Tuesday, October 20, 2015

How to increase velocity? No silver bullet, but it can be done

Ugh... increase Velocity? That question again?


In a discussion forum, I came across this comment that "Velocity is an awful term and one that is often weaponized like it has been in this thread by asking how to get more. It's the wrong question because the goal is to become predictable much more so than faster."

I do not agree that velocity is an awful term, but I do agree that it is often misused (and even weaponized). If you intend to use velocity the right way and are interested in using it as a tool to understand and improve your work outcomes, this post is dedicated to you.

I would like to note that some of my friends are passionate about #noestimate. Whether you go that route or not, is your choice. The important thing is that is the team performing or worrying about numbers and metrics. If velocity discussions are weighing down the team, that is usually a sign of a different problem.

As you go through this post, keep in mind that increasing velocity may not be your real and only problem, and should definitely not be your main goal.

Addressing the "Real Slim Shady"...


If you want to address the issue genuinely please understand there is never really an end.


If the team is rapidly delivering customer value, communicating well and learning from its experience you may only have limited leverage on what the team itself can do. The team may need support from its ecosystem to take it to the next level. At the same time, there will still certainly be things that can team can do to make improvements. and is improving velocity the end goal? Or is it delivering customer value rapidly?
Let's address that.

 

Step 1. Who is asking?


Before you go any further it is important to understand who wants to increase the velocity. Many a time someone higher up in the food chain may want to know how you will improve the velocity. It may also be a software developer on the team, the whole team, the product owner or the scrum master (is that you?) who is interested. Or it may be a stakeholder, project sponsor or anyone else as well. Understanding who is concerned about improvement in team's velocity can help you in several ways in the steps that follow.

Step 2. Why are they asking - their motivation?


Before you jump into a solution, identify why they are asking the team to increase its velocity. What's in it for them (who want the team to improve)? As you dig deep into the reason for asking, you will likely discover the intent (good, bad, helpful, unhelpful) of the person asking the question. You will become sensitive to their needs and discover their situation. Are they interested in improving communications and flow of value or do they have a personal gain in mind by highlighting the problem? Answering why will help not just with framing the response better but also with taking more informed actions. Whatever the reason is, think about how that can help you drive improvements.

Step 3. Draw insights - Where are the problems, what are the solutions?


Based on your study of your context, would it be fair to say there are potential opportunities for improvement? Where could the team do better?

Imagine a team delivering really really fast but the customers are not responding favorably. Could it be the lack of quality or delivery of functionality that customers do not need? In either case, increasing velocity may not be the best solution. We need to think smarter.

Try these questions first:

  1. Is the team delivering new features and improvements frequently?

  2. How are customers reacting to the product deliveries and latest features?

  3. Is the team generally enjoying their work and having fun?


If you spot problems while answering these questions, search deeper. Here are just a few examples of what you could explore.
People questions:


  • Do measures such as velocity drive positive behaviors in your organization? If not what can be done to fix that situation?

  • Are the roles well defined? Is there sufficient empowerment? Are the roles of PO, SM, and Development team clearly understood?

  • How is communication between the team members?

  • How strong are the leaders of your team, project and the organization?


 
Process questions:


  • How do you assess progress towards an important goal? Do you have an objective method?

  • Are team members working on multiple diverging goals and functionalities and have high work in progress?

  • How are releases managed and rolled out - small chunks or big rocks?

  • What is the team doing to continuously learn from its journey?

  • How are (Scrum) meetings running?


 
Technology questions:


  • How is the team doing on the account of technical debt?

  • Is there sufficient test automation and continuous integration?

  • How is your quality? Do you have a huge bug backlog?


Draw insights together with the entire team to understand all perspectives and see what can be done. Dig into the backlog, measurements that the team thinks are relevant, such as velocity, cycle time, defect trends, automation, build failure rate etc. Looking for problems, and asking why over and over again, should give you new insights on what can be done.

Shortlist a few experiments to try out.


Step 4. Sell, Try, Feedback, Repeat (Go to step 1)!


As you go through this exercise, you may come up with a range of solution(s). Some solutions may require senior management intervention, others may need more involvement from the team. Channel the conversations around the solutions, convince the stakeholders to try them out. Clarify how you will be reviewing the outcome - set a time-frame for that. Try the solution and get more feedback!

You may have to repeat these steps several times, to continually get the results you are hoping for (including but not limited to an improved velocity, as that may not be the real goal you are after).

Good luck!

No comments:

Post a Comment

14 Essential Software Engineering Concepts for Engineers and Managers

There are many terms and concepts that are important for an engineer to be familiar with, in order to effectively build software. This post ...