Believe it or not, it is often a very good thing to find that you have some bottleneck in your production system. A body of knowledge about how to take best advantage of bottlenecks has evolved into what is now known as the Theory of Constraints (TOC). It is said that Jeff Bezos, who founded Amazon, made the original book on TOC required reading for his managers; and Jeff has done quite well for himself!
Table of Contents:
Expected Results from TOC
The Cost of a Bottleneck
Five Steps to Better Throughput
Step 1 - Identify
Step 2 - Exploit
Step 3 - Subordinate
Step 4 - Elevate
Step 5 - Start Again
The End of the Beginning
The goal of most For Profit businesses is to make money. Not for Profit's have a different goal but the idea is the same. We want to maximise something!
A business makes money in 3 basic ways:
- Increase throughput: Throughput is defined as the cash income after deducting truly variable costs. Profit is not as good a measure as throughput because profit is counted before we have the cash in hand whereas throughput requires us to have the money in our hands and therefore takes into account problems associated with accounts receivable.
- Reducing investment: Investment in TOC terms is defined as cash that is tied up in the business. This includes not only the Capital Investment in the business but also Working Capital like Inventory and Accounts Receivable. Some investments, like Working Capital and Accounts Receivable, can be turned into throughput but we can't count it until it is.
- Reducing operating expenses: TOC considers Operating Expenses to be all expenses of running a business except the Raw Materials that go into production, which are true variable costs. TOC considers most labour to be an Operating Expense rather than a variable cost, which is different to normal treatment. It does this because it would be comparatively rare for someone to be laid off simply because production took a short downturn and so labour costs are not truly variable.
TOC can have a very considerable impact on the first two of these three factors leading to making money and therefore can provide a sustainable increase in cash in the bank.
It achieves this for several reasons:
- Increases the speed to market of innovation which allows your business to beat its competitors.
- Improves on time delivery meaning you collect cash earlier.
- Produces products that are desired by customers in both price and functionality.
- It is worth always remembering when considering how to improve your business that you cannot save your way to improved growth and cashflow. Any improvement in cashflow and profit, and ultimately money in the bank, must come from an increase in production.
But, there are always things that prevent more throughput in your production system. TOC calls these bottlenecks “constraints”.
The function of a “bottle neck” is to constrain the amount of liquid that runs out of a bottle to make it easier for you to control the rate of flow.
Similarly, the throughput of your production line cannot be more than the (usually) single factor that restricts the flow of work through your production system.
A Simple Constraint Example
Let’s look at a simple example of the impact of a constraint or bottleneck.
Consider a pump pumping water along a pipeline into a tank. Part way down the pipeline there is a valve that is presently set to 80% open.
We want to increase the amount of water that goes into the tank.
If we increase the pumping rate, that won't have the effect of improving the amount of water into the tank because we can't squeeze much more water past the valve that is closed to 80%.
If we increase the size of the pipe above or below the valve, we won't increase the amount of water going into the tank by a significant amount because again, the bottleneck is the valve closed to 80%.
Therefore, it is obvious that the constraint/bottleneck in this example is the valve shut to 80%. We can increase throughput through our system by simply opening the valve from 80% to a maximum of 100% and therefore increase the throughput of the system by 20% immediately.
The same example applies to your business. There is some constraint that governs how much throughput that you can have irrespective of how well the other parts of your business are operating. Until you manage this constraint, improving other aspects of your business is futile because they will not improve the overall through put of your business.
A business constraint example:
To illustrate how a constraint operates, let's consider a restaurant.
Any restaurant has several components in its "production line". It starts with a booking system to take reservations then moves through wait staff to seat customers and take orders, onto the kitchen to prepare the food and ultimately the consumption of the food by the restaurant clientele at their table.
We want to increase throughput at our restaurant. The only real way to do this is to improve whatever limiting factor there is. If we over-engineer in the wrong place, it simply leads to more production which is stored as inventory.
Let's consider our restaurant. What is its most likely constraint?
- It is probably not going to be the booking taking system as it will take as many orders as we like. An improvement here will not increase throughput.
- It could be the kitchen if it is under-staffed or poorly configured.
- It is not very likely to be the wait-staff because your own experience probably is that they usually have a lot of idle time on their hands during a serving session. Adding more staff will not increase throughput and profitability. In fact, it will damage it. This is an example of the futility of adding more resources to something that is not the constraint. In essence, we have increased our inventory of wait-staff-time without improving throughput. We will come back to this point later because it is important.
- But it could easily, and most likely, be the number of tables in a restaurant. If you have 25 tables that can seat a maximum of 4 people, then any sitting at your restaurant is limited to a maximum of 100 people. Since the amount of space that you have in the building is probably constrained, so are the number of tables and therefore the number of patrons and therefore the restaurant throughput.
You might be able to improve the throughput, given this constraint, by:
- Having more than one seating in the evening.
- You could open for a breakfast and lunch and even morning and afternoon tea session providing you did not have to increase the labour hours. Given labour is essentially a fixed cost and already paid for by the very act of opening to trade, any extra work for them in the same working hours is pure profit (after removing the true variable cost of the meal Tools).
- If you had additional unused capacity in your kitchen, you are certainly not going to have them preparing food that no-one eats. But you could add a take-away arm to your business to further increase throughput.
In this practical example we have identified that the number of people that we can seat is the most likely constraint and we have also identified several ways of managing that constraint so that we can improve throughput.
You may think that this example doesn't apply to your business because you are not a production system. But in fact, most businesses are a production system of some sort and TOC and the issue of constraints is likely to apply.
Let's look at some further examples:
- An assembly line or a set of repeated operations: This is one of the more common applications of TOC. If you have one machine in a series of machines that has limited capacity, it will be your constraint and no matter how hard the other machines work, they will not lead to any more throughput in the assembly line.
- Innovation and build-to-order businesses: If you are producing an innovative product or you are building something to order, you will invariably find that just one or two of the departments in a chain of tasks to produce the order will be your bottleneck.
- Professionals office: Accountants, lawyers, architects and so on are equally susceptible to a constraint. Although we cannot predict exactly what it might be, very often it is going to be related to the amount of time that their professional staff have available to sell. A lawyer can only sell so many hours in a day and no matter how much work comes in or how much that lawyer has by way of support staff, at the end of the day the throughput of their system will be governed by the amount of time the lawyer has.
- Accommodation: Most accommodation is going to be constrained by the number of rooms that it has. You might think that it is the number of customers, but in busy times when they have more customers than the number of rooms, it very quickly becomes apparent that the number of rooms they have is a constraint. Therefore, they want to have those rooms in service and available as much as possible.
- Project Management: Most projects are a series of steps. Some of those steps are going to be more constrained and more of a bottleneck than others. No matter how well aspects of the project management aspects of the cycle work on either side of the constraint, the project will not proceed faster than the bottleneck can service the project. This can often be a person with specialist and in demand skills. If several projects compete for their time, all of them will be constrained by the same resource.
- Coffee shops: Could be constrained in two or three ways. Firstly, it might be constrained at the cash register. If people see the queue for the cash register out the door, they might go to another coffee shop nearby rather than wait in the queue. The secret here would be to have more cash registers and make them wait once they have ordered and they are more likely to hang around. A second constraint might be how many cups of coffee the coffee machine can make simultaneously. If you only make one cup at a time, you are going to be far more constrained than a coffee machine that can produce four at a time. If your coffee shop is one where people sit to consume their coffee rather than take away, you may be constrained by the number of tables; as in our restaurant example. This is particularly interesting because some coffee shops make a thing of encouraging people to sit at their tables, consume their coffee, use their computer, consult with their smart phone. Very often, there is only one or two people on a multi-person table. Therefore, the coffee shops ability to sell coffee is somewhat limited by the length of time people spend sitting at the table. Therefore, encouraging people to sit at their tables and work, and perhaps consume only one coffee per hour, may be a strategy that has a detrimental impact for coffee shops.
- Retail: There could be a number of different constraints for retail but one very obvious one is the number of people who come through the door. If potential customers are not entering, they have no opportunity to buy no matter how attractive the product on display is. One way to alleviate this constraint is to introduce an online website selling the same product and then you are open to many more customers 24/7/365 days world-wide (if appropriate) coming through your door.
- Real Estate: Quite probably a Real Estate Agency is limited by the number of properties that they list. They can't sell a property unless it is listed so their principal constraint, and most likely their main focus, will be getting more properties to list.
- IT Department: Very often in an IT Department there is one person who seems to have his or her finger tips on every aspect on the way the IT Department operates in your business. While that person is handling one problem, they are not available to handle others, so we have the same problem as with Project Management where a person becomes the constraint.
- Construction: There could be a number of different constraints in the Construction Industry. It might be getting sufficient trades people on time or even available. It could be that the weather has an impact and this would be especially so in countries with a short construction season due to heavy snowfalls in winter or monsoonal rains in summer.
- Hospitals and Surgery: These could vary. With hospitals, it might be the number of beds available to service patients. If we don't have a bed, the patient can't be admitted. For surgery, it might be the availability of operating theatres or alternatively some of the services assisting operating theatres, such as radiology or appropriate nurses.
The above has given several examples. Necessarily, they are general in their description as we do not have intimate knowledge of every one of those industries. However, you do!
Take a few moments now to pause and consider what is the likely constraint in your business. Until you can find it, you are not going to increase the throughput and profitability of your business. You run the risk of running the rest of the business inefficiently, generating surplus inventory and other costs that have the effect of dragging your profitability down.
Earlier we discussed why there is a neck on a bottle (i.e. bottleneck). It exists entirely to control flow of the liquid contents into some other receptacle. If we didn't have that neck to restrict how much liquid can flow through, we have a much more difficult task trying to regulate flow.
We have also learned that there is very little point in trying to improve every step of the flow of activity through your business as a vast majority of them will have little impact on the total throughput. That is governed, almost exclusively, by the ability of your constraint to have work pass through it.
Trying to improve every step of your business is likely to worsen the situation because inventory from the production of those individual stations will build up as it can't get through the constraint fast enough. Optimising every production stage is known as "Local Optima" and it is not a good thing (see Local Optima article).
Throughout Profit Savvy we talk about the 80/20 Rule whereby 80% of your production is undertaken by 20% of your resources (see The Amazing 80/20 Rule article). However, the impact of TOC is rather more like a 90/10 Rule than an 80/20 because the constraint has such a huge impact on the throughput of your business.
Expected Results from TOC
It is very difficult to predict, for any production system, what sort of results you will get from better managing your constraints.
There are a lot of anecdotes about the productivity increases that are gained from TOC if you look around on the web.
One more scientific analysis that has been done indicates some of the opportunities that TOC might offer you to increase your throughput and therefore your profit.
There is a Case Study in the literature for a business named Sanmina-SCI. This is an $11billion electronics components and assembly firm. It released a comprehensive study, with statistical evidence, on the impact of TOC on their business. They conducted an internal evaluation comparing traditional Six-Sigma and Lean Implementations (two other workplace optimising techniques) with and without TOC being used to focus the efforts of the Lean Six-Sigma tools.
The results were quite stunning.
While all 21 plants saw improved results from implementing these work flow improvement techniques, the 6 plants using TOC to focus efforts generated 89% of the total improvement! The other 15 plants together generated only 11% of the improvement.
Clearly, by focusing on the constraints, TOC leveraged the improvement efforts to produce 15 times greater improvement than the other optimising efforts alone.
By way of illustrating the cost a bottleneck can have on your business, let’s consider a production line that costs $2million per month to operate. In TOC language, these are "Operating Costs".
Let us further assume that the bottle neck is operating at a maximum 5 hours per day, 5 days per week and 4 weeks per month, a total of 100 hours per month. The cost of the plant, therefore, is $2million divided by 100 which is $20,000 per hour. Because nothing else can operate faster than the constraint in an assembly line, the entire assembly line, and therefore the entire operating costs, hinges on the productivity of the constraint.
If we could increase the constraint to working 3 shifts, or 15 hours per day, that would improve its productivity to 300 hours per month meaning that the operating costs per hour of the plant reduces from $20,000 per hour to something like $6,000 per hour. The improvement in the throughput of the constraint has led to an improvement in the operating costs per hour of the entire plant.
In contrast, an hour saved anywhere else in the production line has little or no impact on the operating costs per hour.
We are about to demonstrate how you improve the throughput of your constraint and therefore your throughput of your entire business. TOC teaches there are 5 steps to optimising the constraint on your business to maximise throughput.
- Identify: work out where the constrain is
- Exploit: make use of the fact it has been located
- Subordinate: focus attention on improving the constraint’s throughput
- Elevate: make it the subject of major focus on improvement
- Start Again
The first step is to focus on finding where and what the constraint is.
Sometimes that might be obvious but at other times it can be quite subtle.
If you are fortunate to have an assembly line type business, you will often identify the constraint by the amount of work piled up in front of it waiting to be processed.
If you have a business that doesn't readily lend itself to this type of visual analysis you are going to have to think more carefully.
There are, typically, 6 types of constraints:
- Physical constraints or the availability of resources.
- Managerial policies and procedures.
- Market and marketing based constraints.
- Human and equipment capabilities.
- Quality and quantity of inputs.
- Suitability of operating techniques.
Alternatively, we can consider constraints might sometimes be physical and other times managerial.
For businesses that are producing a physical product, it might be comparatively easy to find the bottleneck.
It can sometimes be identified by a pile of uncompleted work in progress stacked up in front of it waiting for its time to go through the constraint stage of the cycle.
Alternatively, it might be identified by the fact that units waiting downstream of the constraint are sitting idle waiting for input. This might also apply to upstream units that have produced sufficient products to keep the constraint operating and therefore waiting for more demand to be placed on them as the constraint draws down on the stock piles.
One might start by looking for the busiest part of the production cycle on the basis that it is trying its hardest to keep up. As other stages in the production cycle are not the constraint, logic says they should not we working as hard to keep up.
However, it might be more difficult to find these constraint points in the production cycle than one might first imagine.
First, very often operators are told to keep producing no matter what the cost so that they are as "efficient" as possible. This leads to over-production at various stages in a system and is known as the "Local Optima" problem (see our Local Optima article). Because everything is busy, it’s hard to find out what is the busy constraint.
Second, the reason for the apparent lack of any idle machinery can be that most production operators very quickly learn that it is not a good idea to be apparently sitting around doing nothing whenever a Boss or Supervisor comes by. They will often make out they are busy so that they appear to be working when in fact they are not working productively (see Parkinson’s Law article). You can often pick workers with time on their hands because they are taking a break, talking to each other or a Supervisor or making small attempts to clean or service their machine. People quickly learn that they should at least appear to be busy or they will end up with more work to do somewhere else or a scolding from a Boss.
Third, the whole science of Cost Accounting is designed to make each part of an operating system function more productively. Cost Accounting can add dramatically to the inefficiencies and expenses of systems. A cost accountant will work out, with seeming precision, what it costs to run parts through a stage of the production process. The uncertainly surrounding cost accounting means that most of this is essentially guesses. Bosses, driven by cost accounting, will try to get those figures as low as possible by producing as much as possible. We are back to the destructive ‘local optima’ problem. Cost accounting makes a single point more efficient rather than making the flow through a production system as productive as possible.
Fourth, other optimising techniques alone may not be optimal. Above we saw how Six Sigma and Lean optimising techniques fared very poorly when compared to the same techniques combined with TOC. These techniques optimise each machine. TOC optimizes an entire production system.
When you can identify a physical constraint, any improvement in the productivity of the physical constraint is almost 100% translated into improved throughput of the whole business because that constraint was what was holding it back. Therefore, a 20% improvement in the constraint means virtually a 20% improvement in throughput throughout the entire system. This demonstrates how great an impact focusing first on the constraint can have on your overall business.
In some production systems, there will be parallel constraints.
By way of example, consider a dental surgery with several dentists. The time taken by each dentist for the same procedure will vary and the flow of different types of procedures to individual dentists will also vary. It would be most productive then to have a Buffer in front of each dentist that is tuned to their particular dental specialty and to the speed at which they work. Because there is usually an ongoing relationship between a dentist and a patient, we need parallel buffers as we can’t merge them into one.
Having said that, sometimes a parallel (and often inefficient ) constraint can be expedited by having a channel or throughput that avoids it.
An example of this might be in a grocery checkout queue where there is a separate line for people with "7 items or less" and a main queue for people with large trolleys of purchases.
Alternatively, most of us have experienced the frustration of separate queues in front of separate supermarket checkouts and the inefficient process that is involved. Many checkout type operations may well proceed much more efficiently if there was a single queue that then breaks off to the separate checkout counters as the next one became available. It is almost certain to improve the temper of the consumers as well!
Management practices elsewhere than the production line may have a significant impact on the throughput. The constraint may in fact not be one of the production elements but rather one imposed by management.
The first group of these are Policy Constraints.
It is quite possible that the company has introduced restricted overtime, for example, as a "Cost Saving Measure". However, if this means that the constraint is not operating as much as it could potentially, and therefore choking off the amount of throughput, paying some overtime or a second shift at the constraint, and thereby increasing its productivity, may lead to dramatically improved throughput. See the example above where a machine operating 100 hours per month was increased to operating 300 hours per month by adding 2 more week day shifts. In that example, we could potentially further increase productivity by having 7 days by 3 shifts on the constraint. It is difficult to know, abstractly, at what point the increased costs of running the constraint overcomes the increased throughput but that would be easily calculated in the real world.
To control inventory costs, the Company Accountant may have limited the amount of money that can be spent on the raw materials necessary for production. If the constraint is lacking one of the raw materials required for it to operate at its maximum capacity, it will be choking off throughput. It may be a false economy to limit the flow of raw materials to the constraint.
As discussed above, there might be a policy requiring all machinery to operate at full capacity all the time. This won't affect the constraint but will affect the amount of inventory and work in progress and the excess of these as a result of local optima will tie up the cash of the company. The business has paid the cost of these raw materials but won't get any repayment of the outlay until the constraint can process all the product.
In some production systems, it might be possible to produce variations on the main product that do not need to run through the constraint and can be sold separately. It may be that your business sells an assembled product but could also sell several parts to other companies who wish to assemble the same product. Providing the parts can be manufactured without going through the constraint, these are additional sales that the company can make that work around the impact of the constraint. This is an example where a focus on Local Optima could safely be ignored.
It could be that your throughput is not limited by any of the processing stages in your business. It might be that the market’s demand for your product is the problem.
Demand might be restricted by the fact that whatever you produce is going out of fashion and there is less and less demand. Even though you can produce plenty of the product, you can't sell it because of falling demand.
A common problem is a sales person’s natural desire to make a sale and therefore agreeing to all sorts of feature bloat to the product which slows down its manufacturing, wastes time and resources determining how to produce the new feature in the product and generally complicates the process. If you have lots of variations on a product, consider applying the 80/20 Rule (see The Amazing 80/20 Rule article) to get rid of the low achievers in the product range.
It may be that the business has too many projects in play at any one time. By trying to work on many projects at once, multi-tasking comes into effect and that is a well-known cause for reduced productivity (see Multitasking article).
There are several types of businesses - especially in the professions - that rely heavily upon referrals rather than advertising to have prospects enter their sales funnel (see Sales Funnel article) and lead to new clients. Examples include lawyers, dentists, doctors and accountants. If their referral sales funnel is not wide enough, it doesn't matter how good their production systems are. They simply won't have enough work.
A similar problem can happen to other professions that do advertise but are not generating sufficient demand for the services that they have. Whereas a professional might be the constraint when they run out of time, they are no longer the constraint when they have time on their hands because their marketing is not working properly.
Disruptive innovation can also cause problems for existing industries and might become a constraint. As examples, consider Uber replacing taxis and social media advertising replacing TV advertising.
It might also be that your product has a finite demand and no matter how well you can produce it, there simply isn't enough demand to purchase it. An example here might be artificial hearts. If you don't need an artificial heart, you are not going to buy one. This type of constraint might be particularly common with specialist equipment.
Human and Equipment Constraints
People can be constraints:
- They simply might not have the skill set or drive necessary to improve your throughput
- The constraint might be you; the business owner. Are you open to new approaches, and opinions? Do you delegate or does everything have to come through you and can only move at the speed that you have time to give attention to the decision?
- Most people don’t like to change. As an entrepreneur, you are an outlier as you like change – that’s why you are an entrepreneur. Most people don’t and will revert to old habits and work methods at every chance until you can get them out of their present groove and into another. Getting them actively involved in the decision making about new processes is a tested way to get ‘buy-in’ to new approaches. After all, they are probably the best informed about the cause of the local problems anyway.
Naturally, equipment can be a constraint.
But often there is only a single piece of equipment that needs to change to improve throughput. And that is the constraint.
Upgrading any other machine is a waste of money. Increasing its productivity will not have any impact on throughput if it is not the constraint-machine.
Even increasing the constraint-machine’s throughput by buying a bigger capacity model is not something you should do at the outset. If you simply increase its capacity, you have not removed the inefficiencies that caused it to be the constraint in the first place. Upsizing is the sledge hammer approach and has not solved the problem.
Even worse, it might mean that through sheer brute force capacity upgrades in machine A, the constraint moves to machine D elsewhere in the production line and you have to start again.
By definition, you are always going to have a constraint. Better to know exactly where it is and focus intensely on it.
Quality and Quantity of Inputs
If you have poor quality inputs, you will spend a lot of time replacing and repairing defective outputs from your production system.
This rework is disproportionately expensive when compared to the original production of the item, it is worth minimising as much as possible.
Quality should be a major focus of your production system and staff should be empowered and actively encouraged to put any defective inputs to one side rather than incorporate them into a production step that later needs to be undone to rectify a fault.
If your inputs are in short supply, that clearly can be a limiting factor. There can be all sorts of reasons for this that we can offer solutions to here.
Suitability of Operating Techniques
Almost without doubt, it will be possible to introduce improvements to operational efficiencies in your production.
Whole methodologies like Lean and Six Sigma have developed to identify and introduce such improvements in efficiencies.
As we saw in the Sanmina-SCI case study above, 89% of the improvements came in 6 factories that combined these techniques with TOC.
In other words, streamlining efficiency anywhere else but the constraint is a complete waste as long as the constraint remains operating at less than its maximum efficiency.
Change Induced Constraints
If your workflow has been around for any length of time, it might have been optimised for a different operating environment that was in place those several years ago.
Time moves on but our workflows do not necessarily change for changing conditions.
If production has started to fall from what it was, possibly one of the following have changed and introduced a constraint that was not there before.
- new quality requirements
- unfortunate by-products of cost cutting initiatives
- new technology
- new machines
- new strategies
- new raw materials or components
- new management with new rules
- new competitors
- new regulations & laws
Summary of Step 1
At the start of this section, we identified six typical groupings of factors that can be constraints.
- Physical constraints or the availability of resources
- Managerial policies and procedures
- Market and marketing based constraints
- Human and equipment capabilities
- Quality and quantity of inputs
- Suitability of operating techniques
When we first introduced the concept of ‘constraints’ above, we suggested that you try and identify the constraint in your business. If you were not able to identify it before, see if you are now able to identify the constraint knowing about the 6 different generic types.
Once you have identified the constraint in your system, the next step is to take advantage of that information and exploit it to maximise the throughput and therefore the profitability of your business.
On the face of it, you could simply add more capacity to the existing constraint. If you like, this is the sledge hammer approach to overcoming the problem. Possibly, in some instances, it might be worthwhile but very often it will come with a necessary increase in investment to finance the additions to the capacity of the constraint. All this means is that the bottleneck remains as a financial constraint on your business even while productivity might go up.
A better approach, at least initially, is to maximise the throughput of the constraint without dramatically increasing the investment that you make into the constraint. That is not to say that you cannot increase such operating costs as labour if necessary.
Another risk of adding more capacity to the existing constraint is that some other part of your production system will become the constraint. Even with investment in the existing constraint, you will nevertheless continue to have a constraint somewhere in the system; it is a fact of life. There must always be some part of the system that is a bottleneck on capacity irrespective of how much you spend on improving capacity.
In the remainder of this section, we will focus on how to take advantage of the existing constraint and maximise its throughput and therefore your profitability.
Remove Unnecessary Flow
There may be work flowing through the constraint that doesn't necessarily need to.
One example of this is defective parts.
If your system results in work that is defective in some way and needs to be discarded, or "re-worked", then running any of this unnecessarily through the constraint simply takes up capacity that could be used for parts that are up to standard. By putting a quality control station - either real or virtual - in front of the constraint to filter out this material you will gain some throughput in the constraint.
If your business is project or innovation related, there are likely to be a number of stages in the project or innovation cycle that will be unlikely to survive later pruning of non-viable projects. Some experiment will not work out and that avenue will be dropped. If you can avoid running likely nonviable projects/innovations through a bottleneck, you will increase the availability of that bottleneck and therefore its throughput. Better to can them before the constraint rather than process them through the constraint and then can them. This type of problem is particularly likely with human resources that are a bottleneck within stages of a project or innovation.
If there is an alternative way of processing the parts that doesn't require them going through the constraint, filtering some of them off to the alternative route would reduce the amount of work at the constraint and, providing the alternative route has surplus capacity, it will not cost any more if that surplus capacity is all that is being consumed.
Sometimes when there is time pressure on getting a product completed people begin expediting and switch the flow of work between products in order (ideally) to produce some of each of the products more rapidly. This is multitasking and is likely to be quite inefficient (see Multitasking article). As soon as you begin to multitask, almost certainly you will introduce inefficiencies into the flow of work through the constraint and thereby increase, rather than decrease, the problems.
Sometimes a business ambitious to get additional sales copies a competitor’s product and launches a me-to-product into a competitive space. A me-to-product is unlikely to sell better than the competing market leader. They have all sorts of resources at their fingertips to compete with you entering their market; not the least of which is the ability to offer price discounts until your attack collapses. In the meantime, if your me-to-product is running through the constraint, it is consuming valuable resources at the constraint which may not turn into sales, and therefore throughput, at the end of the production cycle.
There may be several possible alternative products competing for time with the constraint at any point in time. By prioritising what does make its way through the constraint, we can increase the value of the production and therefore the amount of throughput.
By way of example, if we are producing 5 units per hour with a value of $10 this gives us a total return per hour of $50. Whereas, a higher flow rate of 8 units per hour but a lower return of $5 per unit gives a total return of $40. Therefore, even though the 5 units per hour appears to be less productive, it is in fact more productive than the item with a flow of 8 units per hour. In reality, making these sorts of decisions is not likely to be quite as straight forward as that because eventually all of the products need to get through the constraint irrespective of their total return. However, if there was another way of processing some of the lower return units - perhaps through an alternative stand-by machine - even if that was a slower unit we could still improve productivity.
Some mini scheduling improvements that are likely to be quite productive are:
- Prioritise: try to schedule high priority work first thing in the day so that it is most likely to be completed. During the day, variations, interruptions and Murphy's Law will gradually edge their way into the production cycle. It becomes harder and harder as the day progresses to predict the delivery time to and from the constraint.
- Biggest: try to do biggest piece of work first.
- Uncertainty: do the one with the least predictable outcome first followed by more predicable ones.
That way your day gets simpler as it progresses.
An example of this quoted in various books on TOC in surgery. The most complicated surgery, with the greatest uncertainty about how long they may take and what might go wrong, can be scheduled at the start of the day. Any remaining part of the day can then be chopped up into smaller pieces with surgery that is more predictable so that the day can be fully used.
Alternatively, if the surgery cycle begins with the predictable surgeries first, then at the end of the day it is faced with the larger and less certain surgeries and is therefore much more likely to blow out into overtime to complete the surgical tasks.
There are several methodologies that have been developed to improve the operational efficiency of items of plant and collections of machines. Two examples are Six-Sigma and Lean. Earlier in this article, we reported on a combination of Six-Sigma, Lean and TOC being some 15 times more productive than plants that were tuned with Six-Sigma and Lean only. Clearly then there is room to improve the operation of the constraint by a combined application of these various optimising techniques.
Even before one gets to these sophisticated operating techniques, there are other opportunities to improve the throughput of the system.
Firstly, multitasking should be eliminated in so far as possible. It always introduces inefficiencies that may easily outweigh the gains from attempting multitasking as a solution to the problem (see Multitasking article).
At various stages in its operation, the constraint will need input from various other sources.
If it is a machine, it is going to need some maintenance time. If it is a human resource, they are going to need a break from time to time or they are going to take annual leave.
To minimise the "outages", from a shortage of resources, other staff can be cross trained to jump in as needed to support the constraint or provide a relieving operation. For example, staff could stagger their lunch times if it was important to keep the constraint operating. Stand-by staff would also help to overcome unexpected absenteeism. It might also be possible to juggle a shift change such that some staff from the old and new shift are available and therefore the constraint keeps running.
Given that we have identified the constraint as the most important part of your production system, it makes very good sense to "slave" every other step to feeding the constraint.
Production tasks upstream of the constraint must be designed to always have sufficient product waiting at the constraint to keep the constraint fully operational.
On the other hand, we do not want more output from the upstream stages than necessary because that is simply inventory or work-in-progress. You have paid for the materials and time to produce this inventory and work-in-progress but you will not realise any money from it until it is finally paid for by the consumer.
An element of this that many people struggle with is that we don't want these upstream machines to work at their full capacity; only at the capacity necessary to feed the constraint.
This means that we must make it very clear to the people at these upstream, non-constraint, tasks that it is ok to stop work when no more parts are required. Clearly, we do not want a lot of idle time but it is quite possible that these people can do other tasks or "make and mend". By rearranging who does various upstream tasks it might be possible to have the same operators working on several stages if the total time working on each of those stages is less than their time available.
Rather than setting support services, like maintenance, engineering and training, to be scheduled during prime production time for the constraint these "non-productive" tasks could be scheduled for a time when the constraint does not need to be operating.
Later we will be discussing the TOC concept of ‘drum, buffer, rope” which is designed to ensure the upstream systems keep the constraint well fed and with minimal surplus inventory.
As a rule of thumb at this point in the discussion, professionals suggest that it is likely that operating a constraint in a production system above 85% of its capacity will lead to increased delays and interruptions. This is also supported by the 80/20 Rule (see The Amazing 80/20 Rule article) which argues that the last 20% of a task will consume 80% of the resources. By running upstream tasks at less than maximum capacity we reduce the likely impact that any hiccup in the upstream capacity will have on the throughput of the constraint.
By this stage, we have stream-lined the throughput of the constraint as much as possible. If necessary, we can now consider increasing its capacity by adding additional resources.
Although it might be wise to have a standby resource to replace the constraint if it stops producing for some reason, it is not necessarily the best approach to remove the constraint entirely.
Your system will always have at least one constraint because there will always be at least one thing that controls the throughput of the system.
There might well be a strong argument that "the devil you know" is better than removing one constraint to have another one pop up somewhere else. It may take you some time to realise that the constraint has moved and during that time your throughput may reduce and certainly won't grow any further.
If the constraint does move, then you ideally should start the whole process over again and optimise the new constraint.
You are going to need to remain quite vigilant to ensure that the constraint does not move unexpectedly. You should get used to looking for the signs that a new constraint is developing so that you keep improving the productivity of the system indefinitely.
Because the idea of a constraint will be foreign to many people, and the idea that it is ok for upstream workers to have some idle time on their hands, it is going to take quite a mind shift in both operators and managers to implement a TOC system to manage their constraints.
People do not like change and they will always tend to revert to what they know and what they are used to even if it is demonstrably less productive than the alternative. People resist change because it introduces uncertainty. Small business operators are used to uncertainty and that is why they are the business owner rather than someone working in the business. Trying to get people to manage more uncertainty will be very difficult.
A brute force approach by management may work but at the end of the day it is likely that the results will be far more productive if workers are led to discover what they should be doing themselves.
Very often an operator is aware of the inefficiencies in the work they do but have learned to live with them because that is what they have been "told" to do. Giving them the opportunity to identify these inefficiencies empowers them to take control of their own operations and is likely to turn up a number of solutions that are not going to be obvious to supervisors and managers.
People are always tuned into the largest radio station in the world - WII-FM - which stands for “What's In It For Me”. Irrespective of whether a process is good or bad, people need to be shown why it is in their interest to do something a different way and preferably to discover that different way themselves so that they "own" the process.
There is very little point in undertaking this improvement effort unless we can be sure that it is working and that we are in fact improving the right things.
Elsewhere, we have argued that improving the productivity of a non-constraint will not move the total throughput of the company at all.
Minimally, we need a metric to measure the throughput of the entire production cycle whether it manifests as money, or some other "currency" if you have a Not for Profit operation (see Metrics article).
It would also be a good idea to set some achievable improvement goals (see Goal Setting article) so that people have a feeling of achievement when these goals are reached.
The End of The Beginning
This has been a particularly long article that introduces something Profit Savvy considers to be of fundamental importance for any manager.
In our Pyramid of Priorities for working on your business, the second layer is “optimisation”. By that we mean fixing what you have before rushing off down some other avenue.
TOC is a major optimisation tool and an important part of this second layer. Its siblings are the 80/20 Rule (see The Amazing 80/20 Rule article) and effective Time Management (see Time Management article).
We strongly encourage you to spend some time reflecting on TOC and where the constraint lies in your business. Finding that and tuning your business to maximize the throughput of your business is very likely to lead to substantial improvements in your throughput.
To skip this step and go straight to tweaking the third layer of your business is to go looking for a solution to a problem you have not yet identified.
Further articles can be found at the Theory of Constraints Menu.
Because the concept of the Theory of Constraints has had such an impact on business, there are a lot of resources available for you to further your knowledge and understanding.
Wikipedia – Theory of Constraints.
The founder of the Theory of Constraints, Eliyahu Goldratt, was an innovator on many levels. He pioneered the concept of a business "novel" which used a story of people experiencing real life business problems to demonstrate how improvements could be made. These business novels are very easy to read and very thought provoking. His initial one, "The Goal", has been a world-wide best seller (The Goal: A Process of Ongoing Improvement, January 1992). Allegedly the founder of Amazon, Jeff Bezos, requires his managers to read this book. Subsequent to Dr Goldratt writing his foundation books, in the several areas covered by TOC, a number of other authors have written eBooks. These are easily purchased and downloaded onto a Kindle app on a mobile or desktop device.
The TOC management innovations cover:
- Critical Chain Project Management
- Throughput Accounting
- The Theory of Constraints
- Inventory Management
Depending whether you believe the efficiencies of the Theory of Constraints, there are several simulation games that you can play yourself or you can watch YouTube videos of the simulation games in progress. Some examples include:
- Demo of Goldratt Maintenance Depot Simulation by Dr Alan Barnard - Link to game.
- Beer Game Part 2 - This is part 2 of an abbreviated version of a presentation by Tom Looy of Tacit Knowledge at the Agile 2009 Conference. The presentation is entitled TheBeer Game: Agile, Lean and Theory of Constraints Applied to the Agile Story Card Wall - Link to game.
- Beer Game Part 3 - Link to game.
- Beer Game Conclusion - Link to game.
This is a simulation of the "Die and Matchstick Game" from the book "The Goal" by Goldratt and Cox. WITNESS simulation has been used to illustrate the game Alex Rogo uses to improve his business - Link to game.
There are a large number of YouTube videos on the Theory of Constraints. We have collected a few below that may give you a good introduction:
Theory of Constraints 3 Bottle Demo to Improve Flow suggested by Versli Lietuva - Link to video.
The Theory of Constraints and Throughput Accounting: Matt McCune at TEDxBOZEMAN - Link to video.