, the need
for 'lot control' is translated by them to mean the same as serial number
control for their machine tool manufacturers. This is another example
of a tool which is not up to the job.
There are two problems with this 'serial number' approach. One is the
tracking of materials themselves especially if there are fluctuations
in quality during the actual production run. The second has to do with
control over the use and sale of the materials with various quality
parameters.
First of all a serial number assumes that only a quantity of one is
produced. The one item may have been produced alongwith others, but
they are all identical. Same parts making up the machine; same quality.
If the 'serial number approach' is to work for the chemical producer
then they must be operating in a batching mode where the lot number
will represent one quality specification for that batch. There can be
no fluctuations in quality within the batch.
But for other producers who run campaigns, there will be a problem.
During the campaign quality parameters may change slightly. Transition
(or twilight) materials between various grades made during the campaign
need to be identified. The lot number might represent a day's production,
but there are sub-lot requirements to track the varying quality outputs
during the day. Without an ability to work with sub-lots, the continuous
flow chemical producer will be 'playing tricks' on the system to get
it to support their information needs.
The second problem is not the identification of the various lots, but
the need to control the use and sale of the lots. Once the product is
produced we may not know the quality disposition of that lot till some
time. Some physical testing may take several days. Yet the materials
are in inventory, using up valuable storage space while the lab works
through its testing routines. The chemical ERP system should be able
to simultaneously track the materials from an accounting point of view-you
do own the products that you have just made-while also preventing unauthorized
use or sale of the product till disposition has been made. This implies
that there is a layering of information within the chemical ERP system
which separates the financial implications of actions (producing goods,
receiving goods) from the physical realities (goods in storage awaiting
disposition, available in three days; goods in storage approved for
immediate sale etc.). Yet the ERP system must make sure that the two
views-financial and operational-are inextricably tied together under
the covers. This design approach is not very obvious to the person trying
to select the ERP system which will best fit his needs, but it is a
fundamental requirement. After all, the ERP system is for multiple masters-finance
and operations. And the ERP system must be the unifying force which
keeps the financial picture exactly parallel to the physical picture
of the operations.
Fatal Flaw # 4: Only one way to make a product
The complications of the 'physical picture' mirrored in the ERP system
do not stop with how lots are tracked. Another complication comes from
the evolution of the chemical processes themselves. Within the chemical
enterprise, if not within a single production site, there are various
technologies which can be used to produce the same end item. This is
a reality in the chemical world, but is not typically accommodated in
the traditional ERP systems design. For the machine tool maker there
is only one way to assembly his machines.
For the chemical producer, having only one statement of how product
is made is not sufficient. If there are old processing lines and new
processing lines which make the same end item, then both are valid.
There differences are, however, critical and significant. Their cost
structures are different. The yields may be different. The run rates
may be different. The formulation may be different.
If the ERP system is to fit the chemical producer's requirements, there
needs to be an ability to state all methods of production. If processing
times and formulations vary from one process to another, then these
must be clearly delineated, visible and available to planners, schedulers,
operators and cost accountants. The planners and schedulers will need
to know what capacity will be used to produce the required end items.
Without this visibility they will be looking at the so-called standard
capacity as overloaded, and the so-called alternative capacity as idle.
They will also need to know if there are different material inputs required
from one process to the next. As they plan to load production on various
processes the planning results will then drive the procurement of the
right materials based upon the correct formulation for those lines.
The basis for performance measurement will be set as the 'standard'
for the scheduled process. Thus yields and run rates will be properly
measured against the right basis. But these measurements need to be
augmented by the cost analysis of the actual production run.
One of the 'hard-to-see' but essential requirements of the chemical
ERP system is to develop a weighted average cost for a product when
there are multiple valid processes which produce the product. With this
information, the company can measure its performance against the projected
run rates and usage, as well as against financial standards. The system
must be able to calculate the weighted average actual cost (actual cost
of each process during the accounting period, averaged according to
the relative volumes produced from each process). The system must also
be able to calculate a projected weighted average cost. This means that
the calculation will be driven from the stated volume of production,
and also how the planner predicts to load that total demand upon the
various processes available to them.
Fatal Flaw # 5: No fixed cost analysis
The chemical enterprise is one of very high capital costs. The plant
and equipment are a significant part of the cost of any product produced.
The most important measures which top management uses have to do with
these assets: Return on Assets (ROA), Evaluated Value Added (EVA), Utilization
and Yield (UY). Not to put too much of a negative spin on this, but
downward shifts in yield mean that the plant was being utilized but
the output of good (saleable) product was less than projected. A downward
shift in utilization means that the time available to produce product
has been lost forever, and the cost of that time must be borne by all
other products which are produced.
Typical ERP systems provide very meager tools for the chemical cost
analyst. This is because the typical ERP system is a job costing system.
The machine tool maker costs the assembly done on a particu
That configuration of machine tool may not be made again for months.
The cost of that machine tool is primarily materials and labor. Adding
up those costs and collecting them against the job order provides their
management with the performance measurements which they are looking
for.
Not so for the chemical producer. You might have made the same product
many times during the month. Using different processes because of fluctuations
in feedstock quality; availability of one technology over another. If
you are going to get your arms around performance in an operating environment
like this, you need specialized costing tools.
Some ERP vendors will encourage you with their costing power derived
from data in the general ledger. This so-called cost-center costing
works with allocation logic which parallels the movement of materials
from one cost center to the next in the manufacturing process. This
tool will be sufficient for you if you have only one, or a very few
products flowing through these cost centers. If you are introducing
more and more products and their variants into your processing facilities
you will soon outstrip the power of a G/L-based costing system.
