One easy way is to use points to rank the projects. Points help you see the relative business value and ignite discussion about the rela- tive value of each project. When you rank with points, you’re separat- ing business value from funding. The number of points you assign to a project is a representation of its value to the organization, not the fund- ing you will provide. This separation of value from funding works in a similar way that separating project sizing from duration helps project staff estimates better.
Start with a large total number of points. You will assign a unique num- ber of points to each project, showing its relative value to other peo- ple. The larger the total number of points, the easier it is to see each project’s relative value. If you have up to eight projects, you might be able to use just 10,000 points. If you have more than eight projects, start with 100,000 points. If you have thirty or more projects, partition them in some way—by division or team or by internal or external—
RANKORDER THEPROJECTS IN THEPOR TFOLIOUSINGPOINTS 67
because it’s close to impossible for human beings to understand that many projects and their relative value at one time. If you can, partition the ranking to no more than ten at a time; five to seven projects at a time is best.
Now, assign a unique point number to each project. The number of points you assign to a project helps you see each project’s relative rank in relationship to all the other projects. Since I said not to rank alone, I’m assuming the “you” here is a group of you. Don’t expect everyone to be in complete agreement with everyone the first (or even second or third) time you try to assign points to any specific project. Each person will benefit from the discussion of how to decide how many points a project receives. Consider adding the issues discussed later in this chapter to the point discussion; see Section 5.4,Rank the Projects by Risk, on page 73; Section 5.5, Use Your Organization’s Context to Rank Projects, on page74; Section5.6, Who’s Waiting for Your Projects to Be Completed?, on page76; and Section5.7,Rank the Work by Your Products’ Position in the Marketplace, on page77.
If you have two projects that are critical to the success of your orga- nization, you might decide to assign one 5,001 points and the other 4,999 points. (Or, if you don’t mind points left over, you could assign one project 5,000 points and the other 4,999 points.) That would show everyone that no one needs to work on any other projects and that these two must be completed before considering work on any others.
The project with 5,000 points needs to be completed first. You would figure out how to create two teams to work on these projects simul- taneously. I’m not saying to create one team to work on both projects simultaneously; that’s multitasking. But I am saying that if these two projects are by far the most important work you can do for the organi- zation, then you would staff these two projects to the exclusion of all other projects and have the two teams work on them concurrently.
With just two projects, if you have only enough staff to work on one project at a time, you can even ask the project staff to work in one- week or two-week timeboxes, alternating on each project. If one project becomes more valuable, you can decide then to have the staff work on just that one project for more than one timebox, assuming you review the portfolio after every timebox. See Section 7.1, Decide When to Review the Portfolio, on page107for more information on how often to review the portfolio.
What if instead you have a situation like this?
RANKORDER THEPROJECTS IN THEPOR TFOLIOUSINGPOINTS 68
Total Points 10,000 550 Project 6
770 Project 5
Project 4 780
Project 3 1,000
Project 2 3,000
3,500 Project 1
Points Project
Here you have two projects that have a relatively higher priority and a whole bunch of other projects with low point values. You have more choices now. One great choice is to fully staff the projects worth 3,500 and 3,000 points. Now you see who you have available. If those people can start work on the 1,000-point project and make sufficient progress without interrupting anyone from the top two projects, great. Staff the 1,000-point project. But if they can’t make progress without needing help from the top two projects, either don’t start that third project, because it’s not that valuable, or ask those people to work on one of the top two projects. Or, ask them to work on the next project down on the list. When you have two clear winners in the ranking and a number of other much less valuable projects, do what you can to complete the first two ranked projectswithout distraction.
You might find that those “extra” people can work on the top two proj- ects in ways you might not have considered before, such as breaking the product backlog items for each project into smaller chunks so you can have more people working in small groups on small features. Or, maybe the “extra” people can pay off some technical debt somewhere or something else that does not require an interruption for the top two projects. Don’t overstaff one of your top two projects, and don’t under- staff any other project just to keep people busy.
Is there an ideal team size? Maybe. Schwaber inAgile Project Manage-
LEFTOVERPOINTSPROVIDEMETADATA 69
ment with Scrum[Sch04] says seven people give or take two is the right number. In The Mythical Man Month: Essays on Software Engineering [Bro95], Brooks discusses a ten-person team. Katzenbach, in The Wis- dom of Team: Creating the High-Performance Organization [KS99], says the number is “less than ten.” Weinberg1 says there is a factor of three with teams, and teams larger than nine break into groups by them- selves. My experience with teams is that teams smaller than five people may not have enough people to finish features, and teams larger than nine break apart into subgroups.
If you add more people to a project in the hopes of finishing it faster, you may well slow it down. Every time you add more people to a project, you increase the number of communication paths. Don’t move someone on to a project just to keep them busy. Optimize at the team level to ensure finished projects so you don’t create bottlenecks as inThe Goal[Gol04].
To see more about productivity, take a look at Section 10.10, Measure Capacity by Team, Not by Individual, on page158.
Keeping people “fully productive” if they can’t add value to the most valuable projects is not keeping them productive or adding value to the organization—it’s splintering the efforts of the people who are adding value. To see more about productivity, take a look at Section 10.10, Measure Capacity by Team, Not by Individual, on page158.
This is the whole point of going through the aggravation of relative rank- ing of all the projects in the portfolio. You know what’s worth your time to start. You know what’s not worth your time to interrupt. You know which projects you have to staff now and which ones can wait until later.