Cross-Portfolio WIP Limits
Sriram Narayan
Impact Intelligence | Product/Digital/Tech Performance | Author: Agile Org Design (Pearson)
A cross-portfolio WIP limit serves to cap the number of projects (or initiatives or epics) that may be in flight across all workstreams at any time. Capping helps make sure that we finish what we start. In the absence of such limits, you might have witnessed a situation of too many projects in progress for a long time and too few going through to completion. The rest are cancelled or put on hold to make room for newer projects in the next funding cycle. What a waste.
How do these limits work in practice? Let's take an example.
Example
Say we have five streams of work, each with a WIP limit of 20, serving three portfolios. A cross-portfolio limit of ten projects would mean that no more than ten projects shall be in progress at any time. The limit would hold even if the ten projects are unable to saturate the available capacity in the workstreams. In other words, cross-portfolio limit may kick in even when individual workstream limits haven’t. On the other hand, it is also possible that workstream limits are reached even with only a couple of projects in progress.
How do we arrive at a number for a cross-portfolio WIP limit?
Extreme Limits
What if we went with an upper limit is 20 x 5 = 100 for the situation mentioned above? We could, in theory, have a hundred projects crawling along in this situation with one story in progress at any given time. None of them would be completely stalled. Each workstream would have twenty stories in flight belonging to twenty different portfolio items. And that would be disastrous for getting anything done.
The other extreme is a limit of one. You can imagine the ruckus if you put all but one business stakeholder on hold. Besides, too low limit a limit will result in too much spare capacity and hurt throughput.
Realistic Limits
Realistically, you need to be able to accommodate at least one in-flight project per important business stakeholder or department or business unit. That doesn’t mean each of them will always have something going on. Business priorities might skew the distribution. But that’s one way to think of a realistic lower limit for cross-portfolio WIP.
To arrive at a realistic upper limit, we make use of the available data. Say the average project is 36 stories and the average workstream throughput per sprint is 12, 19, 16, 15, and 12 for the five workstreams respectively. Roughly speaking, that means the delivery organization has enough capacity to process two (74/36) projects worth of work per sprint.
What would a cross-portfolio limit of five mean in this situation? Starting from scratch, if you began work on five projects, they might all get done by the end of the third (36 x 5/74) sprint, assuming the folks upstream and downstream can keep up. That’s a delivery cycle time of six weeks per project assuming a two-week sprint. How does that compare with your current cycle times? If it is much better, you have room for a higher limit.
Start somewhere in the middle of this realistic range and observe the effect on project completion rates relative to start rates before you revise the limit. This might take a while.
领英推荐
Critical Thinking Section
If you have been following this series of posts closely, you should be able to deal with these questions. Are workstream limits and cross-portfolio limits independent of each other? If not, which one influences the other? How exactly?
Taking Stock
We are now five posts into this series on improving system performance by addressing system usage:
Cross-Portfolio limits (this post)
In the next post, we’ll see how to take this knowledge about "regulating system usage to improve system performance" to business stakeholders.
Until then, take care and prosper.
Sriram
Enterprise Architect || Friction Fixer || Business Analyst || Senior Product Owner
3 年One of the core problems of WIP limits is getting stuck with projects when there is an opportunity to do some thing that has just come up and has far more impact or ability to influence the entire organization. how do we handle it, apart from the common sense argument of put the WIP to hold and replace it with this. any out of box approach here? The first para in the article touched a raw nerve of the operations person that iam. Thank you for writing this and helping with one of the approaches.