Generally if a Scrum Master is off on vacation the role should temporarily fall to someone on the team to cover the standups, attempt to remove impediments, and generally keep things moving.
What doesn’t work that well is having another Scrum Master who has no relation to the project step in help. They don’t have any context to work from and at most they’re going to mechanically record status updates on tasks. Maybe they can try to remove some impediments, but again they lack context.
Given a choice leave it up to the team who fills in the Scrum Master role while they’re on vacation.
We’ve been holding code reviews for close to a year now and the process has evolved from the first days. We’ve seen many of the common benefits:
- Better code quality and readability.
- Catching some defects early.
- Mentoring opportunities around design.
We’re still a bit informal about it and probably reviewing less code than we should, but overall I’m pretty happy with the results.
That said I’m starting to weed myself out of more code reviews because I simply can’t keep up and I don’t need to be a bottleneck in the process. I really enjoy the break of a good 30-60 minutes of just reading and examining code and maybe running the corresponding tests, but I don’t have much time for it. Luckily we have enough senior developers now who have gotten used to the process and are able handle code reviews on their own projects.
As a manager code reviews can be important, but so are standup meetings, one-on-ones, staff meetings, removing impediments, planning out future project opportunities, and recruiting.
A few days ago a customer/product owner asked us for an estimate of the completion date for a project. None of the developers responded, so the Scrum Master is pressing for an estimate.
The state of our incomplete information is thus:
- There is an incomplete product backlog, it contains almost every item for a first release with estimates in ideal days. It also contains all the backlog items for the second release, but with no extra information and no estimates.
- At least 16 reports are just completely missing from the product backlog altogether.
- We have use cases for almost everything in the first release, but not even high level use cases for the second release items.
The answer was easy we looked at the product backlog and came up with about 150 days for the first release backlog items. Our current velocity is about 60 ideal days per Sprint. So the estimate would be 2.5 more Sprints as of today and assuming everything in the Product Backlog still needs to be completed.
The Scrum Master thanked us for that effort, but then asked if we could just get a ‘gut check’ estimate on both releases. Our answer so far is still no, not enough information. Until we do an estimating effort on the second release items and get some information on the reports we can’t honestly give any sort of ‘gut’ estimate.
The plan tomorrow is to go through the second release product backlog items and estimate them one by one, probably just by asking questions of the product owner. This may take longer than the 1 hour planned for, but the team isn’t going to get an estimate any other way. ‘Gut’ estimates are just a slide back into the scary nether-world of Waterfall.
I took a personal day this past Friday after realizing I was just burnt, fried, and generally stressed and not that useful. I pondered whether I should or not for far too long. The reason for the difficult decision–I didn’t want to give up control of several projects.
I’d been planning to take a personal day for the past several weeks. Our fiscal year ends in September so the last few weeks tend to be insanely busy despite any amount of preparation. I ended up coming in on at least one weekend day for 3 weeks in a row on top of several really late nights during the week. Every time I thought things would calm down enough to safely take a personal day, some new project emergency came up.
So this Friday I finally just went ahead and took the day. Friday included a planning meeting on a new Proof of Concept (POC) project for the second two week Sprint. I had planned to lead the meeting, but I finally decided I really needed to delegate it to a PM and one of the developers on the project.
Delegation is still something I can tend to struggle with especially when it comes to things like running Agile projects and acting as the Scrum Master. I’d told the PM on the project for the past two weeks that I’d get her up to speed on how Scrum worked, but I completely failed at making any time to do so.
I have at least one developer who was willing to learn the Scrum Master role as well and had experience in at least 7 prior planning meetings on another project. He was perfectly capable of selecting the backlog for the Sprint, establishing a Sprint goal, and helping to brainstorm out all the tasks. Helping to guide the PM in this case wouldn’t be an issue.
The work required to initiate everything was two simple emails. One to let everyone know I wouldn’t be in for the day and a second to bullet point the high level agenda for the planning meeting that they’d have to arrange and facilitate.
So Friday they pulled off the meeting on short notice and I trust it went fine. Once delegated you have to let go and let your people run things. They may do things differently, they may make some mistakes, but that’s the whole point of learning. And in my case learning that I need to delegate more, not just when I take the day off.
I got in a nice cozy nap with the two cats and rested up over the weekend. Monday actually looks pretty inviting.
I budgeted this year for dual 20″ monitors or one 24″ monitor for each developer. With prices dropping now maybe we can go to a single 30″. Keeping your IDE, a browser, a terminal window, and an email client certainly helps with productivity.