We are entering our 21st Sprint as a team tomorrow. 21 2-week Sprints that we all meet everyday at 11:00AM in the project room. We’re a bunch of dudes working on the back-end systems of our websites: caching, framework, performance etc. 5 PHP developers, 1 QA, 3 contractors, 2 SysAdmins, 2 DB developers, 1 Product Owner and a Scrum Master. That’s a whole 14 people, quite a few if you ask me…
Problem description
Since I joined the team, I have always found our Daily Scrum meetings a little, well, boring. I have tried shaking things up a little but without success so far. However recently I had the feeling people come to the room because they must and not because they want to, entering the room as if they were going to be beaten up to death. As a result the actual communication is of no value whatsoever to anyone. That annoys the heck out of me and I don’t seem to know what I could do…
Work environment analysis
Here is a sumary of the team’s work environment, this should give some clues already on the root cause.
* We have several projects running at the same time – team members are usually split between these projects.
* Some team members are not fully dedicated to the team (they are shared with other teams).
* The backlog is managed virtually, same with the various Burndowns.
* Lack of meeting rooms generally make it hard to split things up.
* Lots of pressure and high priority work going on: we can’t take the time to talk about things!
Generally I think some guys don’t want to go into details because they either dont want to hold off others or actually think that others don’t care about what they are doing because they are focusing on other things. 14 people…
All that ain’t fun!
What can I do, as a Scrum Master, to sort things out?
I have to be creative to sort these problems while staying within the framework provided by the company. Here is a list of things, in no particular order, which come to my mind as I write this piece:
* Organize the Daily Scrum in projects: start with project A, the with project B etc. The expected result is more coherence in the communication flow instead of jumping from one project to the other.
* Do 1 Scrum per project instead of 1 big mammoth Scrum – split the team into smaller groups while keeping the lead developer in all these.
* Have a physical product backlog and burndowns in the project room in addition to the virtual one (thanks Dan).
* Change the time and location of the Daily Scrum meeting.
* Do one session to remind everyone of the purpose of these daily meetings.
* Work closer with management and stress again the importance of not working on several projects simultaneously.
That’s all that comes to my mind for now… Next step is do a thorough cost/benefit analysis and work together with the team to find a solution that everybody is committed to.
What would you do to break the routine in Daily Scrum meetings and make things fun for everybody? Have you ever experienced a similar scenario?
Another day, another interesting discussion at work.
This time, I guess you got that, we talked about the data teams’ Sprint Burndowns should be based on. Should teams have their Sprint Burndown based on Story Points or on the sum of all remaining estimated tasks in hours?
To answer this question, it’s important to understand what the goal/role of a Sprint Burndown is. The Sprint Burndown is a tool that:
* shows the team’s progress of a running Sprint
* indicates the team’s performance during the Sprint
* should show whether a team is on a good way to meet its commitment or not
* shows where blocks in the development process may happen
Let’s do a small comparison of 2 approaches…
1. Sprint burndown based on story points
* doesn’t require the team to estimate tasks
* might be considered as less overhead by the team
* may not explicitely show bottlenecks in the development process
* may have the tendency to create burndowns like this
2. Sprint burndown based on time
* requires initial tasks to be created and estimated at Sprint start
* may put negative pressure on the team by introducing another level of estimations (and measuring against these estimates)
* may be considered as too much overhead by the team
* may show bottlenecks/blocks more explicitely than with Story Points
* will have the tendency to create burndowns like this
In summary, the purpose alone of the Burndown should already give a good idea in which direction to go. The team/company’s culture and how deep you are in your implementation of Scrum may be 2 other aspects to take into account when making that decision.
Now Jeff Sutherland recently wrote “The best teams I work with burn down story points. They only burn down when a story is done.” (Source) - Are you there yet?
I thought that was never going to happen, but I’m back online and with a new hosting provider this time. As I mentioned a while ago, I was quite dissatisfied with the service provided by 1blu.de and finally switched to all-inkl.com. They have an offer ending end of January: no setup fees and first month free… Hell yes!
The move is now done – although not as smoothly as I had hoped it would!
Let me know if you see anything acting weird on the site…
Happy reading! :)
Today the Product Owner of one of our teams popped by our desks to seek some advice on estimations. He seemed quite puzzled:
“For user story X, the team correctly identified that the work was at the database level. But when doing the actual estimation, they stepped back and left the expert database developer do the estimate. When the DB developer spoke out, I got some signals from the rest of the team that the estimation might have been inflated – still none of them said anything.”
I see 2 problems here:
1. Why isn’t the entire team estimating the user story *together*, even though it only involves DB work?
2. Why is the DB developer inflating the estimation?
In my experience, developers – as basic example – are not willing to estimate anything that isn’t related to their code, is a common pattern in teams starting Scrum. “How can I provide an accurate estimate in a field I am no expert in?” How have I solved it? Usually I try to insist on the fact that there are no *right* estimates and that the goal of all this is to generate a discussion that will help everyone to have a better idea of what the deliverables of a user story are. Estimates are not written in stone!
Inflating estimates is a little trickier… In this case, I know that the DB developer is actually part of another team which means he may already be overloaded with work from this other team. Inflating to avoid the burnout? Why not… How would I solve this? First thing that comes to my mind is to try to hire, which may not be solving the root cause. If that is not possible, I would suggest the team to inform themselves on the subject matter to become less dependent. Writing SQL queries is surely no rocket science when you can do .NET!
Questions to other Scrum or Agile practitioners: Did you encounter similar scenarios and how did you handle them? If not, how would you handle them if you did? Share your experience with us!
