In my last posting (Part 397), which was on Throughput Accounting (TA), I made
reference to one of my blog readers asking me a question. This blog reader asked me to write a posting
on the Theory of Constraints, but more specifically on how I might go after
breaking constraints. He had finished
reading my book,

*The Ultimate Improvement Cycle – Maximizing Profits Through the Integration of Lean, Six Sigma and the Theory of Constraints*and because he was new to TOC and TA, he was interested in learning more about what happens if you weren’t able to break the initial constraint. In order to answer this question, I needed to make sure everyone understood at least the basics of Throughput Accounting (TA), hence the reason for my last posting.
For those of you who have not read my book, UIC, it is a book on the
integration of TOC, Lean and Six Sigma and how I recommend implementing it.
The Theory of Constraints is what I refer to as "the missing link" when in comes to improvement efforts. And for those of you who are not very familiar with the Theory of
Constraints, let me review Goldratt’s 5 Focusing Steps which is the heart of TOC:

*Identify*the System Constraint or that part of the process that is limiting throughput- Decide
how to
*exploit*the System Constraint or how to make the most out of it - S
*ubordinate*everything to the above decision. That is, the non-constraints should run at the same speed as the constraint to avoid a WIP build-up. - If
necessary,
*elevate*the System Constraint. Elevation of the constraint usually involves spending money to, in effect, break the constraint. *Return*to Step 1, but don’t let inertia create a new System Constraint. The inertia in this case is basically becoming complacent once the initial constraint has been broken.

These five steps are referred to as TOC’s Process of On-Going Improvement
or POOGI.

For those of you who have read my book and might be applying the lessons in it, by now you have probably
completed all of the steps needed to “break the initial constraint” and
hopefully you have succeeded in breaking it. Breaking the
constraint simply means that the process that was initially limiting the amount
of new revenue from entering the company, the one that was constraining throughput,
is no longer doing so. This simply means that your output has matched the demand requirements of the market.

You’ve probably significantly reduced process and product variation and processing and cycle time variability by eliminating or significantly reducing things like unplanned down time, scrap, rework, and differences in work methods between operators. If a supplier was your problems, you've no doubt fixed this as well. You’ve probably eliminated much of the waste in your process by eliminating non-value-added activities through things like constraint changeover and set-up time reductions, limited line balancing activities, cellular manufacturing and you may have even gotten rid of

You’ve probably significantly reduced process and product variation and processing and cycle time variability by eliminating or significantly reducing things like unplanned down time, scrap, rework, and differences in work methods between operators. If a supplier was your problems, you've no doubt fixed this as well. You’ve probably eliminated much of the waste in your process by eliminating non-value-added activities through things like constraint changeover and set-up time reductions, limited line balancing activities, cellular manufacturing and you may have even gotten rid of

*efficiency*as your primary performance metric. At least I hope you have. You may have also improved the flow of products through your process through the use of things like Drum-Buffer-Rope scheduling, establishing a WIP cap and implementing protective buffers, pull systems and establishing one piece flow or optimizing the batch size.
But what if, after all of these improvements, like the reader who
asked me his question, you have not been able to improve the throughput of your
process enough to satisfy the demands of the marketplace? In other words, you still don’t have the
needed capacity in your constraint operation to fill the order requirements of
your customers. What do you do now? The fact is you have a decision to make that
calls for a management action. If after
all you have done to this point, through exploitation of the constraint, has not
resulted in the constraint no longer limiting the system’s performance, then
you must move to Step 4 and

*elevate the constraint.*Quite simply elevating the constraint means you must take a different kind of action to improve the capacity and throughput of the constraint if you are to satisfy market demand. You’re most likely going to have to spend money to make more money. You’ll have alternatives to accomplish this, but how do you know which one is the most attractive?
Let me take a step backwards for a moment and introduce those of you
who haven’t read my book to the graphical representation of my UIC. The following figure is how I presented this
integration in my book.

What you see here are three concentric circles with each on
representing the TLS components. I
purposely placed TOC’s circle in the middle because for me it is the nucleus of
TLS. I say nucleus because it is the
missing link in many improvement initiatives.
TOC provides the focusing mechanism for using Lean and Six Sigma. In the next graphic, I have listed the
actions and tools that I recommend for each step in the TLS methodology.

Now, let’s get back to answering the reader’s question.

Although there are several important things you must consider, one is
more important than all of the rest. In Step 1 of the UIC I tell you to
identify both the current and next constraint.
Before you make any decision that involves spending money like adding
additional resources, (i.e. people, equipment, etc.), we must always consider the impact of the next constraint
(or other potential constraints) on your decision. For example, suppose that in order to break
the current constraint, you decide to purchase a very expensive piece of
equipment. You justify the expenditure
on the basis of future throughput through the current constraint and you
determine that the ROI is acceptable.
You then purchase and install the equipment and in so doing you were
successful in increasing throughput at the current constraint. When this
happens, predictably a new constraint emerges.
But suppose that this new constraint limits throughput to the extent
that your overall capacity is still less than market demand. This typically means that the ROI assumptions
you made to justify your purchase probably weren’t correct.

