Summary

Project-based organizations (PBOs), undertaking unique, novel, and transient work to deliver bespoke products or services to their customers, require a different approach to their management than the functional, hierarchical, line management approach that has been successfully adopted by organizations with stable products and technologies. The functional organization can adopt an internally focused, discrete approach with:

• Well-defined job roles, designed to meet the internal needs of the organization

• Stable command and control structures

• A focus on increasing efficiency.

The PBO needs to adopt versatile, customer-focused processes to do the work required to meet the customers' bespoke requirements. We have observed that, within these constraints, PBOs adopt different strategies for their management approach, dependent upon the size and number of projects undertaken and the size and number of clients they service. They also adopt different tactical approaches, particularly the type of team structures used, for different stages in the project life cycle.

Organizations undertaking a few large projects tend to adopt traditional project management approaches, whereas organizations undertaking many smaller projects tend to group the projects into larger programs of work and adopt emergent program management techniques. Because the large projects truly are transient, bespoke command and control structures are adopted project by project. It is also appropriate to adopt isomorphic team structures appropriate to the stage of the project. Organizations undertaking programs of small projects find the programs are more continuous in their existence, and so they can adopt more stable command and control structures. They tend to use specialist, surgical teams to deliver the individual projects.

Organizations with a few customers tend to have a one-to-one relationship between the projects or programs and their customers. This becomes inappropriate where there are many customers, because the number of interfaces is the product of the number of customers and number of projects or programs. To avoid this, organizations create an "internal market", so the projects or programs supply the internal market, and that supplies the customers. This reduces the number of interfaces to the sum of the number of customers and the number of projects or programs. We speculate that some form of transactional analysis (economy of transaction and organization) would determine when it was appropriate to make the switch from one-to-one relationships to the internal market.

Organizations undertaking many projects for many customers tend to operate this internal market in many ways, depending on whether the size of project or number of customers dominates.

• The virtual factory adopts the isomorphic structures of the firms (in a versatile network) undertaking a few large projects for a few dominant customers, but using the broker rather than the project manager to manage the interface with the client.

• Companies undertaking many related internal development projects to be used by many internal departments create large programs working for those many customers, with a program director managing the program and a promoter (champion) managing the interface with the many users.

• One company undertaking many less-related development projects, created an internal market so that the projects had just one customer for whom to work. They deliver the product to that customer, and the customer passes it on to the front office.

• Matrix management tries to design a solution specific to this scenario, but when the operation becomes large, the need to maintain customer focus means that individual project teams need to be limited to doing a few projects for a few customers.

One consistent message coming from the firms we have studied and projects we have observed is organizations need to recognize different approaches and when they are appropriate. They should adopt different approaches for different businesses within a larger group, and try not to adopt a one-size-fits-all approach.

Elsewhere (Turner and Keegan 2001), we have given a transaction cost perspective of the governance structures adopted to manage these different scenarios. We now plan to investigate appropriate project contract types associated with these governance structures, and identify learning strategies adopted by clients and contractors to carried forward learning from unique, novel, and transient projects to long-term relationships, especially those associated with partnering and alliances.