STOP the estimation game!

Context

One of the most difficult things to change in an organisation is the way they estimate software delivery and approach predictability. Yes, I said organisation and not just development team because estimation and prediction involves everyone:

• Managers use it as a decision making tool
• Managers use it to measure performance and success
• Products Owners use it to help ordering the Product Backlog
• Development Teams use it to plan the work to be done

Even finance, marketing, HR and procurement are affected by the way estimates are done, but that is even a more tricky topic.

Poison

The way a lot of organisations estimate software is completely poisonous. It destroys the most elemental values of innovation and delivering real value to customers.  Success becomes getting better making guesses instead of getting better delivering value to customers.

People prefer to become better predicting instead of becoming better delivering

Problem

It is proven the fight for better estimates does not improve delivery whatsoever. It does not make anyone happy (especially the customer). People always feel miserable estimating and then playing the game of  "Oh, wrong estimates again... so lets start the blaming culture once again :[."

Being better doing estimates is not the innovation differentiator any more

Besides that, only one person in a million is good defining software estimates. The human brain simply does not cope with it. The variables are so many that it is completely impossible to predict all of them and understand their delivery impact.

But why continue doing the normal estimation game if it is so obvious it does not work?

• The estimation process becomes just another weapon managers can use to blame someone
• It gives a short-term sensation of control (to everyone actually...)
• Everyone else is doing it as well
• Some people just like to continuously be miserable and waste time

Breaks my heart

• Performance and success based on story points
• Teams planning their work using hours
• Extensive planning sessions for big chunks of work
• Blaming culture because of "wrong" estimates

Result

• No focus on what the customer really needs
• Quality decreases enormously

The game to play is not the predictability one; it is rather the adaptability one

Advices

• Break down work items into a small or medium size
• Measure throughput instead of story points
• Measure lead time of your flow streams (cycle times are a nice to have because sometimes it is difficult to initially measure them; therefore start with lead times first)
• Putting hours against story tasks is just an incredible waste of time (by the way, just create high level tasks)
• Create product roadmaps and do proper user story mapping
• Define a product vision and testable strategy statements
• Understand the difference between a story and a hypothesis
• Understand the difference between vanity and actionable metrics

Summary

Trying to change how an organisation approaches predictability sometimes is just a lost cause since it is driven by the executive level. However, it does not mean we just give up. At least, educate yourself so that you can be prepared to make improvements when possible. I unfortunately have been meeting a lot of Scrum Masters and Agile Coaches with a traditional mindset concerning predictability...

Agile coaches and Scrum Masters, please keep in mind:

• A level of predictability needs to be achieved, but the way to get it can and must be simplified
• Predictability must be oriented to what customers really need over when someone things the customer needs something 
• Predictability is not a measure to understand success 
• Learn about no estimates
• Learn about lean startup and lean analytics
• Learn about the difference between resource efficiency and flow efficiency

Estimating is like gambling: you are just making bets