The point is that unless you consider the throughput of the next
constraint as you are justifying the ROI for this expenditure, it is quite
possible that this new constraint will limit your predicted throughput and
negatively impact the return on investment used to justify it. The return on investment you assumed with
this purchase is limited by this new constraint, therefore making this
investment would be an example of wasteful spending. Once again, before you consider how you’re
going to break the current constraint, evaluate the limits imposed by other
potential constraints. I’ll present a
couple of examples later on in this posting.

Increasing the capacity of the constraining operation can be
accomplished in several different ways or a combination of each. Capacity is defined as “an upper limit on the
throughput of a production process

^{1}.” So how could we increase the capacity of the constraint operation without purchasing additional capacity in the form of more equipment? One way is to increase the length of time that the constraint operation runs compared to the non-constraints. For example, if we are currently not utilizing the constraint process during lunches, breaks, meetings, training etc., we could provide operator relief during these periods of time, thus never permitting the constraint operation to sit idle. This would require cross-training to assure other operators are qualified to run the work station, but it would be a lower cost alternative.
Another solution might be to temporarily use overtime in the constraint
operation to raise its effective capacity.
While other stations continue to operate in the normal 8-hour shift, we
might consider running the constraint operation longer. If we do consider this alternative, then we
must balance this cost against the cost of new equipment so that we make the
right decision. Another alternative
might be to hire additional operators and add an additional shift in the
constraint operation, but again, we must balance this against the cost of
purchasing additional equipment.

If using limited overtime or adding additional resources (i.e. adding
an additional shift) is too costly, then one other alternative might be
outsourcing part of the constraint’s work load.
That is, instead of purchasing additional equipment, contracting with an
organization outside the company might be a viable alternative. In general, however, I am not always a fan of
outsourcing simply because it is easy to lose control of your products. In fact it is possible that the product being
produced outside the company could have worse (or better) quality than that
being produced in your own process. What
customers are looking for is consistency, so any action that results in
inconsistency should be avoided. In
reality, however, sometimes it is the best, least cost option. The lesson here is that as long as the next constraint’s throughput is
substantially higher than the current constraint’s throughput, then it’s
probably a good idea to spend the money.
Let’s look at a few examples.

Suppose after all of the improvements you’ve made to your process, your
constraint output was 10 parts per hour.
You are considering the purchase of a very expensive piece of equipment
that should double the throughput to 20 parts per hour. You evaluate the anticipated next constraint
and find that its throughput rate is 12 parts per hour. Even though the new machine would double your
throughput rate to 20 parts per hour, you are limited by the rate of the new
constraint operation of 12 parts per hour, unless you can easily increase the
rate of the next constraint. Your
challenge then becomes one of finding a less expensive alternative that raises
your throughput to some level above 12 parts per hour for this potential
constraint. Otherwise you have spent
money when you shouldn’t have. On the
other hand, if the anticipated new constraint’s throughput is 15 or 16 parts
per hour, then it’s probably a good decision to purchase the new machine, but
you still have to consider this rate in your ROI calculation. The point is, before making a purchase to
break the current constraint, consider its “true” return on investment.

To summarize, breaking the current constraint is really the same as
increasing the capacity of the operation that is limiting the throughput of a
given process. If the constraint is an
internal resource, this simply means developing ways to squeeze more time out
of that resource for productive work so that we can meet the demands of the
market. This might be any or a
combination of the following alternatives:

- Purchase additional equipment
- Hire additional resources
- Utilize overtime
- Add additional shifts
- Outsource or contract with an outside operation to produce product
- Relieve for breaks, lunches, meetings, etc.
so as to never let the constraint sit idle.

If the constraint is external to the operation, then your options are
obviously different. For example,
suppose the constraint is market demand or a lack of sales. Clearly this means that you have excess
capacity, so in this case activities will always be geared toward increasing
the competitive edge factors that result in increased sales. These activities include improvements in
cost, lead times, quality, customer service and flexibility.

No matter whether the constraint is internal or external to the organization,
elevating the constraint usually always means spending more money. It also means that you have done everything
possible to squeeze the most out of the constraint before you spend the money. Again, it’s important to remember that
different alternatives cost more or less than others so whichever one you
choose must be done so after careful deliberation and analysis of the entire
system. Sometimes alternatives are more
attractive than others in ways that can’t be measured financially. For example, if breaking a constraint results
in a process that is easier to manage, then this might be the best alternative
as long as throughput increases. Let’s
look at a couple examples.

