High usage and little risk with agile development
Conventional project management has its limits
Frequently a detailed planning, including all the working steps and functions to be implemented, is developed for web projects. In practice, however, this approach is seldom successful. About every fifth project (18 percent) fails; almost half miss their deadlines and/or exceed budgets by a long shot according to a study by the Standish Group on the success rates of IT projects.
The reason? The so-called ”waterfall method“, where all the working steps and results are prescribed in detail, is rarely suited for software projects. On the one hand, many requirements cannot be foreseen and, on the other hand, the precise amount of effort for the individual steps can hardly be predicted. Agile project on the other hand follows the approach that matches human nature best – adaptive planning.
Sprints & feedback loops for constant optimization
With agile development the software is continuously planned, tested and developed. The aim is to attain an executable version of the software as quickly as possible. This version is then extended and improved at regular intervals together with the customer until a status is achieved that is satisfactory for everyone.
The individual objectives are described in so-called stories, following the pattern “As xxx I can xxx to achieve xxx.” The stories are then implemented in accordance with their priority in short manageable steps, so-called sprints.
Together with our customers, we at Infopark manage the prioritization of the stories. The conclusion of every sprint results in a new, executable version. In the subsequent acceptance test it is determined, together with the customer, if the specifications have been fulfilled and which functions should be implemented in the next sprint.
Thanks to these feedback loops, we are able to continuously improve the product characteristics. Furthermore, it becomes clear early on whether or not the originally formulated targets have to be modified or surrendered because the expected benefit is too low.