The defining features of agile methodology are iterative delivery, incremental value, adaptive planning, customer collaboration, cross functional self organising teams, timeboxed work, continuous feedback and a sustainable pace. Together they describe a way of working that expects change rather than resisting it.
Key Highlights
- Agile is defined by how a team responds to change, not by which ceremonies it runs. A team holding daily stand ups while working to a fixed twelve month plan is not agile.
- Iterative and incremental are different features. Iterative means refining the same thing repeatedly; incremental means adding usable pieces. Agile uses both.
- Self organising teams decide how to do the work. What to build remains a Product Owner decision, and confusing the two causes real problems.
- Adaptive planning does not mean no planning. Agile teams plan more often, in smaller pieces, rather than planning less.
- Working software is the primary measure of progress, which is why percentage complete reporting sits awkwardly with agile delivery.
- Sustainable pace is the most widely ignored feature, and ignoring it produces exactly the quality decline agile is supposed to prevent.
Two Meanings of Feature, Cleared Up First
The word feature carries two distinct meanings in agile, and searches for it return both.
Feature as a work item. In SAFe and similar scaled frameworks, a Feature is a specific size of work sitting between an Epic and a User Story. It is a defined artefact with acceptance criteria and a place in a hierarchy. If that is what you need, our guide to Epic vs Feature vs User Story covers the hierarchy, and SAFe Featurescovers the scaled version specifically.
Feature as a defining characteristic. The attributes that make an approach agile rather than traditional. That is what this article covers.
The rest of this piece deals with the second meaning: what actually distinguishes agile working from the alternative, and how to tell whether a team has adopted the features or only the vocabulary.
If you are heading toward a Scrum Master role, these characteristics are the foundation everything else rests on. OurfreeCSM practice test is a quick way to see how solid your understanding currently is.
The Defining Features
| Feature | What it means in practice | What its absence looks like |
| Iterative delivery | Work repeats in cycles, each refining understanding | One long pass with no repetition |
| Incremental value | Each cycle produces something usable | Nothing usable until the end |
| Adaptive planning | Plans revised as evidence arrives | Plan fixed at the start and defended |
| Customer collaboration | Continuous involvement of the people who will use it | Requirements gathered once, then silence |
| Self organising teams | The team decides how to do the work | Work assigned and sequenced by a manager |
| Cross functional teams | The team holds every skill needed to finish | Constant handoffs between specialist groups |
| Timeboxing | Fixed durations, flexible scope | Deadlines that move when work is unfinished |
| Continuous feedback | Short loops between building and learning | Feedback arrives at the end, or after release |
| Working software as progress | Progress measured by what functions | Progress measured by documents or percentages |
| Sustainable pace | A rate the team can hold indefinitely | Crunch before every deadline |
| Transparency | Work and problems visible to everyone | Status filtered as it travels upward |
| Continuous improvement | The team regularly changes how it works | The same process for years |
The rest of this article works through the ones that are most often misunderstood.
Iterative and Incremental Are Not the Same Thing
These two get used interchangeably and they describe different behaviours.
Iterative means going over the same thing repeatedly, improving it each pass. Build a rough version, learn from it, build a better one. The classic illustration is painting a portrait by sketching the whole canvas roughly, then refining the whole canvas again, and again.
Incremental means building in usable pieces. Finish one part properly, then add the next. Painting the portrait one completed section at a time.
Agile uses both, and this matters because each solves a different problem.
Incremental alone delivers pieces early but assumes the original plan was right. If the first three completed sections turn out to be wrong, they were built correctly and pointlessly.
Iterative alone allows learning but may deliver nothing usable for a long time, since everything stays in a rough state until late.
Combining them is what produces the agile pattern. EachSprintdelivers a usable increment, and the feedback from that increment changes what gets built next. Value arrives early and direction can still change.
A team delivering increments but never changing course based on them has the incremental half without the iterative half. This is extremely common and it looks agile from the outside.
Adaptive Planning Is More Planning, Not Less
The most persistent misunderstanding about agile is that it means planning less.
Agile teams plan constantly. They plan every Sprint, re-plan daily against the Sprint Goal, refine the backlog continuously, and revise longer term direction as they learn. In aggregate that is considerably more planning activity than a traditional project, which plans thoroughly once and then executes.
What changes is the horizon and the commitment. Detail is high for the immediate future and deliberately low further out, because detail invented for work six months away is guesswork that will need rewriting.
This is why traditional status reporting fits badly. Asking an agile team what percentage complete the project is assumes a fixed denominator. In a genuinely adaptive approach the total scope is expected to change, so the percentage measures against a number that is itself moving.
