New Product Development as a Complex Adaptive System

"To design and to manage processes accordingly will
require comprehension of how decision rules, agent variety,
and agent connectivity influence the pace of process
self-organization and emergence" (McCarthy et al. 2006: 453)

New product development (NPD) is the process of bringing new products (and services) to the market. It involves identifying needs, generating ideas and then producing designs which are prototyped, tested and brought to market. Given that these are the fundamental steps, what is the significance of the way the process is organized in terms of different behaviors and associated outcomes? This is a question I explored with colleagues in a research project that was published in the Journal of Product Innovation Management.

We identified three classic frameworks used to organize new product development processes: linear, recursive, and chaotic (see Table 1). These frameworks are used by managers and the members of new product development teams to structure, understand and control different process behaviors, so as to generate different innovation outcomes.
Table 1 –New product development frameworks (adapted from McCarthy et al. 2006)

Linear NPD Process
Recursive NPD Process
Chaotic NPD Process
Process structure
- Relatively fixed, discrete and sequential process stages.
- Connections, flows and outcomes that are comparatively deterministic.
- Concurrent and multiple feedback loops between stages.
- Iterative behaviors and outcomes that are more difficult to predict.
- Initial process stages appear to be random, while the latter stages are relatively more stable and certain.
- Highly unpredictable and unstable.
Process control
- Highly planned using backward scheduling with reviews and decisions at each stage.
- Driven by market pull forces.
- Dynamic, nonlinear and fluid in nature.
- Driven by market pull and technology push forces.
- Goal directed, autonomous, emergent and rule-breaking in nature.
- Driven by technology push forces.
Process outcomes
- Incremental and continuous innovations produced on time and within budget.
- Radical and continuous innovations that often deviate from time and cost constraints.
- Highly radical and discontinuous innovations that are free from time and cost constraints.

It is generally assumed that it is impossible or unlikely that an individual NPD process can switch or adapt between these frameworks to produce different behaviors and outcomes at different times. However, in the companies we studied, we found evidence to the contrary. In particular, we observed that the NPD process was likely to adapt:
  •  as the changes and levels of instability imposed on the process by its managers and customers increase;
  •  when the number and diversity of people in the process increase;
  • when the connections and interactions between these people increase; and
  • as the informality of the process rules and decision stages increases.
These observations are interesting because they show that an NPD process, given the right circumstances, can exhibit the complex adaptive system properties of nonlinearity, self-organization and emergence. Together these three properties are the basis of adaptability in any complex adaptive system (CAS). In the case of the NPD process they provide the “ability to switch or toggle between behaviors that range from linear to chaotic to produce corresponding innovations that range from incremental to radical” (McCarthy et al. 2006: 441) (see Figure 1). Armed with this knowledge, companies can adjust the make-up and control of an individual NPD process to generate different CAS properties and associated innovation outcomes.

 Figure 1 – New Product Development as a Complex Adaptive System

In sum, a complex adaptive view of new product development allows us to understand which features of the NPD process (and many other organizational processes) can be designed and adjusted to generate different behaviors and outcomes. For a process adaptability to occur, its agents (people) should be significant enough in terms of their number, diversity and connections, while having the ability to act freely and in unpredictable ways.

Check out my related talk:

Further Reading