An agile technique that my students are almost always surprised to find as part of our agile training courses is Planning Poker, and its subsequent application in a real environment.
In summary, it is a team estimation technique that asks the people who are going to do the work to reach an agreement on the approximate effort necessary to complete the work. This agreement is based on a third party’s description of the desired result; in the game, this “third party” role is played by a person who does not participate in the estimation. All team members have a voice and a vote.
To play, each team member has a set of cards, usually with the numbers 0, 1, 2, 3, 5, 8, 13, 20, 40, 100 (a modified version of the Fibonacci Sequence). The number on the card represents the number of effort units, or Story Points. After a description of the desired results is provided (the what and the why), the team has a conversation about the work needed to perform that functionality (the how). Then, each team member chooses a number individually, and, in unison, they reveal their selection to each other. The number chosen corresponds to the number of estimated Story Points. If the results are similar, an agreement is easy to reach. If the results are more spread out, the differences are discussed and an estimate is made again.
Image source: http://store.mountaingoatsoftware.com/products/planning-poker-cards
How to apply Planning Poker is detailed in many other articles so beyond the explanation above, I’ll skip going into the details. I want to focus on the why, rather than the how.
This technique, popularized by Mike Cohn, is effective because it emphasizes the point of who should be making the estimates. As noted above, I would like to highlight the fact that the estimate is made by the people who are going to do the work. It is not an estimate made by a single expert, the “boss”, the Project Manager, or the commercial department.
Estimates will always have some kind of bias, but some scenarios will produce more bias than others, either voluntarily or unconsciously. Being aware of the biases and understanding how to break them is key to success.
The question is: What problems are we trying to solve when applying Planning Poker and Story Points?
Here are a few!
The Boss Effect
Because all team members give their estimate at the same time, nobody shares their estimate before anyone else. This prevents a natural tendency to be influenced by the opinion of the “boss”. The boss can literally be the boss, or they can be a dominant team member. Planning Poker allows for the presence of a hierarchical structure within a team to not affect the process or the result.
The Expert Effect
This is very similar to the previous case. If an expert tells us their estimate before anyone else gives their opinion, it will surely influence others’ decisions.
I know what you are thinking (because I get this a lot in class), why do we want an inexperienced estimate? Isn’t it better for the expert to do it directly? Maybe. Maybe not… teams must remember that group estimates will always be more beneficial. Discussions with the expert can allow for less knowledgeable people to gain insight that can be applied to future estimates and projects.
Also, an expert might not always know the details of a task – a person with less experience might have more specific business unit knowledge. In my experience, there is always a positive aspect to team reflection and discussion of work.
The Group Effect
Also similar to the two cases above. However, this influence is due to majority instead of a single individual. If each team member exposes their estimate one after another, it is very possible that someone might modify their answer. A very silly example, but… consider how many times you have changed our mind in a restaurant after hearing what others at the table ordered.
Time Spent Making Estimates
The Planning Poker process forces teams reach an agreement without delaying the decision, and only for the work they are going to do next.
By using Story Points, players are able to think more abstractly, instead of using the time-based units such as “days”, “hours/man” – which can overload and slow down a person’s thought process.
Once the team agrees on the effort that is equivalent to 1 Story Point, it makes estimates by affinity, triangulating and comparing the effort between them. The result is more accurate, and it’s faster to reach an agreement. It’s like in the casino; we don’t bet with dollars, we bet with chips. We know how much money a chip is worth, but betting with them instead of with bills allows us to think faster and not as much.
Precision vs. Accuracy
By using a set of pre-defined cards, there is a limit on the possible subset of values. This, in addition to speeding up the process since we do not argue for accuracy (we have to choose between 8 and 13, we cannot choose 9, 10, 11 or 12), helps us to be as accurate as possible.
I highly recommend using Planning Poker and Story Points to avoid these common problems.
Oh, by the way… I didn’t want to end without commenting that there are other approaches to solve estimates and predictability of delivery, such as the #NoEstimates movement, but that’s for another article that my colleague Victor wrote!