The productive version of the question is what has been delivered, what is being worked on now, and what has been learned that changes the direction. Those have real answers.
Teams that claim to be agile while defending a scope baseline agreed nine months ago have kept the vocabulary and dropped the feature.
Self Organising Does Not Mean Unmanaged
Self organisation is the feature most often misread in both directions.
It means the team decides how to accomplish the work. Who takes which item, in what order within the Sprint, using which technical approach, and how the team's own practices operate. Those decisions belong to the people doing the work, because they hold the most relevant information.
It does not mean the team decides what to build. That is a Product Owner accountability, and priority is set outside the team. Nor does it mean there is no accountability, no standards, or no leadership.
Two failure modes follow from misreading it.
The first is a manager who continues assigning tasks and sequencing work while calling the team self organising. The label changes and nothing else does.
The second is a team that treats self organisation as freedom from direction, resisting priorities set by the Product Owner. That is not self organisation, it is a missing boundary.
Genuine self organising teams operate inside clear constraints: a goal they did not set, a priority order they did not choose, and a quality standard they agreed to. Within those, they decide.
This works only when the team can actually finish work without external help, which is why cross functional composition is a related and necessary feature. A team that cannot deploy without another group is not organising itself, it is queuing.
Working Software as the Measure of Progress
This feature has consequences most organisations underestimate.
If working software is the measure, then documentation, design artefacts, completed analysis and detailed plans are not progress. They may be necessary, but they do not count as delivery.
The practical effect is that a team six weeks in with extensive documentation and nothing running is behind, even though it may feel productive. Traditional approaches would record substantial progress there.
This is also why the Definition of Done matters so much. Without a shared standard for what working means, teams declare items complete that are not tested, not integrated, or not deployable. Progress then becomes a matter of opinion, and the feature stops functioning.
The measure has a useful side effect: it is very hard to fake. Documents can describe intentions. Software either runs or it does not.
Sustainable Pace, and Why It Gets Ignored
The Agile Manifesto asks teams to maintain a constant pace indefinitely. It is the feature organisations abandon first.
The reasoning behind it is practical rather than ethical. Tired teams write worse code, make more mistakes, and take shortcuts that create rework. A fortnight of overtime frequently produces a month of defects. Speed borrowed under pressure is repaid with interest, which is the same dynamic as technical debt.
Recognising the failure is straightforward. If the last two days of every Sprint involve long hours, the team is not delivering a Sprint's worth of work, it is delivering slightly more than that and covering the gap with unpaid effort. Velocity then reflects the overtime rather than the capacity, and planning built on it is systematically wrong.
The correction is uncomfortable because it means taking less into a Sprint and letting the shortfall be visible. Most teams prefer the overtime to the conversation, which is precisely why the feature disappears quietly.
Continuous Feedback and Improvement
Two related features, and both depend on acting rather than gathering.
Continuous feedback means short loops between building something and learning whether it was right. Sprint Reviews with real stakeholders, releases to actual users, automated tests that report in minutes rather than weeks. The purpose is to shorten the distance between a mistaken assumption and the discovery that it was mistaken.
Continuous improvement means the team regularly changes how it works. The Retrospective is the mechanism, and the feature only exists if something actually changes as a result. A team running retrospectives for a year while working identically to how it worked at the start has the ceremony without the characteristic.
The pattern here repeats throughout this article. Every one of these features has a visible form and an actual behaviour, and organisations adopt the visible form far more readily than the behaviour. That gap is where most disappointing agile transformations live, and it is a large part of what the agile mindset refers to.
Timeboxing and Transparency
Two features that support all the others, and both are structural rather than cultural.
Timeboxing fixes the duration of an activity and flexes the scope inside it. Every Scrum event has a limit, the Sprint being the outermost. Its purpose is not efficiency but forcing a decision at a known point, since work otherwise expands to fill whatever time exists. Applied properly,timeboxing is what keeps adaptive planning from becoming indefinite deliberation.
The feature fails when teams treat the limit as a deadline and extend it whenever work is unfinished. Once that happens the constraint no longer constrains, and planning loses its rhythm.
Transparency means the state of the work is visible to everyone who needs it, including the uncomfortable parts. A board showing real status, a backlog anyone can read, impediments raised rather than absorbed quietly.
Transparency also has a lower profile form that matters day to day: estimates presented as ranges rather than single numbers, risks named while they are still cheap to address, and unfinished work shown as unfinished rather than nearly done. Each is small and each removes a place for problems to hide.
Its practical test is simple: does bad news travel as fast as good news? In many organisations status improves as it moves upward, with each layer softening the message slightly. By the time it reaches a decision maker the picture is wrong, and decisions get made on it.
