I recently was quoted as saying, I’ve yet to encounter an organization that was excited to put in the time for agile to be successful when they first get started.
Often Scrum is adopted, and subsequently adapted for what the organization can tolerate in terms of the time commitment. So we are all talking about the same thing here, Scrum meetings consist of daily stand up (or scrum, or huddle, depending on your implementation), sprint planning, sprint demo, backlog grooming, retrospective, in some implementations, mid-sprint review/checkin.
General consensus for standup or scrum meetings are that it should last 15 minutes. Hardly anyone disagrees on this, though lots of people have trouble implementing this. Run well, standup shouldn’t go over 15 minutes. If they do, it’s time to get curious.
When it comes to sprint planning, while there is some disagreement on the amount of time groups should spend, I can almost guarantee an hour is unlikely to be sufficient. Lots of businesses do not want to spend more time however. If you’re only spending an hour on sprint planning, you’re losing out on a lot. You’re crippling your team.
There are a few theories on the appropriate amount of time, from 4 hours for a 2 week sprint, to 1 day for a 1 month sprint. I’ve found 2 hours *can* work, however, it depends on the work being discussed, and the team’s need.
Backlog review can take an hour, or it could take more depending on the backlog, the product owner and the expertise in the room.
Retrospectives should take as long as your sprint planning meetings should. You don’t get into the heart of how a team can do better or the problems that they are facing in the course of an hour. I’ve found two hours is really the sweet spot for organizations unwilling to spare more time. It’s enough time to knock out a check in activity, get the team warmed up to discuss things and start digging into the real things that get in their way or help them succeed. You can wrap up with a check out activity and sometimes don’t even need all that time.
It’s an interesting quandary. I see organizations that say they want to get better and see the benefits, but they don’t want it to cost them any time. Things worth doing rarely work that way.
A team I have been working with recently has reached a milestone. When I first started working with this team, most work would get to “development complete” but would linger in QA for multiple sprints or would not pass Product Owner approval so the work never got to “done.”
This particular team had experienced a lot of turnover and strife. I don’t know all of the details, but I can say that even just having people move in and out of a team can create enough instability to make holding to the team’s Definition of Done (DoD) difficult.
I think most people on the team understood what their Definition of Done was, even if it wasn’t something they said out loud or had written on the wall somewhere (but you can do such an activity in a retrospective, and I highly recommend it).
This team was writing automated tests, code reviewing check-ins, ensuring new work passed the test suite and built successfully. Everyone seemed to understand in concept that QA had to test their work and bless it, and ultimately the work needed Product Owner approval, but that’s not what was being measured.
The act of measuring something impacts your results. People naturally start to key on what you measure. This team was being measured on how much work they could get to development complete when I first started working with them. I knew instantly we needed to adjust it, but getting to shift the team towards the right measurements couldn’t happen overnight. I needed to better understand why this was happening, and then determine what to do to move towards what should be happening.
For any Agile team, the Definition of Done (DoD) is very important. It becomes how you know you’ve met your sprint goals and commitments and it gives a guideline for what the team considers to be completed work. It’s also how you measure the team and the work at a high level, for sprint planning purposes and at retrospective time to understand what trends the team experiences and why.
Every team’s DoD may differ, but generally, done does not mean “development complete.” It’s rare to be development complete and have fully releasable software on the first try. When development work is passed on for testing and quality checks, there are inevitably bugs/defects to fix or improvements to be made in the work. The Product Owner or Stakeholder may request tweaks or changes before approving the work, or alternately may accept this iteration’s pass at functionality while requesting tweaks in the next iteration.
Done usually means the work in question has met a level of quality the team has agreed on. Quality Assurance, Business Analysts and Product Owner(s)/Stakeholder(s) have approved the work. Acceptance criteria are met, and automated testing and/or checks (performance/load etc) have passed.
Generally, done for Agile teams should mean the work is releasable. Releasable, but not necessarily released. Meaning if you were to stop work after a given sprint, the product you have developed so far would be able to be pushed live and provide value. Some teams release small improvements often. Some teams group development work into releases containing multiple sprints worth of work. Regardless when you release, at sprint end, your work is ready for real world use.
Definition of Done can be skewed when there are things happening that [someone or the team] is trying to address- for example none of the work involved in getting to QA or Pending Product Owner Approval counts as completed (in terms of moving up the field). If you have enough work that lingers like that just shy of done that leaves a team demotivated because no one acknowledges their hard work. Or someone isn’t addressing the concerns coming from the Quality Analyst or Product Owner. Or someone isn’t protecting the process. This is a process problem AND a people problem. Sometimes it’s people not understanding the process. Sometimes it’s absence of process.
On teams where work is frequently not meeting the definition of done, Scrum Masters / Agile Coaches must be looking at why and working together with the team to find solutions.
Your very first step when encountering a team that isn’t performing the way you think it should, is to start looking at what’s going on with an open mind. Find out what they consider as their DoD – either formally, through a retrospective, or informally, through observation and casual discussion.
Observe. Take notes. If possible, talk to the team about what you are seeing. Inevitably, there are reasons why the team is where they are, and as in the Retrospective Prime Directive- that’s a result of everyone’s actions and inaction – they were doing the best they could at the time. No need to start blaming anyone. Just start addressing the problem in a constructive and adaptive way.
Ask questions to discover what may be preventing the team from getting to done.
Is the definition of done too difficult to meet? Too ambiguous?
Is the team spending enough time discussing the work and how to approach it in sprint planning?
Are stories fully defined or do various roles have to track down information during what should be development or testing time?
Are acceptance criteria too ambiguous, resulting in confusion over whether the work is complete, or whether the work passes the expectations of it.
Does the stakeholder/product owner understand incremental software development and accept working software (a few most valuable features that work well rather than pressing to get all of the features that barely function)?
Does the team have all of the skills and experience needed to complete work? Does the team have the right balance of skills?
Is there a workflow problem? Are the right people involved at the right agile ceremonies to keep work moving (especially review of defects and work failing testing as well as product demo and stakeholder approval).
Is the team overloading the sprint?
Where does the work tend to get stuck, and why?
Are team members remaining blocked without hope of help getting unblocked?
Does the team fully grasp the agile sprint lifecycle?
Is someone protecting the agile process?
This is just an example list. Remember, look at what’s going on with an open mind. After asking questions, start looking for solutions. It’s easy to complain, but it’s much harder to take problems and start finding solutions. It may help to use a Retrospective Method such as Known Issues, AFTER clarifying the Definition of Done (DoD), to collaborate as a team in finding ways to solve the problem for the team.
As an example, for one team I worked with not too long ago, we identified one main reason they didn’t reach DoD was the work was not fully defined. People doing the development had to spend time tracking down and clarifying with Stakeholders what was desired. People testing the work had to spend time walking through with the developers what the heck the software was supposed to be doing. Stakeholders didn’t understand that the work could be further improved in a future sprint and didn’t want to release before everything was perfect. All of these things left work just shy of done at the end of the sprint.
When problem solving how to address the DoD issue with this team, there was one root cause and 3 resulting problems to consider. The first thing they decided was they wanted to write more robust acceptance criteria. The second, that they would not begin work without clear acceptance criteria and they would consult these acceptance criteria in determining whether the work was complete. The third was that QA would use this criteria to evaluate the work. And lastly, they had to, over time, help their Stakeholders become more comfortable with the idea that the software was ever evolving and improving.
Every team is different but getting to done is the general goal of an agile team during the sprint. If your team is struggling with this, it’s time to examine why.