The ideal chemical costing system will allow you to understand the impact
of not only volume throughput changes, but the changes in product mix
as well. Some chemical companies have departments which do nothing but
work with the 'what ifs' of changes in volume and mix at specific plants
as well as shifts in production volume and mix from plant to plant within
the enterprise. Given the volumes of production coming through some
of these facilities, small improvements in the unit costs mean big differences
to the bottom line.
Fatal Flaw # 6: Manage materials above all else
The make up of costs, however, is not just dependent upon the absorption
of fixed costs by the volume of production. Each product's cost has
other elements which may be significant to see when analyzing the performance
of various operations. Certainly there are material costs. But the typical
ERP system places too high a priority on understanding, managing and
controlling materials. In fact, they place materials as the top priority
in the system and with good reason. There are thousands of parts in
a machine tool or an automobile. The majority of the money in the business
is tied up in the material costs of the products they are making. Their
primary concern should be the smoothing out of the flow of materials.
For the chemical producer typically the material costs are not the most
significant cost. The cost of capacity is usually the highest cost.
Therefore the prime driver of the business is the optimization of the
utilization of the capacity. The chemical producer's ERP system should
not force them to plan materials prior to planning capacity.
But capacity is not the only significant cost element in chemical production.
Commercial chemical production is often a very high energy consumer.
Whether the energy is electrical, natural gas, fuel oil or steam, whether
the energy is procured from external sources or produced on site, the
cost is significant and sometimes the supply of the energy is a limiting
factor on one's ability to produce. Another limiting factor on the chemical
facility's ability to produce may be the authorized volume of waste
production in any given period of time.
If the ERP tool is to fit a chemical producer's requirements, there
needs to be an ability to track all consumption and production, to plan
for, track and cost all inputs and outputs of the processes which take
place in their operations.
Fatal Flaw # 7: Only one output
A bill of material-based ERP system has one architectural assumption
which can not be changed: the manufacturing process takes many things
and makes only one thing. From this architectural foundation grows the
total system. If your chemical processes are essentially 'assembly'
processes (blending operations, for example), then this model of your
process will suffice. However, if your chemical processes are complex
producing by-products, co-products and recycle streams, a bill of material-based
ERP system will fail you in the end.
Of course, the typical ERP suppliers have tried to adapt their systems
to accommodate the 'multiple outputs' reality. They have fallen back
on the roots of computing-mathematics. What they would have you do is
to enter your output as an input but with a negative quantity required.
And mathematically this will work since a negative times a negative
will result in a positive. In other words, if you negatively add something
to the process of producing something else, then the negative add is
really a produced not a consumed quantity. Confusing? Do you think any
of your operators will have trouble with this concept? But, it must
be admitted that the arithmetic does work.
The problem is that there are other limitations which are-like all the
fatal flaws-beneath the surface. First of all when will this negative/negative
be produced? The bill of material system has only two possible answers.
It will be 'produced' at the same time as all other ingredients are
consumed, or it will be 'produced' at the end of the process. The first
is the logic of the system without the ERP supplier modifying any of
the logic. The second will require the ERP supplier to change the logic
of the programs to understand that a 'negative quantity per' will mean
that the 'lead time' for the negative/negative is really zero (ie equal
to the due date of the produced item). Remember that every time the
ERP vendor adds additional logic to the system to play a trick on the
original logic processing time is used. Eventually, too many tricks
will impact the performance of the overall computer system.
But what if your by-product is really produced after the first processing
stage several hours before the finished goods are produced and several
hours after production starts? There is no answer for this in the typical
ERP system. And if you are doing just in-time scheduling, anticipating
this by-product's availability for other processes (even as a fuel source
for your energy plant) this imprecision in the ERP planning and scheduling
logic is burdensome if not unacceptable.
And what about the cost of by-products and co-products? With the negative/negative
approach you must enter a value for these items into their master records.
This will work if there is a set value (Net Realizable Value) for the
item which will credit the cost of production. Remember, however, that
if the by-product is actually a waste stream which will require additional
expense to clean and dispose of, then the value you enter will be a
negative-yes, another negative.
If the output is not a by-product, but a co-product the cost calculations
from a typical ERP system will not assist you in understanding the relative
cost of the various co-products. What you really need is a system which
will allow you to direct the system to calculate processing costs up
to the point of the co-product's production (it may come out of the
process before the last processing stage) and to apportion costs based
upon some % distribution. Alternatively, you might want to use a volume
or a weight distribution factor. Either way, the deal system will follow
your rules.
Co-products present another problem for the typical ERP system. Since
they are managed by the system as a negatively consumed material, there
is no ability to forecast or master schedule the co-product. Depending
upon your products and processes, this may be a severe limitation of
the bill of material-based ERP tool. And, by the way, you will have
to forget about recycle streams. Bill of material systems will not allow
them.
Little by little, it all adds up
The selection of the right tool is not easy. But knowing what you need
is the first step. These 'fatal-flaws' are just some of the areas where
little things which don't quite match the way you run your chemical
operations can begin to spell big trouble for a company. The power of
an ERP system is the easy access to, and interchange of, information
through all parts of the organization. If a group has to play tricks
on the system to get it to work, if they are making mistakes because
they don't understand how to use the system, or-worst case-they develop
their own 'side systems' which work better and then half heatedly fill
in the blanks for the accountants in the ERP system, your hard work
and investment in your ERP system will be a wasted effort.
It does not have to be like that. Spend a little time studying what
you need and what the ERP vendors have to offer. There is the right
tool out there for you.
SUSAN J CONNOR,
Sr VP (Chemical Industry),
Marcam Solutions Inc.
Courtesy: DSQ Software Ltd.
Fatal Flaws
New Update