Consider the simple five-operation process in the figure below. If we consider only the current capacity of
each process step, then it is clear that Operation B is the capacity constraint
because its capacity is less than all of the other steps. Assuming we have completed our exploitation
step, it should be clear that before any additional improvements in throughput
can occur, we must

*elevate*Operation B.
Because the market demand for this product is 20 parts per hour the
improvement team has come up with a plan to increase the capacity of Operation
B from its current 10 parts per hour to 20 parts per hour. The plan involves the purchase of an
additional machine for Operation B. The
team reasons that if the profit on each part, assuming no scraps, is $10, then
the additional profit per hour is 10 parts per hour x $10 per part or $100 per
hour of additional profit. Since the
cost of the new machine is $200,000, the team concludes that the ROI is approximately
250 work days, or approximately 50 weeks.
The team presents their recommendation to the Steering Committee and it
is rejected. Why did it get rejected?

Although with the purchase of this machine the capacity of the current
constraint would be increased from 10 parts per hour to 20 parts per hour, the
capacity of the line is now constrained by Operation C at 15 parts per hour
which is five below the 20 parts per hour of market demand. This implies that the ROI of the equipment is
not the 50 weeks calculated by the team, but rather the time is doubled to
approximately 100 weeks because of Step C’s limitation. Also, the team did not consider the increase
in operating expense associated with hiring additional resources to run this
new machine. Clearly if the decision had
been made to purchase the new machine, it would not have been in the best
interests of the company.

The team then decided to look at the expense of working overtime on
Operation B. The cost per hour for
overtime on Operation B is 1.5 times the hourly rate of the operator which in
this case was $15.00/hour. The team
reasoned that in order to break this constraint (i.e. reach current demand) the
process has to produce an additional 80 parts per day on Step B and 40 parts
per day on Step C. The process operates
for 8 hours per day so the total number of parts that must be produced on
overtime would be a total of 80 parts.
Since the current capacity of the constraint (Step B) is 10 parts per
hour, then an additional eight hours of overtime per day would be required to
effectively increase the capacity of Operation B to the demand. But because Step C is 5 parts per day below
the current demand, then it too would have to expend 4 hours per day of OT to
meet demand. The cost of this overtime
would be 12 hours/day x $22.50 per hour = $270 per day. But what happens to profit?

We already stated that profit is $10 per part so the profit they are
currently achieving is 80 parts x $10/part or $800 per day. An additional 80 parts per day would result
in an additional $800 in new revenue (i.e. $10/part x 80 parts). So if we subtract the OT cost from this
number we see that the profit on these 80 parts is $800 - $270 per day extra
labor charge = $530 per day. So the new daily profit would be $800 + $530 or $1,330/day.
If we assume that there are 240 working days per year, then the total profit
for the year would be 240 days x $1,330/day = $310,200. abor cost per year would be 240 days per year
x $270/day or a loss of $64,800 per year.
Realistically, it would make much more sense to simply add 1 full shift
for Operation B and 4 hours of OT for Operation C. It would be very difficult to sustain 8 hours
per day of OT on Operation B. 4 hours
per day would be difficult as well, but this could be done for some period of
time. The question that must be answered
here is this surge in demand something that will continue?

What other alternatives might we look at? One alternative is to outsource or
subcontract the work to an outside company.
Suppose that the subcontractor charges $5 per part produced including
transportation costs and they are able to produce the additional 80 parts
required to reduce the deficit to zero parts.
Again, assume that we use the subcontractor 240 days per year. The total additional cost to the company would be $5 per
part x 80 parts per day x 240 days per year or $96,000 per year. Since we are making $10 per part in profit,
then our profit on these 40 additional parts per day would be ($10 per
part)-($5 per part) = $5 per part. This
translates into an additional annual profit of $5 per part x 80 parts per day x
240 days per year or $9,600 per year.

Let’s look at our process again.
In the figure above I have added an additional bit of information that
might help us with our decision. Notice
that in our constraint operation we are currently losing 2 parts per hour due
to scrap. The team has come up with a way of reducing this scrap level to zero,
but it requires that the process be slowed down to an effective average
capacity of 9.5 parts per hour. Is this
a good decision? Obviously if we gain 2
parts per hour in scrap avoidance by reducing the effective capacity by 0.5
parts per hour, then there is a net gain of 1.5 parts per hour, thus making it
a good decision. It doesn’t get us to
the needed 20 parts per hour to satisfy market demand, but it gets us
closer. There are various scenarios that
could be worked on, but you do have options.

I didn't include the benefits of relieving for breaks and lunches, but doing this would not cost the company any additional money since this would be done by existing employees and would result in an additional 10 parts per day.

The calculations used in the preceding discussions could probably be
questioned or be seen as incomplete, but the point of these exercises is there
are different ways and combinations of ways to break the constraint. The key is to think through and make rational
decisions on the best way to do so.
Remember, when you break the current constraint, a new constraint will
immediately appear so you must be ready when this happens. Always, always consider the entire process,
including the most likely next constraint when making capacity or throughput
improvement decisions.

Bob Sproull