The Case for Poker Cards in Sprint Planning

18 February 2013

Dele Oluwole
Qual IT

During sprint planning in Scrum, the team can use poker cards as a unit of measure for estimating the overall size of a user story, feature, or even associated piece of work.This tool helps the team assign points, in a relaxed manner, to stories or tasks. Having points assigned to every story will, in turn, help the team estimate how long it will take to complete each task, what story can be triangulated, and what goes into the next iteration. This tool is a consensus-based technique for estimating and coming to agreement on the projected workload in a sprint backlog.

A study by K. Molokken-Ostvold and N.C. Haugen ("Combining Estimates with Planning Poker: An Empirical Study," Eighteenth Australian Software Engineering Conference, April 2007) found that estimates that teams reached when using the planning poker process were more accurate than the more optimistic estimates they obtained by combining individuals' estimates.

For me, the poker session is the best part of sprint planning. I try to objectively balance its sense of fun with the responsibility I have as a member of a cross-functional team in a dynamic environment. Sprint planning without "poking" your card is like going on a cruise without standing on the ship deck to view the beauty around you.

Can there really be complete or successful sprint planning without collaborative estimation? If each team member estimates individually, and the time estimated for every story is put against their names with the aim of working as a member of a team to meeting individual estimated targets, would that be working cohesively as a team in the spirit of Scrum/Agile?

When I help organizations reengineer and drive down Agile principles and procedures, I try to remain objective, calculated, and timely with my input, especially when the team includes verbal Agile skeptics who could drown out the implementation of these principles. Recently, in a supposedly Agile environment, I stuck out my neck to make it clear that the team should use poker cards and assign points to every story cohesively in order to determine velocity. Surprisingly, some ScrumMasters still have Waterfall habits in the Agile environment, and such ScrumMasters drive down the Agile practices in haphazard manner.

If a Scrum team faces the problem of "chickens" (those with ancillary roles) dominating the "pigs" (those with core roles) in sprint planning, the ScrumMaster could defer the estimating part of the planning till the end, when they can exclusively decide the story points. Only core Scrum roles that are involved in the actual work should assign points to stories. The ScrumMaster must make it a prerogative to make this happen.

I believe that poker is an integral part of sprint planning and can help solve this and many other problems. According to Mike Cohn ("Agile Estimating and Planning"), the reasons why poker is beneficial to a Scrum team are:

  • It combines individual estimates through group discussion, leading to overall estimates.
  • The emphasis is relative rather than absolute.
  • Estimates are constrained to a set of values so the team doesn't waste time in meaningless arguments.
  • Everyone's opinion is heard.
  • It's quick and fun.

Remember, the velocity of a team is the number of story points assigned to stories that are finished over a specific period of time, mostly one to four weeks. Let's say, for example, that a team completed 10 stories worth 2 points each during a 2-week period. The velocity is 20 story points every two weeks. The velocity is a good measure of the overall throughput of the team.

Knowing your velocity helps a lot with planning. I therefore cannot imagine sprint planning without real velocity to use as a benchmark. And to determine velocity, the team will need to decide the size of a task, influenced by how hard it is to complete the task and what it involves.

It's only after assigning points to a story that the story can then be compared to a similar one during planning: "This story is like that story."

It's also open to team members to triangulate, or estimate by comparing a story to multiple other stories or a group of like-sized stories on the table or whiteboard.

These benefits and others will be missed when a team plans for a sprint without proper estimation produced by using poker cards or another method of determining velocity and assigning points to stories consensually.

Article Rating

Current rating: 4.5 (2 ratings)

Comments

ABHISHEK KUMAR SRIVASTAVA, CSM, 2/18/2013 4:55:41 AM
Good overview on planning poker Dele. One interesting thing I observed as scrum master during planning poker, "how newly joined team members" or "junior team members" get influenced by "senior team members". If I ask for size of a user story, new / junior team member may say size 2, then they look at senior team member card with size 5 or 8, and they will also change their estimate to 5 or 8 too. It's a big challenge; sometimes new team member has a smarter way of doing things.
Dele Oluwole, CSM, 2/18/2013 7:41:01 PM
Thanks ABHISHEK.

The rule is that cards are placed face down and flipped over at the same time to avoid the kind of influence you have mentioned. As the ScrumMaster please reiterate the need for independence when triangulating a task..
The rationale behind the use of Planning Poker is to forestall the influence of the other participants, especially the strong team members.

Planning Poker is expected to force people to think independently and put forward their numbers or cards simultaneously. ScrunMasters can only achieve this by advising that all participants show their card at the same time.

I wish you all the best in your planning.
Nathan Kettles, CSPO, 2/20/2013 1:23:37 PM
Great post. I believe planning poker is a incredibly useful tool for collaborative estimating, and we have used it at a few teams I have been involved in.

You must Login or Signup to comment.