The product architecture will create needs for more intense communications around the undefined interfaces within the design. The less well-characterized the interface, the more likely we are to have questions about it. Thus, an architectural interface will act as a region of the design which requires heightened and accurate communications. This heightened communications cannot be achieved if we place such an interface at either an organizational boundary or a geographic boundary. Communications will be worst if we compound the problem by placing and undefined interface on top of both an organization boundary and a geographic boundary.

Almost every company would like to improve communications on their development teams. Too often they think this requires getting more information to flow faster inside the organization. More people are invited to meetings. More decisions are documented. More memos are sent to more people. Do communications improve? No Ironically, the difference between companies that claim to have good communications and companies that have bad communications is never the quantity of information that is exchanged. Companies that have bad communications can still be buried in information. The key driver of good communications is the amount of usable information communicated compared to what is needed to make good communications.

A boundary is a formal way of dividing authority in an organization. There are two ways of communicating boundaries inside the organization. Unfortunately, the most common approach is trial and error. Some people refer to this as discovering the location of the invisible electric fences... There are two practical problems with using trial and error. First, from a technical perspective, it is slow. Second, touching electrical fences does not encourage the team to use initiative in making decisions.

Like any system, the organization is composed of parts that have relationships to each other, and these relationships or interfaces determine the value of the organization. Without them the efforts of a group of people could never exceed the sum of their individual abilities. We create relationships between elements of an organization because we believe we will create some value that is not present in the individual parts. If we create this value, then we created a useful organization. If we do not we have failed in designing the organization.

After many years of working in product development, I have concluded that the idea of best practices is a seductive but dangerous trap. Best practices are only 'best' in certain contexts and to achieve certain objectives. A change in either the context or the objective can quickly transform the 'best practice' into a stupid approach. For example, a dedicated prototype facility may be a superb practice in one business but an economic disaster in another... The great danger in 'best practices' is that the practice can get disconnected from its intent and its context and may acquire a ritual significance that is unrelated to its original purpose. This distortion appears to arise because of our tendency to start believing the label 'best' as an absolute judgment. Why would you choose not to do something the 'best' way? It requires great mental clarity to recognize that the label 'best' might be misleading.