The word iterative comes from the Latin itero, iterare: to do a second time, or repeat. In programming, iteration means to repeat a set of instructions, and an iterative design is one where repetition sits at the heart of the process, established as an alternative to the problems of Waterfall's fully upfront planning.

How it differs from Waterfall

In Waterfall, each phase of the development cycle is gated: the whole application has to be designed and pass a phase gate review before coding begins, and coding has to pass its own gate review before testing. Iterative methodology, by contrast, is about continual looping of a design through a sequence of testing conditions to gradually improve it. Feedback is encouraged, and modifications are made in small, incremental steps, so developers can learn from earlier changes as they go.

How the lifecycle works

The model starts by defining and implementing part of the software requirements, then reviewing them to identify further requirements, informing subsequent versions until a final release. Each release is developed within a fixed time block, called an iteration, aimed at a particular set of requirements, with access back to earlier phases where changes are needed. By the end of the cycle, a usable system is produced.

Iterative methodology is an essential building block of frameworks such as Scrum, Kanban and XP, which are in turn central to Agile software development more broadly.

Advantages and disadvantages

On the plus side: problems get caught early, a full specification isn't needed at the outset, the product improves gradually as defects are identified, there's less time spent documenting and more time designing, working functionality can be reached early, and testing within small iterations is easier to control. On the downside: it can still be resistant to change in the original plan if not handled carefully, doesn't automatically build in team communication, and needs strong skill and risk analysis to work well, along with more active management involvement than some teams expect.

When it's worth using

An iterative model suits projects where the major requirements are known but there's scope for additional functionality to evolve, where some aspirational elements of the design may change, where specific resources need targeting to particular iterations, or where the team needs to learn certain technologies as development progresses. It can cope well with changes in requirements and help eliminate design flaws, both of which are common in most software projects, and can provide swift, usable returns for projects with a limited budget.