How To Implement Scrum in 10 Easy Steps - Step #3: Sprint Planning (Requirements)
by Kelly Waters
If you've followed the first 2 steps in this series, you should now have your product backlog in order and have estimated its size using Fibonacci points.The next step - Step #3 - is to plan your Sprint.
Sprint Planning Workshop
Call a Sprint Planning meeting. Make sure the meeting is attended by the whole team. Include all roles. Business Analysts if you have them. Testers if you have them. ALL Developers on the Scrum team for the product. And very importantly the Product Owner.
The first thing you must do (in your first Sprint Planning meeting) is decide on your Sprint duration. This decision should be taken as a team.
Decide Your Sprint Duration
This is an important decision. Scrum suggests 30 days. It might be right. But this is one point that seems to be widely adapted by agile teams practicing Scrum.
The optimum Sprint duration depends on many factors. Something I read recently suggested that a development team's 'cycle time' is a direct reflection of the maturity of their processes. I think I would agree with that statement.
A team with immature processes will find the intensity of Scrum and the overhead of Sprint Planning, Testing, Deployment and Review quite onerous for a short Sprint cycle. Whereas teams with very mature processes (for example automated testing, automated deployment, and teams who've become very quick at Sprint Planning), a short cycle might be very comfortable.
I'd suggest the range is between 1 week and 1 month. 1 week is probably the shortest that will ever be practical, although if you really master agile practices, why not ship each new feature when it's ready? (if that's appropriate for the product). 1 month should certainly be the longest.
For fast-moving products or markets, such as web-based products - where there is central deployment and no rollout or user training - 1 month seems like a lifetime! Personally I like 2 week Sprints for fast moving products.
Mike Cohn is one of the key exponents of agile development. See here for Mike's article giving further advice about how to select the optimum iteration length.
Keep Sprint Duration Consistent
Whatever Sprint duration you choose to go for, my advice is to keep it consistent.
This, in fact, is more important than the length itself. Because it's this consistency that allows you to get into a rythm. It's this consistency that makes your process very repeatable. And therefore helps you to get into your stride as a team. And it's this consistency that allows you to start understanding how many Product Backlog points you can typically do in a Sprint.
Once you've decided, you can now set up a Sprint Planning Workshop as a recurring appointment before every Sprint.
Select Target Backlog for Sprint
Now you've decided on your Sprint duration. Next you must decide on the goal for the Sprint...
Looking at the top section of the Product Backlog, what would seem to be a reasonable goal to set for the Sprint? Can you express an objective that sums up the goal for the next Sprint, or at least pick a section of items/features from the top of the Product Backlog that the team thinks can be achieved in the Sprint duration?
Select your target backlog for the Sprint. Make this decision as a team.
Include a bit more than you think can be achieved. It's important to prepare more items during planning in case the team finishes early. These items can be clearly identified as stretch tasks and the Product Owner should not expect them to be completed. These are the things you will only do if the Sprint goes better than you expected.
In future Sprints, you will be able to use your Scrum team's previous Velocity to help with this decision. Velocity is the number of Product Backlog Points delivered in a Sprint. This tends to fluctuate wildly early on when adopting Scrum. But it will settle down as the team get into a rythm, and in future provide you with a reasonable norm to base your target backlog on.
Clarify Sprint Requirements
Take each item on the Product Backlog. It's important to go through them methodically, one item at a time...
Once you've decided, you can now set up a Sprint Planning Workshop as a recurring appointment before every Sprint.
Select Target Backlog for Sprint
Now you've decided on your Sprint duration. Next you must decide on the goal for the Sprint...
Looking at the top section of the Product Backlog, what would seem to be a reasonable goal to set for the Sprint? Can you express an objective that sums up the goal for the next Sprint, or at least pick a section of items/features from the top of the Product Backlog that the team thinks can be achieved in the Sprint duration?
Select your target backlog for the Sprint. Make this decision as a team.
Include a bit more than you think can be achieved. It's important to prepare more items during planning in case the team finishes early. These items can be clearly identified as stretch tasks and the Product Owner should not expect them to be completed. These are the things you will only do if the Sprint goes better than you expected.
In future Sprints, you will be able to use your Scrum team's previous Velocity to help with this decision. Velocity is the number of Product Backlog Points delivered in a Sprint. This tends to fluctuate wildly early on when adopting Scrum. But it will settle down as the team get into a rythm, and in future provide you with a reasonable norm to base your target backlog on.
