Should We Have Longer Sprints?

 

by Christian Bisson

I’ve recently been part of a discussion concerning changing the length of sprints from two weeks to three weeks, and the product increments (“PI” from SAFe) from 10 weeks to 12 weeks. Hearing the arguments throughout the meeting made me realize how the impact the sprints have on teams is greatly underestimated. Also, it’s important to note that in this case, the sprint length is aligned for all teams—meaning all teams need to change.

In the discussion I was in, the arguments for having longer sprints were that it would reduce the number of meetings (therefore deliver more value), and that we would have better sprint reviews. Let’s review those arguments, and other factors to consider

The Number of Meetings

Assuming we are only referring to “agile” meetings, it’s true that the events (ceremonies) will occur once every three weeks instead of two weeks. However, aside from the daily scrums, the length of each of those meetings is expected to be extended accordingly. 

For example, a good rule of thumb for sprint planning length is about one hour per week in the sprint, so a two-week sprint would have a two-hour sprint planning meeting, and a three-week sprint would have a three-hour meeting—making it an average of one hour per week, and thus not really saving time. The same goes for the review and the retrospective.

The daily is the exception; that would remain at a maximum of 15 minutes every day, so no gain there either.

Better Sprint Reviews

In theory, since sprints would be longer, teams should deliver more within them (more on that under “Predictability” below)—and that will allow teams to present more accomplished work.

Depending on your circumstances, one could even argue that stakeholders would need less travel time to attend the review since it’s once per three weeks instead of two (although these days, even that argument has lost its value!).

But let’s look at the other side of the medal. The review is a key event to gather feedback from stakeholders and obtain precious information to move forward. That now happens less often, and could risk gaps in communication. In some cases, releasing an increment of work is not possible without having approvals within the review, meaning that value could be delivered slower.

So for this argument, I would caution analyzing your circumstances properly before deciding it’s a good idea to change the length of the sprints.

Predictability

This argument was the “con” discussed in our meeting, which I felt was underestimated. For organizations and teams, having good predictability means delivering all (or almost all) of what we forecast to deliver within a specific timeframe. This is key for planning product roadmaps, aligning dependencies, forecasting budgets, etc.

Predictability in our world is a challenge though; as we know, things change, and what we thought would bring value is only confirmed by empirical information that allows us to adapt as needed. This means that the longer the forecast is needed, the less chance we have to be accurate.

Don’t get me wrong—teams that have gained a certain maturity are expected to get to a point where they can accurately forecast the work they’ll accomplish within their sprint, and moving to a longer sprint would be viable for them. 

But the maturity of the team may vary, and in our case, we were talking about brand new teams that had a decent amount of carry-over from sprint to sprint already, so it was a no-brainer that they would struggle with longer sprints.

Scope Stability

Another element to take into account when selecting the length of sprints: Are the circumstances related to the scope stable enough to be able to focus on a sprint that doesn't (or barely) changes? The longer the sprint, the longer a potential “change” would have to wait—and if it can’t, then the sprint’s scope risks being changed.

Estimating and Splitting Stories

In order to properly plan and deliver sprints, it’s a good practice to split stories so that they fit within a sprint. Therefore, the smaller the sprint, the more it’s important to split stories—and that’s a habit teams develop as they grow.

On the flip side, if a team has not yet developed a good habit of splitting stories, a larger sprint will give it the impression that “it can be delivered within a sprint.” This results in a higher risk of having a backlog of large items that the teams struggle to deliver.

Conclusion

The length of a sprint largely impacts the ability for teams to forecast and deliver accurately. It’s not a decision that should be taken lightly. It’s also important to involve team members in the discussions.

What factors do you consider when deciding the length of your sprint? Any lessons learned to share?