Transparency is a precondition for the rest. Adaptive planning requires knowing what is actually true. Continuous improvement requires naming what is not working. Self organisation requires the team seeing its own performance honestly. A team without transparency can perform every other practice and still not learn anything, because the information the practices depend on is distorted before it arrives.
Customer Collaboration in Practice
The feature that organisations most often claim and least often implement.
Customer collaboration means the people who will use the thing are involved throughout, not consulted once at the start and again at the end. In practice that means attending Sprint Reviews, being available for questions during the Sprint, and having their feedback change what gets built.
Three substitutes commonly appear in its place.
A proxy who represents the customer without access to them. A business analyst relaying requirements is better than nothing and is not collaboration, because the feedback loop still runs through an intermediary who is guessing.
A single upfront requirements exercise. Thorough, well documented, and treated as settled. This is the traditional approach with agile scheduling layered over it.
Internal stakeholders standing in for users. Common in enterprises, where the people in the room are colleagues rather than the people who will actually use the software. Their opinions are useful and are not user feedback.
The reason this matters more than it appears is that every other agile feature assumes the feedback is real. Iterating on inaccurate feedback converges confidently on the wrong answer, and does so faster than a traditional approach would.
Getting genuine access to users is frequently an organisational obstacle rather than a team one, which makes it a legitimate impediment for a Scrum Master to raise. It is also one of the more valuable things the role can fix, and it is covered inCSMCertification Training as part of stakeholder engagement.
How to Tell Genuine Agile From the Label
Six questions separate the two, and none of them ask about ceremonies.
When did the plan last change because of something you learned? If the answer is never, adaptive planning is absent regardless of how the work is scheduled.
What did you deliver in the last month that someone actually used? If nothing, incremental delivery is not happening.
Who decides the order in which the team works on Sprint items? If a manager, self organisation is not happening.
Can the team finish an item without another team? If not, cross functional composition is missing and dependencies will dominate.
What changed as a result of the last three retrospectives? If nothing, continuous improvement is ceremonial.
When did the team last work late to hit a Sprint commitment? If routinely, sustainable pace has been abandoned.
Teams answer these uncomfortably more often than not, which is useful. The purpose is not to grade anyone but to identify which specific feature is missing, because that is a far more actionable finding than a general sense that agile is not working here.
Building the capability to run this diagnosis and act on it is a substantial part of the Scrum Master role, and it is covered in CSM Certification Training alongside the framework itself.
Where These Features Come From
The features are not arbitrary. They trace directly to theAgile Manifesto, written in 2001 by seventeen practitioners who had independently arrived at similar conclusions about why software projects failed.
The Manifesto states four value preferences and twelve principles. The features described in this article are what those look like when a team implements them.
Customer collaboration over contract negotiation produces the customer collaboration feature. Responding to change over following a plan produces adaptive planning. Working software over comprehensive documentation produces the progress measure. Individuals and interactions over processes and tools produces self organising teams.
Worth noting is what the Manifesto does not say. It does not prescribe Sprints, stand ups, story points, backlogs or any specific practice. Those come from individual frameworks such as Scrum, Kanban and XP. The features are the underlying characteristics, and the frameworks are competing implementations of them.
That distinction explains why teams can follow a framework precisely and still not be agile. The practices are a means of producing the features. Performing the practices without producing the features is the most common failure in agile adoption, and it is what most agile myths are built on.
Which Features Matter Most
If a team can only fix one thing, some features return more than others.
Start with incremental delivery. A team producing something usable every couple of weeks will discover most of its other problems naturally, because the attempt to finish something exposes dependencies, unclear requirements and quality gaps.
Then cross functional composition. Without it, incremental delivery is blocked by handoffs no matter how well the team works internally.
Then continuous feedback. Delivering increments nobody looks at produces the illusion of agility without the learning.
Then the rest. Adaptive planning, self organisation and sustainable pace are important and tend to follow more easily once the first three are working.
The sequencing matters because these features reinforce each other. A team delivering usable increments every two weeks generates the feedback that makes adaptive planning possible, and adaptive planning is what makes self organisation meaningful rather than decorative. Attempting them in the wrong order produces effort without compounding benefit, which is why so many adoptions plateau after the first few months.
The common mistake is starting with ceremonies, because they are the easiest thing to introduce. Stand ups can begin on Monday. Cross functional team design requires an organisational decision. Teams reach for what is available rather than what would help, which is understandable and largely explains why so many adoptions stall at the surface.