Clarify Sprint Requirements
Take each item on the Product Backlog. It's important to go through them methodically, one item at a time...
The Product Owner presents each item and explains how he/she sees it working from a functional perspective.
The whole team discusses the item in detail. The whole team asks questions about the feature in order to establish what it should do and how it should work.
The outcomes of this discussion should be captured on a whiteboard or flipchart, or someone could write notes on a laptop as the discussion progresses. Interactive or printable whiteboards are ideal for this process.
You can use whatever form of writing requirements you want to. But the important principle in Scrum, and in any agile development methodology, is that you write requirements feature by feature, just before they are developed.
You can use whatever form of writing requirements you want to. But the important principle in Scrum, and in any agile development methodology, is that you write requirements feature by feature, just before they are developed.
Write requirements in a way that is lightweight and visual. Agile requirements should be barely sufficient. The fact the features will be developed and tested within the next few weeks, and by the team that were present, makes this possible.
Consider writing 'User Stories', a concept from XP (extreme programming). It's beyond the scope of this article to explain user stories in any detail. But the basic concept is to write features using this construct: As a [type of user], I want to [do whatever], so I can [achieve what goal].
Consider writing 'User Stories', a concept from XP (extreme programming). It's beyond the scope of this article to explain user stories in any detail. But the basic concept is to write features using this construct: As a [type of user], I want to [do whatever], so I can [achieve what goal].
The story can be backed up by a sketch of the UI, a wireframe or visuals. Annotate the sketch to describe the functionality. Backed it up with statements about how it will be confirmed (or tested). This will help to identify scenarios up front, before it's developed.
For more information about user stories, Mike Cohn has written a book, User Stories Applied.
Next...
Once you have clarified the requirements for all the Product Backlog items targeted for your Sprint, the next step is Step #4: Sprint Planning/estimate tasks...
See also:
Next...
Once you have clarified the requirements for all the Product Backlog items targeted for your Sprint, the next step is Step #4: Sprint Planning/estimate tasks...
See also:
How to implement Scrum in 10 easy steps:
- Step #1: Get your backlog in order!
- Step #2: How to estimate your product backlog
- Step #3: Sprint Planning/clarify requirements
- Step #4: Sprint Planning/estimate tasks
- Step #5: Create a collaborative workspace
- Step #6: Sprint!
- Step #7: Stand up and be counted!
- Step #8: Track progress with a daily burndown chart
- Step #3: Sprint Planning/clarify requirements
- Step #4: Sprint Planning/estimate tasks
- Step #5: Create a collaborative workspace
- Step #6: Sprint!
- Step #7: Stand up and be counted!
- Step #8: Track progress with a daily burndown chart
- Step #9: Finish when you said you would
- Step #10: Review, reflect, repeat...
- Step #10: Review, reflect, repeat...
'Implementing Scrum' PowerPoint Presentation
10 Key Principles of Agile Software Development
5 October 2007 06:10
It does seem that optimal sprint duration depends loosely on many factors. It's something that teams need to adapt in order to give themselves room to work. I like your word "comfortable".
However, I don't think the cycle time is actually very malleable. I think one factor sets the bounds: the time required for testing and stabilization.
I just posted a piece on stabilization, with this theory: The length of the stabilization period actually determines the length of your release cycle / iterations / sprints. You should allow about twice as much time for new development as stabilization. So, if your stabilization period is four days, you will want 8 days for development (12 days total), rounding up to 2 weeks.
8 October 2007 06:48
Hi! I’m trying implementing the Scrum method for my team. I’m a newbie in this, but it seems, that Scrum will be really effective for my team. Your tips are very helpful by the way. Thanks! I’ve also found this piece of software you can use for Scrum; it’s pretty agile and works for us so far.
9 October 2007 18:13
We look at duration from the perspective of the number of items that will be delivered. Not sure if that is a good approach, but it seems to work here.
The situation is altered if there are any mission critical items that need to go at short notice, which then goes outside the SPRINT.
3 January 2010 14:00
If we detail requirements one at a time, just in time, we may miss conflicts or duplicates in the requirements.
If we do all design for requirements one at a time, we are more likely to miss common solutions and duplicate code.
We certainly don't want to go waterfall, and document ourselves into paralysis, but some requirements specification and top level design should be done up front for iteration features.
In XP you actually do some requirements detailing up front, since TDD story tests are requirements.