Take all you want, but eat all you take. – Jay Packlick
I learned about the Agile Buffet Rule from the ever loquacious @jpacklick. This is not to be confused with the Warren Buffett Rule or even the Jimmy Buffett Rule. The concept, as it applies to Agile, is simple: Take as many stories as you want from the backlog, but make sure you have the available capacity to complete the work you take. I’ve noticed a trend at some companies where teams are not completing all of the work they have committed to, but calling the sprint successful. Often they are close to completing all their work; however I was taught to finish everything on my plate. Maybe it was the way I was taught Scrum in the old days, but I still feel if you don’t complete all the work you commit to, your sprint was not successful. It is very empowering for a team to be able to use velocity and capacity to plan their work, and have the costumer’s confidence that you will be able to finish what was committed. For Agile to be successful, you need to to win the trust of the business. You do that by delivering what you promise. It should be uncommon when you don’t finish all the work you signed up for. To be clear, I don’t have any problem with teams adding stretch goals to the sprint, so long as everyone understands they are stretch goals, and not part of the sprint commitment. You have a ten pound sack of potatoes and that’s all that will fit. Don’t try to put fifteen pounds in a ten pound sack. When I’ve observed teams committing to more work than they have in the past, the common question I ask is: what is different about this sprint versus the last one? In other words, what has changed since the last sprint that enables you to commit to a higher velocity?
So that begs the question: how do we ensure that we complete all the work we commit too? We can break the answer into two parts: Planning and Executing the Sprint.
The obvious answers are to make sure you are not going over your velocity or capacity thresholds (velocity in points and capacity in ideal hours). Your velocity is an average of your points over time (sprints), which should be relatively stable. So do not take more story points than your velocity. Likewise, if you are doing your planning via WIP, Cycle Time, etc. the same rule applies: Don’t take more than your WIP allows, unless something has changed.
Your capacity takes into account the number of hours available for this sprint. So make sure you are accounting for any changes in team personnel, holidays, vacations, etc. One way to achieve this is to create a capacity sheet which takes into account a simple formula for each team member: number days in the sprint x number of ideal hours per days x % dedicated = available capacity. Here’s a quick explanation of the variables:
- Number of days in the sprint: This is the total number of days, less any holidays, training, conferences, vacation days, etc.
- Number of ideal hours per day: This is the amount of hands-to-keyboard-time you actually do on project work each day. It could be based on an 8 hour work day, less meetings, breaks, lunch, phone calls, responding to emails, etc. I’ve seen teams use as little as 2 and as high as 6. Never more than 8 unless they plan to work 12 hours days. To give you some relevant data, from a study we did at Nokia, the average team only spends a 1/3 of their day on actual project work.
- % dedicated: It’s unfortunate, but many organizations continue to loan team members across teams. As a result, you may have team members dedicated less than 100% to the team. Thus, you must adjust their capacity accordingly. I don’t condone this, but it is reality. For those that are interested in why this is a bad idea, you can find it here in Kenny Rubin’s most excellent presentation on Economically Sensible Scrum.
The end result is the total number of hours, each team member has available to commit to the sprint. During Sprint planning, when tasks are created and assigned, the Scrum Master (or other team member) can reduce the number of hours available, per team member, each time a task is assigned. You are free to sign up for tasks until your total reaches 0. Here’s and example of a capacity worksheet I typically use:
Sprint Execution (aka How to Finish what’s on your plate)
On what he thought of his team’s execution: “I’m in favor of it.” – The late Tampa Bay Buccaneer’s Football Coach John McKay
Using the information radiators in Scrum is paramount. Use of the burndown chart and other information radiators (Cumulative Flow Diagrams, Cycle Time, WIP, etc.) to make timely decisions is critical. If it becomes clear that you are not going to be able to complete everything on your sprint backlog, when should notify your customer/product owner? The answer is immediately. The more time the PO has, the better the options become. Don’t make the mistake of waiting, thinking things will improve, when the data is telling you otherwise. As pilots, we are taught to trust our instruments and not our gut. Why? Because wishing the situation was different won’t help. The data is telling you that failure is on the horizon, it is better to admit it and come clean. Doing anything else erodes trust. Be decisive and be proactive. What strategies can we employ to make sure we finish the sprint? Here are some suggestions (in no specific order):
- Remove impediments – This strategy should be obvious, and hopefully, is being done on a daily basis.
- Swarm – Can we gang up on unfinished work?
- Add people/experts – perhaps we are struggling with a difficult story and the expertise exists outside of the team. With this strategy, you are essentially trying to increase your velocity.
- Move the stories to different owners – It could be that the person working on the story, isn’t the best one suited for the work.
- Work overtime – Not ideal, but can work as long as it doesn’t happen very often.
- Move into a team/war room – Sometime getting together in a crisis can help bring ideas out faster.
- Out of ideas? – Ask for help. This is not the time to be shy. Ask the team, ask your manager, ask your coach, the Internet, etc.
The key of course, to making use of these (or other) strategies, is early detection. None of these will work if you procrastinate.