ties are less flexible because they form the center or the core of the system or organization and changing them will affect all other parts that are built as the stapling layers of an onion.

The decentralized parts and solution on the other hand, operate on the outside and can fluctuate easily

This dilemma is one of the most common management dilemmas. It is not constraint to large organizations that focus -- to name one -- on implementing a shared service center, but also small companies are dealing with this problem ... even a computer programmer faces the same dilemma...

Thus, when the programmer experiences a lot of maintenance because of the many exceptions that bring the decentralized program with it, he or she wil streamline the program and centralize the code. When you have done this, your maintenance will become much less of a problem, but there is a new issue: when there is a change required in the central part of the system, you are affecting all the decentralized routines that are used by the various clients. So where in the previous situation you could update any part without affecting the whole, now in the new (centralized) situation you have lost this autonomy. In this situation, a growing number of users, has given momentum to use a central solution in which scale is more important than client differentiation.

The main cause of this dilemma lies in the fact that centralized activities (and systems, etc) are less agile. Centralized activities are less flexible because they form the center or the core of the system or organization and changing them will affect all other parts that are built as the stapling layers of an onion.

The decentralized parts and solution on the other hand, operate on the outside and can fluctuate easily.

Thus, when the programmer experiences a lot of maintenance because of the many exceptions that bring the decentralized program with it, he or she wil streamline the program and centralize the code. When you have done this, your maintenance will become much less of a problem, but there is a new issue: when there is a change required in the central part of the system, you are affecting all the decentralized routines that are used by the various clients. So where in the previous situation you could update any part without affecting the whole, now in the new (centralized) situation you have lost this autonomy. In this situation, a growing number of users, has given momentum to use a central solution in which scale is more important than client differentiation.

The main cause of this dilemma lies in the fact that centralized activities (and systems, etc) are less agile. Centralized activities are less flexible because they form the center or the core of the system or organization and changing them will affect all other parts that are built as the stapling layers of an onion.

The decentralized parts and solution on the other hand, operate on the outside and can fluctuate easily

when there is a change required in the central part of the system, you are affecting all the decentralized routines that are used by the various clients. So where in the previous situation you could update any part without affecting the whole, now in the new (centralized) situation you have lost this autonomy. In this situation, a growing number of users, has given momentum to use a central solution in which scale is more important than client differentiation.

The main cause of this dilemma lies in the fact that centralized activities (and systems, etc) are less agile. Centralized activities are less flexible because they form the center or the core of the system or organization and changing them will affect all other parts that are built as the stapling layers of an onion.

The decentralized parts and solution on the other hand, operate on the outside and can fluctuate easily

is autonomy. In this situation, a growing number of users, has given momentum to use a central solution in which scale is more important than client differentiation.

The main cause of this dilemma lies in the fact that centralized activities (and systems, etc) are less agile. Centralized activities are less flexible because they form the center or the core of the system or organization and changing them will affect all other parts that are built as the stapling layers of an onion.

The decentralized parts and solution on the other hand, operate on the outside and can fluctuate easily

ties are less flexible because they form the center or the core of the system or organization and changing them will affect all other parts that are built as the stapling layers of an onion.

The decentralized parts and solution on the other hand, operate on the outside and can fluctuate easily. And leave the core untouched. They behave like little waves around the central organization. Which makes it easy to respond to changes. Unless you need something to come from the core. Anyone involved with this dilemmas needs to choose between flexibility and efficiency. Flexibility has a price (the price of overlap), but this price is affordable if the solution is only for a short time and developed in an changing environment. Centralizing activities is required when you operate on scale, when the activities are stable … And this is amongst other thing the case when offering infrastructural services.

Some implications:

If you wait to long with reaping the benefits from centralizing activities, you will loose money because of inefficiencies. It will also be more difficult to change later on, when there is little experience in handing over activities. If you continue to long with a central approach you will become rigid; both in the solution you can offer as well in the confrontation to change. New initiatives are best served with some autonomy. Because of its structural character, Infrastructure should be managed and supplied centrally. Most important of all is the c