How These Features Appear in Different Frameworks
The features are shared. How each framework produces them varies, which is why teams moving between them recognise the underlying ideas but not the mechanics.
| Feature | In Scrum | In Kanban |
| Iterative delivery | Sprints of a month or less | Continuous flow, no fixed iteration |
| Incremental value | A Done increment each Sprint | Items released as they finish |
| Timeboxing | Every event has a fixed limit | Applied to meetings, not to work |
| Adaptive planning | Re-planned each Sprint and daily | Re-prioritised continuously |
| Limiting work in progress | Implicit through the Sprint Backlog | Explicit WIP limits per column |
| Continuous improvement | Sprint Retrospective | Regular review of flow metrics |
Scrum produces the features through cadence. Fixed length Sprints create regular points where planning, review and improvement happen whether or not anyone remembers to schedule them.
Kanban produces them through flow and limits. There is no Sprint boundary, so improvement and re-prioritisation must be deliberate rather than automatic, and work in progress limits do the job the Sprint Backlog does in Scrum.
Neither approach is more agile than the other. They are different mechanisms for producing the same underlying characteristics, and the choice usually comes down to whether work arrives in plannable batches or as a continuous stream of varied requests.
The practical implication is that copying practices between frameworks rarely works. Adding a two week planning cycle to a Kanban team without the rest of Scrum's cadence tends to produce the overhead without the benefit. The features are portable, the practices are less so.
Frequently Asked Questions
1. What are the main features of agile methodology?
Iterative and incremental delivery, adaptive planning, customer collaboration, self organising cross functional teams, timeboxing, continuous feedback and improvement, working software as the measure of progress, transparency and a sustainable pace.
2. What is the difference between iterative and incremental?
Iterative means repeating over the same work to refine it. Incremental means building in usable pieces. Agile uses both, so each cycle delivers something usable and what is learned changes what comes next.
3. Does agile mean no planning?
No. Agile teams plan more frequently in smaller amounts. What changes is that detail is high for near term work and deliberately low for distant work, and plans are revised as evidence arrives.
4. What does self organising actually mean?
The team decides how to do the work, including who takes what and in what technical approach. It does not mean the team decides what to build or in what priority, which remains a Product Owner accountability.
5. Is a Feature the same as a feature of agile?
No. A Feature with a capital F is a work item sitting between an Epic and a User Story in scaled frameworks. A feature of agile is a defining characteristic of the approach itself.
6. How can you tell if a team is genuinely agile?
Ask when the plan last changed because of something learned, what was delivered recently that someone used, and what changed after the last few retrospectives. Ceremonies tell you very little on their own.
7. Is agile only for software teams?
It started in software and the features apply anywhere work is complex and requirements are uncertain. Marketing, hardware and operations teams use them successfully. The features that transfer least well are those assuming rapid, low cost iteration, which is harder where each increment has a physical manufacturing cost.
8. Which agile feature is most commonly missing?
Sustainable pace, followed closely by genuine continuous improvement. Both are easy to abandon quietly because neither failure is immediately visible.
9. Can a team be agile without using Scrum?
Yes. Scrum is one implementation of these characteristics and Kanban, XP and others produce them differently. What matters is that the features are genuinely present, not which framework produced them.
10. How long does it take to develop these features?
Practices can be introduced in weeks. The characteristics take considerably longer, typically several quarters, because most of them require changes to how decisions are made and how teams are structured rather than to what the calendar looks like.
11. Do all agile frameworks share these features?
Broadly yes, though they implement them differently. Scrum uses Sprints and prescribed events, Kanban uses flow and work in progress limits, and both aim at the same underlying characteristics.
Closing Thoughts
The features of agile describe a way of working rather than a set of meetings. Iterative and incremental delivery, planning that adapts to evidence, teams that organise their own work and hold every skill they need, progress measured by what functions, and a pace that can be maintained.
The reason this distinction matters is that practices are easy to copy and features are not. An organisation can install Sprints, stand ups and a backlog in a fortnight and change almost nothing about how decisions are made or how the plan responds to reality.
A useful exercise for any team is to work through the twelve features in the table above and mark each as present, partial or absent. Most teams find three or four genuinely absent, and those are usually the source of whatever frustration led them to ask the question in the first place.
The value of doing it as a team rather than individually is that people disagree about which are present, and the disagreement is the useful part. One person marking customer collaboration as present while another marks it absent usually means the team is talking to a proxy and has never examined the difference.
If you want to build these characteristics properly rather than adopting the surface, CSM Certification Training from an accredited Scrum Alliance provider covers the framework and the practices that produce them. Request the course curriculum to see the full agenda and upcoming dates, or start with our free CSM practice testto gauge where you are now.


























