Utah Agile Lehi June Meetup
Photo credit: BIG-D Construction
About the Utah Agile Meetup
If you’re looking for a great forum to discuss Agile methods with practitioners and coaches, check out the Utah Agile Meetup. This group is a great mixer for people with all different levels of experience. Here are some questions and answers from our discussion this month:
How to manage unplanned work
“Our team faces a constant barrage of unplanned work from new feature requests to production maintenance. It gets in the way of our planned sprints. What should we do?*
- Start tracking time spent on unplanned work and reporting it to engineering management.
- Some problems can be solved through automation. Others can be passed off to a support team.
- Split off a maintenance team to insulate resources assigned to planned work.
- Assigned a certain number of placeholder story points for maintenance, being mindful of avoiding too much slack.
- Create classes of service: assigning levels of urgency (positive or negative).
- Remember that urgent tickets don’t fit the nature of SCRUM. Consider moving to Kanban.
- Have a designated firefighter. They cannot have stories on the backlog assigned to them. Rotate this assignment so one person isn’t always in a high-stress situation.
- Defer newly found work to the backlog. Address it with the business and assign to a later sprint, rather than taking it on now. Assess value in what you’re trying to do for the customer.
How to coordinate releases across teams working on a monolith application
“We have about 30 devs on 5 teams managed by 3 scrum masters. Everyone is working on the same, monolith application. Our pipeline is frequently jammed. One team’s testing goes long and throws off the next team’s schedule. Jobs take half a day to usher out one team and make room for the next. We do a manual regression every week. It’s very costly. How can we coordinate our releases better?”
- Automate more of your testing. Absence of test automation causes longer testing times.
- Spread QA across the sprint, not just at the end.
- If the team can do all the work it takes to deliver a story all the way to “done,” make sure all that work fits into a single sprint.
- Have a regular release cadence. In our org, we synchronized sprint cycles: everyone starts and ends at the same time.
- SirsiDynix stopped development for 16 weeks to build automated unit tests, acceptance tests, etc. They got buy-in from the executive team to dedicate the time to reap longer-term pay-offs.
- Focus your testing efforts on the code where there were more developers working and more changes made.
- Roll out to a subset of users first.
- Build feature toggles to turn something off, if needed.
How to reward teams vs. individuals
“How can we reward teams effectively? Is it always better than rewarding individuals?”
- Make sure the performance of the team is measured, not the performance of individuals. Otherwise, you can turn individuals against each other.
- Focus on outcome over output.
- Find ways to reward individuals for team performance.
- Be mindful of the team members’ satisfaction with their current working conditions.
- Introduce peer-to-peer rewards rather than top-down rewards.
- Review the Drexler/Sibbet Team Performance™ Model. See infographic.
- When one team depends on another, find ways to recognize both teams.
- Nominate others for rewards.
What are the biggest pain points or organizations with mature SCRUM practices?
“If an organization already has mature agile practices in their technology delivery team, what issues may arise? What problems might they be trying to solve?”
- Measuring satisfaction and collaboration across the business
- Keeping the business ahead of the technology team’s output in a push model. See the Lucile Ball Chocolate Factory. Focus on a pull model.
- When agile teams get good, they start to pull others forward, causing friction. Pay attention to the points of friction throughout the organization.
- When agile shops are prioritizing with the business and teams are in sync, everybody in the organization is working on the most important thing.
- When executives or middle leadership lags behind, teach them SCRUM or whichever method you are using.
- Most agile consulting engagements arise when the client is rolling out Agile or fighting the threat of disruption from a competitor. “We’re going to start” or “someone is eating our lunch.” They know they need to change.
- The CTO is usually the advocate that has to pitch improvements to the rest of the execs. Once those other parts of the organization “get it,” if they can foot the bill, they’re putting their money where their mouth is.
- Review Value Stream Mapping.
How to Improve Predictability
“My organization is about to embark on a major 2-year development effort. How can we know if we’re on track along the way?”
- Start with end date and work backwards. Break your scope down into smaller release plans.
- Review DSDM – Dynamic Systems Development Method. How many features do you need to hit your first release?
- Review The Morning Game, a great exercise in breaking down and prioritizing scope.
- Define intermediate milestones and track progress against the earlier milestones to measure short-term progress.
- Measure velocity from results of earliest sprints.
- Consider how compensation can impact predictability. If a client pays in installments, it can introduce a lot of scope creep for the vendor.
- Predictability depends on consistency.
- Review User Story Mapping by Jeff Patton.
- Review Story Point Estimating by Mike Cohn.
- Focus on sharing the business vision to get team members to raise problems as they occur.
- Front load the most risky things to have more time to reduce risk.
Is JIRA the best tool for agile management?
“Based on your experience, would you recommend JIRA? What are the pros and cons of JIRA compared with other major agile tools? ”
- Consider geographic distribution of the team.
- JIRA is so configurable with so many add-ons, that it’s not the same from one company to the next.
- The three major players in the agile tools space are: Version 1, Rally (now CA Technologies), and JIRA.
- Start prioritizing what your biggest needs are.
- Get input from scrum masters and administrators.
- Have vendors demo how they can solve your problems.
- Product owners managing their tickets through CardBoard and then transferring to JIRA.
- Set up flexible administration processes to avoid a remote admin from being a bottleneck. Are you allowed to make your own changes?
Stay in Touch
Subscribe to have our latest research and videos delivered to you as soon as they are published.