My team built an XP practice before introducing Scrum. The desire for change arose within the developers and was about sharing, becoming better at our craft and making our productivity and quality more consistent.

Scrum became necessary when our biggest problems were no longer within the team but in how we took on work from the business.

I started thinking about that at the last XTC-NYC when Mike Roberts wondered aloud if Scrum software development can work when a team doesn’t practice at least some aspects of XP. In a similar vein, Scott Bellware has a post about XP deserving more credit for Scrum’s success.

At XP-Spin, Jeff Sutherland said the success you build on is delivering working code at the end of each iteration. To even begin you need work described in achievable stories with acceptance criteria and daily builds.

Mike and Scott are definitely right when I look at my team.

Our present challenge is aligning our development with a clear and achievable business objective. Scrum is the tool for that. One of the things about Scrum is that you can use it for almost any kind of work requiring cross-functional teams.

But that challenge only exists because we were highly productive at creating quality software. The team’s XP practice with its low formality and high discipline gets the credit for that.

Short link: http://jkat.me/homOi8

Tag(s): agile, Jeff Sutherland, Mike Roberts, quality software, Scott Bellware, scrum, software development, xp

Comments: none

We have seven developers growing to eight. The team is funded from two P&L’s so from the outside we’re considered two teams. We work on two projects at a time, one from each P&L.

We used to run one backlog with developers pulling stories from either project. In the sweet spot where both projects ran smoothly this was great. But that proved short-lived. Churn in one project slowed the other and context switching took a toll on the team.

So we split the work into two backlogs. Everyone sits in the same room but developers are “assigned” to one project. As a side note, I actually planned on having two rooms but the team said “no”.

This change has created more focus. Developers know what they’re working on, churn in one project has less effect on the other. The Scrum is easier to run, plan and track and progress appears more regular.

The change is a qualified success. Four may be too small for sustainable pairing. If anyone goes on vacation or is sick, the pairing practice starts to break altogether.

So to retain the focus of two small teams but gain the benefits one slightly larger one, we’ve loosened the assignments. Developers are asked to work on a specific project but they can switch if they feel the hate. The leads can switch people as well.

Management is often about embracing contradiction. We need to find our way to both variety and focus. It’s all a work in progress.

Short link: http://jkat.me/gxMm7d

Tag(s): scrum

Comments: 1

I was recently asked the following question:

“I’ve been dropped into a situation best described as “story card hell.” How do I reduce complexity of the software project planning without losing features?”

The project was to replace an existing system. The behavior was largely known. As a result, the initial planning generated enough cards to feel like, “way too much detail up front.”

My answer:

Don’t lose the work you’ve done but don’t place more value in it than it has.

Step back to think of the application as a phased set of releases. Talk with the team and product owner about a way of delivering the system in meaningful pieces that customers could (and hopefully will) use. Get creative with defining those steps as long as they lead you down a path towards your ultimate goal.

Once you have that road map. Do an exercise with the team to group your existing stories into those releases. There will be a natural tendency to add and change stories. That’s fine but don’t get too distracted by it. The main point is to get a sense of the relative size of the releases and move the bulk of the stories out of your immediate planning concern.

Now focus on the first release. Spend more time on those stories but still not with the scrutiny of an iteration plan. Make/allow the team to live with ambiguity. Set up an agreed upon rule of thumb for how much bigger a “theme” is from a real “story” and let the team know it’s okay to discuss these things at the “theme” level.

Then prioritize your release backlog. Have the team chunk the remaining vague stories/themes into future iterations for that release. This is your release plan. Expect it to change. Again, you’re getting a rough sense of how many iterations in the first release and moving later stories out of your immediate concern.

Now you can focus in detail on the stories for the next one or two iterations without getting drowned in details. Don’t be surprised if future stories become irrelevant or drastically change as you get to the iteration within which they fall. It’s okay.

My answer was based on what we’ve tried to do at Oxygen with coaching on release planning from Hubert Smits.

Here’s the full thread off LinkedIn.

Short link: http://jkat.me/fHVQT2

Tag(s): Hubert Smits, management, scrum, software project, team and product owner

Comments: none

A metaphor for the Scrum Master is a vigilant sheepdog protecting their flock.

At the Fall Scrum Gathering, I met practitioners facing different challenges in their agile practice.

Some faced profound impediments that their organizations were unable or unwilling to address. The effect on the project and team was dire and the Scrum Master had exhausted all avenues to raise alarm.

It’s human nature, unfortunately, to associate an unpleasant message with the messenger. A vocal Scrum Master can be seen as the problem.

In those fraught circumstances a Scrum Master has to balance the interests of the team, the company and themselves. Can the project deliver in spite of the obstacles? Should the Scrum Master accept the dysfunction or not? At what cost?

As Ken Schwaber says in Agile Project Management with Scrum, “A dead sheepdog is a useless sheepdog.” Still, a useless sheepdog is also a useless sheepdog.

As JP Boodhoo says, “develop with passion.” As my friend Luke Melia says, “live with passion.”

Short link: http://jkat.me/hXVzi6

Tag(s): agile, Ken Schwaber, Luke Melia, messenger, scrum

Comments: 3

I’m presenting a paper, Great Scrums Need Great Product Owners, at the Hawaii International Conference on System Sciences.

The presentation is part of the Agile mini-track co-chaired by Jeff Sutherland and Hubert Smits. I co-authored the paper with the senior member of my product team, Ilio Krumins-Beens.

We survey literature to describe pragmatic and collegial relationships which satisfy the definition of collaboration and honor roles while barely tapping or actually working against the potential of a project and its participants. These limited forms of collaboration are common not just in our industry but within agile projects.

A common interpretation of the call for a single Product Owner is that the development team itself should play little part in shaping the vision and value priorities of a product backlog, focusing instead on efficient delivery of those priorities.

One of our major sources are the What’s Worth Fighting for series by Andy Hargreaves and Michael Fullan on collaboration in education.

The disempowering climate faced by teachers is a direct parallel to that faced by software developers disengaged from the business vision and value priorities of their work.

“… [W]e have collectively 45 years of teaching experience and nobody has ever asked us our opinion about anything where it would actually be put into action. And yet I’ve got to have more experience with junior [children] than a lot of the people who are telling me what I should be doing with them. And I think that is very frustrating… I think I could help bring a lot to it and nobody ever asks, no one ever asks what we think. They just go ahead and proclaim and we have to follow.”

Its tragic how talented, experienced people are not allowed to bring their full selves to their work.

We contrast this with unbounded collaboration between a product owner and team. In unbounded collaboration an individual’s contribution is not constrained by their role or status.

It is a collaboration built upon high-performance, mutual respect and deep trust. The product owner walks a tight rope, engaging the team in an evolving product and business plan while guiding the project toward her vision and high-level goals. The team is passionate about the product they are building and feel personally accountable to the product’s success.

We argue that this kind of collaboration is at the heart of agile values and is a characteristic of the knowledge-creating companies upon which lean principles are based.

I’ll write more about unbounded collaboration, it’s opposites, practices we’ve used and what we’ve learned.

Short link: http://jkat.me/i1xpkz

Tag(s): Andy Hargreaves, Hubert Smits, Jeff Sutherland, Michael Fullan, product owner, scrum, senior member, single Product Owner, software developers

Comments: 1