The misconception of estimation

Marcel Stieber
Nov 3 · 5 min read
Let’s quickly estimate the number of peas here.

The idea is fairly simple. We estimate stories with numbers to know their size. This way it’s easy to plan a sprint upfront. It helps us to predict what we can do within a time frame of a sprint and when we will deliver a feature. The number also shows us, if a story is too big and needs to be split. Predictable and reliable. Perfect for the team, management and stakeholders. And there’s the rub. Let’s step back, catch a breath and scrutinize this.

What does it mean, the size of a story? What are we estimating? Effort or complexity? What’s the maximum points to plan for a sprint? I had this discussion several times in different teams. But discussing this when establishing the estimation process is like putting the cart before the horse. There are two chapters for this topic. The first one which I’ll handle in this article is about the estimation itself. The second one is about what to estimate and will be tackled in a second article.

To understand estimation better you need to understand what purpose you are addressing.

Here are some common examples:

  • have a reliable prediction
  • know when something is finished
  • understand the scope of a story
  • everyone is on the same page about the story
  • know if the story is ready to be planned
  • avoid surprises and delays
  • prioritize better

And how does estimation help us with this in real life? A common pattern is this: There is a ticket someone prepared. It’s presented to the team and everyone is estimating. If there is a consensus the number is assigned to the story. If not, there should be a discussion between the highest and the lowest estimation. Pretty often this discussion reduces to “I was not quite sure, so I’m fine with x as well.” Or the higher number is taken to be on the safe side. Another solution I experienced a lot is something like a democratic decision. Sounds bad. It is. The only thing we achieve with this is having an abstract number bound to the story. But our underlying purposes are not satisfied. We have completely no idea if we have the same understanding or any understanding at all. And if we want to improve by this what do we actually learn? We get a benchmark how much we can finish in one sprint which will help us to plan the next. Any improvement done by this approach is fragile as anything which is not fitting into our experience and reference stories from previous sprints (e.g. a completely new epic or the “unlikely” event of a change in the team composition).

Those examples just show that the stories are not properly discussed which implies that the lesson we seem to learn from this is (almost) useless. The only purpose the estimation is serving here is the estimation itself. It’s giving a number for planing that is predestined to uncertainty and lack of improvement. Of course this can somehow work out for you. But let’s dive into the underlying intention of estimation and you will see that the numbers are only a side effect.

Basically estimation is a tool to give everyone’s unbiased rating and triggering a conversation. Especially in a heterogeneous team with extrovert and introvert members or big differences in experience this gives a stage for everyone to be heard. Actually, if this is needed to get everyone’s voice heard, you have a different issue. It should be taken for granted that in a healthy team everyone should feel save (and do so) to challenge the opinions of others.

Ask yourself: who needs to share information to get all team members onto the same page? It could be the Product Owner. Or it could be one part of the team that gives information to the others. Share this information and ensure that everyone is on the same page. Let someone who just received the information rephrase it and summarize what needs to be done and what’s the business value here. Discuss this. Putting a number to the story is just one part of the process to gather all information, to think stuff through and to get the whole picture to everyone. This way you will most likely not set a record for the most numbers estimated in a refinement meeting. But you made a good step forward to reach your goals. It’s a deception to think just putting a number to your story will improve your gut feelings about stories over time. This may be true for a certain degree but a gut feeling will be exactly this, just a feeling about something I do not know any better.


Let’s wrap this up

Estimation as a self purpose could give you the feeling you are doing fine. Your team is working and you are having a proper scrum flow. Congrats, you are deceiving yourself. Estimation is subordinate but gets into focus way too fast. It’s not about meeting your estimation at 100%. It’s about good products that are used by customers and not being perfectly predictable to deliver what you promised. Ask yourself if the number you are coming up with is the thing you need here. Are you giving this to stakeholders? What exactly are you communicating here? What do they understand when hearing a number like 68? Or is it your planing where you have a limit and you just throw in tickets until you reach that number? Does it help you with your planing to have a number or is a common understanding the thing you need? Or is the number a hint if you need to split the story into smaller pieces? Splitting a story just to get a smaller number in your bonny statistics but not delivering anything useful to the user does not make a lot of sense except for your statistics.

Be aware that doing agile does actually not mean to make you faster, it means to slow you down in short term. But to take your steps deliberately makes you faster in the long run. Estimation will still be a guess, not a guarantee. But you start to get better and reduce the guessing a bit, if you start communicating.

Marcel Stieber

Written by

Senior Software Engineer

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade