In many businesses with well-filled order books, I often come into contact with purely capacity oriented thinking. There is so much demand that it can neither be met with the current pool of qualified employees nor the amount of material resources at hand. The solution seems to be perfectly clear: increase capacity. There is the notion that capacity is something like a tap which can be turned on and off, and the problem is solved – this just happens to be the business model of temporary work agencies as well as some consulting firms. Unfortunately, things are not that easy and especially not in knowledge work.
The building up of capacity is indeed a crucial investment in the future which definitely has a positive effect, but this effect only takes hold much further down the line. Constantly increasing capacity on a short-term basis due to shortages is often accompanied by effects that are not intuitive and are thus often ignored: Sooner or later, businesses find themselves in the same situation again even though they have increased capacity. The whole situation of having more projects in progress than capacity and not being able to keep to project deadlines starts all over again. Or as Don Reinertsen und Preston Smith put it: „If we add more resources, sooner or later we will be back in the same situation with more projects than we have the resources to handle, and we will be diluting our effort and delaying projects again.“ The problem is that more work is being started than can be finished – the work within the system is not really limited.
There are two crucial aspects when considering how additional capacity will affect a work system: Throughput and Lead time. The throughput is the amount of work done within a time period. The more throughput found within a work system, the more one can sell – the throughput is thus an economical aspect. The lead time, on the other hand, tells us how quickly one job is completed between two measured points. Time-to-market describes the lead time when measured from idea to completion. Now, what happens when capacity is increased within a work system, for example, by hiring additional employees?
Quite probably, the aspect of performance will paint a rather ambiguous picture. Initially, throughput will increase because the additional employees will be able to complete more orders. But: the lead time will also increase – the work system will become slower and the customers will wait longer for their products. There are two reasons for this:

1. Initial phase: Each new employee needs some time until he is somewhat familiar with the procedures and processes of the company he is new to and until he has familiarized himself with a product or project.
2. Friction losses: If more employees are added to a work system, then the complexity increases as well because the amount of coordination also increases. Often additional capacity goes hand in hand with more focused specialization which, in turn, increases the number of handoffs.

 
Friction loss especially influences the lead time. With the increasing of capacity, one tries to optimize active work. From my own personal experience, I can say that active work only makes up about 2 to 20 percent of the actual lead time. The lead time is made up mostly of waiting times which occur during the handoffs to the individuals involved in the work.
An example: Let us assume that the lead time of a job is 100 days. Now, let us assume that the active work time is approx. 20 percent. Within the graphic below, we see how this active work time is distributed over the entire lead time. Again and again people have to wait for suppliers, information is missing or there is simply no more room in the input queue of the next department responsible and so on.

In a condensed form, the following bar expresses just how small the active work portion really is.

Again, let us assume we are in the fortunate situation that customers are ordering from us like crazy and, for this reason, we have to increase our capacity by 100 percent for this project – twice as many employees are now working for this project. Even if there were absolutely no friction losses due to the additional complexity and the new employees are able to go straight to work without any limitations, the lead time will only be reduced by a maximum of ten percent even though the work time has been cut in half. This effect could be graphically illustrated like as follows:

This minus of 10 percent, however, only represents exactly the case that will never happen: the best case. The next figure illustrates the normal case: Work time is cut in half, but the lead time increases due to processes becoming more complex and because coordination efforts increase. Here, the only question is: How high is the increase to the lead time?

 
This game can only be played to a certain extent. At some point, the lead time has increased so much through continuously adding capacities that it takes forever until work is completed. Guess who is not going to be happy about this: the customer. From his perspective, it is the lead time that counts. This is the dilemma in which many large companies find themselves when they continuously increase capacity: More throughput means more income, but you can easily lose sight of the lead time itself. This is exactly what should not happen, when you are looking to optimize your position in the market.
The problem is:

• In most businesses, it is not clear what is being worked on from the portfolio to the team level.
• More work is being started than can be finished. Thus, the system expresses a starting and not a finishing behavior.
• Furthermore, there is no feeling for what is active and inactive work.
• In the sense of optimization, the focus is always on active work – efficiency -, but people seldom work on optimizing the waiting time.

If someone only increases capacity within an environment, that person sub-optimizes his time-to-market.
What could be a solution?

• You need effective work control above and beyond the boundaries of the team – Kanban on Flight Level 3 or 4 can help.
• The work within the system must be limited and the system must express a finishing behavior. “Stop starting, start finishing” has to be the motto.
• Inactive work and waiting times have to be made apparent.
• The economy of waiting times must be understood so that not only the active but also the inactive work is optimized.