My first response when I heard about Planning Poker® from Mike Cohn in a private training class I attended in 2003 was, “This guy has got to be kidding!”
First, Mountain Goat Software is a wild and unusual name for a company that has grown on me over the years. Pair this with a weightlifter named Mike, who could snap me like a twig, telling me I need to use poker cards to estimate my work was just a little out of place. I remember thinking, “Playing cards with particular numbers on them to hide our estimates from others? There has to be a better way.”
Well once the class was over, I was open to trying the approach. After the first round with a team of new developers to our team, I was hooked. I was typically faster at developing software than the other developers on my team and I was influencing their decisions because they were new to our team. This influence led to bad estimates and in turn more frustration when they missed the estimates and I had to help complete the work after hours.
Using the Planning Poker® approach to team-based estimating has been a life saver for the agile projects I have been a part of over the years. It has survived the test of time on software teams and it fits quite nicely with content marketing teams.
Estimated Work Helps Steer Decisions
I always state this in my posts about estimating so let’s get it out of the way, estimating is hard! We need a number describing the size to make decisions. Without our best effort in sizing our content, we are running the race with our legs tied together and wondering why the leaders keep getting farther away.
Have you ever wished you had another Content Developer on your team? The total scope of our backlog will give stakeholders the ability to see what is possible when we produce content and resource planning. With your team’s estimate of the Content Backlog, attached with data that shows your previous estimates and your ability to deliver on time, you will be able to demonstrate the need for a new resource on your team. Which would you rather lead with in that meeting: we need another resource because we are missing deadlines or we need another resource to be able to deliver more value?
A Sequence Will Help with the Inaccuracy
It is just about hurricane season again so it is a great time to use the analogy of a forecast. The further we are from the shore in an estimate, the wider the trajectory will be. Think about those storms on track to hit Texas or Florida. That is a pretty large gap of uncertainty.
In content, we don’t typically have the luxury of giving such a range, but we need the ability to have one. The solution is to embed the range into the number we use as qualified estimates. This paired with having several estimates will lead to a better overall satisfaction with our estimates because some will be high and others will be low. Frequent review of the estimates, the closer we get to the work will help get a better guess and help us see when and where we are going to land on shore.
For the Agile estimating process in Planning Poker®, the sequence of 0, .5, 1, 2, 3, 5, 8, 13, 20, 40, 100, Question Mark and Infinity is used. Only these numbers can be used. If you see a content item with an 11 on it for the estimate, you are doing it wrong.
The reason we use this particular set of numbers, is our inability to see clearly from a distance. When we are discussing whether an estimate is one day or two, we can clearly draw some assumptions to how long the tasks that make up the item will take because we don’t have very many. When we get up to a larger item, like something we believe is two weeks, can you clearly state this difference between a 10, 11, or 12? Since we have this discrepancy, it is easier to decide is it bigger than 8 and less than 13? If it is, we go with 13.
The Question Mark and Infinity are in the deck for the when estimates cannot be formulated with given scope or information. You select a Question Mark card if there is not enough detail for you to make an accurate estimate or you just don’t know enough about creating this form of content to estimate. A few examples would be if you had a newsletter item, but the Content Owner did not tell you how many articles would need to be created or an item with a 3 minute video and you have never filmed or edited one.
One modification I make for smaller teams or teams with shorter Sprint durations is removing the 40 and the 100. If we estimate items this large, we should probably take the time to break them into smaller chunks. We do this because, there may never be a point to where it is possible to bring this item in and finish within the Sprint timeframe.
Other times I will remove the 0 as well. Teams start to use this to represent content that shouldn’t take very much time. However, since time is involved, we would be forced to borrow from another piece of content. The ½ may be something you want to remove as well because of your quality constraints.
How do you Play Planning Poker®?
Here are the rules to Planning Poker® as well as a link to Mike Cohn’s site dedicated to the game. Just like any typical card game, it is always good to know the rules before you begin playing.
- Hand out cards to everyone who is estimating – You can get cards by either printing a template or purchasing them online. Each member who is a qualified Content Developer who will be working on the item is eligible for estimating. Content Owners should not estimate unless they are members of the Content Development team as well. The larger the group, the less likely this should be.
- Select a Content Item from your backlog to review – This should be based on the priority of the backlog item. Priority is determined by many factors, please review this post for more details.
- Read the information on the card – It is important to refresh your understand of the value you want to create before estimating a card. It is easy to get into a rhythm after producing several items before and forget that every piece of content is different and requires great insight into what the audience member wants or needs while remaining educational or entertaining.
- Ask the team if they have any questions to help clarify what is being created – Do not blindly estimate, you should have the attention of everyone in the process so ask questions and get clarification. If new ways to demonstrate value are decided, make sure you document them on the Content Item.
- Each individual selects an estimate and keeps their card hidden – Each Content Developer will choose the card that best represents his understanding. Typically you round up on these because taking away time means you have to take away needed value.
- Once all estimates are selected, show your cards – Everyone will drop their cards at the same time and a decision will have to be made.
- If estimates are a match, document – Great! We want consensus in our estimates. Write the estimate down either on the paper card or document in the digital system of your choice.
- If estimates are NOT a match, discuss and repeat estimate – At this point someone is low or high so we need to have a discussion to determine who is right. You need to break the item down into the elements that make it up and discuss why you think it is the size you believe. These conversations will get easier over time and the more we document the time it takes us to accomplish tasks, the easier it will get. Once you think you have consensus, either write the agreed number down or pull and flip cards.
- Select the next Content Item and repeat – Hurray! You have now estimated your Content Item and you should keep moving down your Content Backlog.
If you have estimated a significant portion of your Content Backlog for the first time, you can now gauge how much work you have to do. This may be weighty to see for the first time or you may be a little light on ways to reach you audience. Whatever it is, remember this is your first attempt at estimating so you will probably be slightly off.
But don’t worry, you will get better and once you start Sprint Planning and break the highest priority items into task elements, you will begin to see some of the flaws in your estimates. Just pull back up the item, hide the original estimate and have the team give a new number using the data available.