Agile Software Development
What is Agile Software Development?
Agile software development is a group of methods, including Scrum, XP and Kanban, in which cross-functional teams ship working code in brief iterations and revise plans from user feedback, unlike sequential waterfall projects.
Updated November 2026 · 3 min read
Agile software development is a family of approaches to building software in which cross-functional teams deliver working code in short iterations, get feedback from users and stakeholders, and adjust their plans continuously. It contrasts with sequential or "waterfall" development, where requirements, design, coding and testing happen in long, separate phases and users see the product near the end.
The term was established by the Agile Manifesto in 2001, which brought together several existing lightweight methods under one set of values and principles. Earlier roots include iterative and incremental development practices used since at least the 1970s, and specific methods that appeared in the 1990s, including Scrum, Extreme Programming (XP), DSDM and Crystal.
Main Agile software development methods
| Method | Focus | Distinctive practices |
|---|---|---|
| Scrum | Team process and planning | Sprints of one month or less, Product Owner, Scrum Master, sprint planning, review and retrospective |
| Extreme Programming (XP) | Engineering quality | Test-driven development, pair programming, continuous integration, small releases, refactoring |
| Kanban | Flow of work | Visual board, work-in-progress limits, measuring cycle time |
| Lean software development | Removing waste | Small batches, fast feedback, deciding as late as is responsible |
| Crystal | Fitting process to team size and criticality | Lighter or heavier variants depending on the project |
Many teams combine them, most commonly Scrum for planning with XP engineering practices, or Scrum with Kanban (Scrumban).
The Agile development cycle
- Refine. The team and product owner clarify and size upcoming backlog items.
- Plan. The team picks a goal and the work for the next iteration.
- Build and test. Code is written, reviewed and tested continuously, with automated tests running on every change.
- Integrate and deploy. Changes merge often; with continuous delivery they can go to production at any time.
- Review. Stakeholders see working software and give feedback that changes the backlog.
- Retrospect. The team decides one or two changes to how it works.
Engineering practices that make Agile work
Short iterations put pressure on code quality. Without good engineering habits, each sprint adds defects and technical debt until delivery slows. The practices that let agile teams sustain their pace include automated testing at several levels, continuous integration, code review or pairing, regular refactoring and a strict definition of done. These are why the manifesto includes "continuous attention to technical excellence" among its principles.
An example
Consider a hypothetical team of seven building a mobile banking feature that lets customers freeze a lost card. Working in two-week sprints, it delivers a basic freeze and unfreeze in sprint one, released to staff only. Sprint two adds a confirmation screen after staff testers accidentally froze their own cards. Sprint three releases to 5% of customers behind a feature flag, so the rollout can be widened or switched off without a new deployment. Support data shows customers also want to report the card stolen in the same flow, so that moves to the top of the backlog for sprint four.
After eight weeks the feature is live for all customers with the stolen-card option included. A sequential plan would have specified the feature without the confirmation screen or the stolen-card flow, because nobody knew they were needed until real people used it.
Common pitfalls
- Mini-waterfalls inside sprints. Designing in week one, coding in week two and testing in the next sprint keeps the old handoffs.
- No real users in the loop. Reviews attended only by internal managers miss the feedback that justifies short cycles.
- Velocity as a target. Pressure to raise story points inflates estimates rather than output.
- Neglecting technical debt. Teams that never reserve capacity for maintenance slow down sprint after sprint.
Agile software development is the origin of the wider idea of agile, overlaps heavily with DevOps and is measured in practice through agile metrics.
Frequently asked questions
What is the difference between agile and waterfall software development?
Waterfall development runs requirements, design, coding and testing as long, separate phases, so users see the product near the end. Agile software development has cross-functional teams deliver working code in short iterations, gather feedback from users and stakeholders and adjust plans continuously. Agile surfaces needs that nobody predicted up front, such as a confirmation step users only ask for after trying the feature.
Which methods are used in agile software development?
The main agile software development methods are Scrum for team process and planning, Extreme Programming for engineering quality, Kanban for flow of work, lean software development for removing waste and Crystal for fitting process to team size and criticality. Many teams combine them, most often Scrum planning with XP engineering practices, or Scrum with Kanban in what is called Scrumban.
Which engineering practices does agile software development depend on?
Short iterations put pressure on code quality, so agile teams rely on automated testing at several levels, continuous integration, code review or pair programming, regular refactoring and a strict definition of done. Without these habits, each sprint adds defects and technical debt until delivery slows. That is why the Agile Manifesto lists continuous attention to technical excellence among its principles.
Related terms
Agile Manifesto
The 2001 founding statement of agile's four values and twelve principles.
Scrum
An agile framework of short Sprints, three roles and regular inspection.
Kanban
Visualizing work, limiting work in progress and improving the flow.
Crystal Agile Framework
Agile methods sized to team size and the cost of failure.