Agile Approaches in Projects
With agile methods, such as Scrum and Kanban, a project is executed step by step with a self-organizing, interdisciplinary team in cycles (sprints). The idea is to keep a project lean by prioritization, to implement customer wishes quickly and to be able to respond flexibly to changes even in late project phases.
How do agile and classical approaches differ?
In classic or sequential project management, requirements are collected once and fully documented. Based on this, the next steps take place. The full benefit is only visible at the end. Changes that are required in a later stage often cause delays, risks, high costs, and effort.
Scrum, by contrast, relies on incremental, iterative development in predefined time windows, called sprints. After every sprint a result is visible. A backlog contains all the requirements, but these can be changed over time. This allows for the flexible adaptation to new insights and environmental conditions (market, goals, etc.).
When to use agile methods in projects?
Projects that cannot be planned consistently due to complexity, the high dynamics of changes in requirements, technologies, market conditions, and the environment call for an agile approach. The Stacey complexity matrix helps to classify the applicability of agile and classical approaches.
The meaning of the areas in the diagram:
Simple: results and developments are easy to predict and retrospectively, no new insights arise. Classic approaches can be used.
Complicated: If requirements become less clear or the technology becomes riskier, we move into the complicated area. If necessary, prototypes must be built, technology alternatives weighed, special know-how created, or experts consulted. Classic approaches still can be used.
Complex: The un-clearer the requirements and the riskier the technologies become, the more we move into the complex area where cause-and effect relationships cannot be predicted. The change dynamic is high. Here you can use agile procedures like Scrum.
Chaotic: In this area, things happen rather by accident. Cause and effect relationships are neither predictable, nor identifiable. For example, gambling falls into this category. However, through appropriate measures (lean start-up, hypothesis building, experimentation, learning), projects can be moved out of this area by using well-known technologies and creating more clarity in the requirements.