Would life be easier if we did not have to think about planning? Let's face it, whether it is about running our day to day routine or even something as simple as paying a visit to the neighborhood grocery store there is planning involved. Imagine taking a vacation without planning where you will go, what you will do while you are there, how long will stay and what you need to bring with you. Or imagine not thinking about how much you need to earn, to save and to invest to meet your worldly needs.. Well lets just keep those worries aside for a few minutes shall we and think about how planning can actually help! Yes that's right, no matter how much burdensome it might sound, planning can help.
The simplest definition of planning that I have come across is "decide on and arrange in advance". I like this because this definition is light, it does not say how much detail we need to plan in, but an important thing that it does say is that planning happens in advance.
There are many reasons why we plan. Our limited information about the future and vision of what we want to do with that information in order to achieve a specific objective gives us an opportunity to play with creative alternatives and paths to get there. Applying constraints helps us narrow down our search of solutions, breaking them down into chunks or steps that are (hopefully) easy to understand and sequence as necessary. When steps and solutions are not known we put on our experimentation hats and navigate our way to discovering the possibilities.
So what should I tell my teams when they get planning? What are the things I need to keep in mind but don't come easy, specially with Scrum? In my note to self, I include the following.
Plan early but not too early. Have a high level roadmap that (can change and) tells the team the path we will take if things went all-right. Planning too much too soon could mean a lot of wasted effort and repetition. Planning too little could mean little predictability in outcomes for initiatives that span multiple sprints. So lets groom our backlog stories and have them ready for a few (may be two to four) sprints to come. Lets find that right balance for our team where Sprint planning sessions do not become painful and burdensome.
More often than note, teams are asked to meet an "aggressive targets". In other words, there is a huge tendency to ask for deliveries tied to dates without understanding scope and effort involved. These dates are usually based on business constraints and are tight (more often than not). So, organizations and teams find themselves in situations where teams are unable to deliver as some say - "on time" and "within budget". Lets avoid that for our team through scientific and consistent use of measures such as velocity and cycle time. Lets make forecasts based on data so that business understands our non-trivial effort and can make realistic, prudent decisions on scope and priority.
There is much debate on whether to do estimates or not. What seems more important to me than that debate is to agree on an approach that works for our team and for your organization and follow it with discipline. When we estimate, lets spend little time estimating, after all estimates do not produce value for customer, working software hopefully does. Estimates are just that - "estimates", they are "not actuals".
What should the team work on first and what should it work on later? There are multiple factors that decide this, chief among them are dependencies, customer feedback and risk. The product owner plays an important role balancing all aspects while deciding the order so that team delivers value while making continuous progress. So lets keep in mind the problem we are trying to solve, and order our backlog on the basis of that. Lets identify those things that are not clear and run spikes. Lets reduce the risk early as early as we can, so that we do not embarrass ourselves in front of our stakeholders and customers when the time comes.
Lots of times, teams work with other teams to deliver an experience or functionality. Many times this needs dependency management, and back and forth discussions. This can lead to side effects such as reduced velocity, need for additional testing to retain quality, development of interfaces and/or APIs etc. When we get into such situations, lets find out if we can minimize dependencies through techniques like creating mockups, and integrate early and often. Let us ferociously automate testing and deployment (at all levels) so that we can identify and resolve issues quickly.
Lastly, should planning be one time? Absolutely not! The team should be ready to embrace change as they get better insights into customer needs and potential solutions. Lets dispassionately accept new stories, reject old ones that are no longer useful and update ones that need to be updated. Once again, an active Product owner and Team partnership can make a big difference. It is not as important to stick to the initial plan as it is to fulfill needs of the customers.
Related reading:
What is planning?
The simplest definition of planning that I have come across is "decide on and arrange in advance". I like this because this definition is light, it does not say how much detail we need to plan in, but an important thing that it does say is that planning happens in advance.
Why plan?
There are many reasons why we plan. Our limited information about the future and vision of what we want to do with that information in order to achieve a specific objective gives us an opportunity to play with creative alternatives and paths to get there. Applying constraints helps us narrow down our search of solutions, breaking them down into chunks or steps that are (hopefully) easy to understand and sequence as necessary. When steps and solutions are not known we put on our experimentation hats and navigate our way to discovering the possibilities.
Use the plan as a defense strategy
So what should I tell my teams when they get planning? What are the things I need to keep in mind but don't come easy, specially with Scrum? In my note to self, I include the following.
Plan just in time, and just enough..
Plan early but not too early. Have a high level roadmap that (can change and) tells the team the path we will take if things went all-right. Planning too much too soon could mean a lot of wasted effort and repetition. Planning too little could mean little predictability in outcomes for initiatives that span multiple sprints. So lets groom our backlog stories and have them ready for a few (may be two to four) sprints to come. Lets find that right balance for our team where Sprint planning sessions do not become painful and burdensome.
Plan based on data..
More often than note, teams are asked to meet an "aggressive targets". In other words, there is a huge tendency to ask for deliveries tied to dates without understanding scope and effort involved. These dates are usually based on business constraints and are tight (more often than not). So, organizations and teams find themselves in situations where teams are unable to deliver as some say - "on time" and "within budget". Lets avoid that for our team through scientific and consistent use of measures such as velocity and cycle time. Lets make forecasts based on data so that business understands our non-trivial effort and can make realistic, prudent decisions on scope and priority.
Maintain discipline, Estimate or #NoEstimate..
There is much debate on whether to do estimates or not. What seems more important to me than that debate is to agree on an approach that works for our team and for your organization and follow it with discipline. When we estimate, lets spend little time estimating, after all estimates do not produce value for customer, working software hopefully does. Estimates are just that - "estimates", they are "not actuals".
High priority, High risk first..
What should the team work on first and what should it work on later? There are multiple factors that decide this, chief among them are dependencies, customer feedback and risk. The product owner plays an important role balancing all aspects while deciding the order so that team delivers value while making continuous progress. So lets keep in mind the problem we are trying to solve, and order our backlog on the basis of that. Lets identify those things that are not clear and run spikes. Lets reduce the risk early as early as we can, so that we do not embarrass ourselves in front of our stakeholders and customers when the time comes.
Minimize hand-offs, eliminate technical debt..
Lots of times, teams work with other teams to deliver an experience or functionality. Many times this needs dependency management, and back and forth discussions. This can lead to side effects such as reduced velocity, need for additional testing to retain quality, development of interfaces and/or APIs etc. When we get into such situations, lets find out if we can minimize dependencies through techniques like creating mockups, and integrate early and often. Let us ferociously automate testing and deployment (at all levels) so that we can identify and resolve issues quickly.
Plan often and monitor, welcome change..
Lastly, should planning be one time? Absolutely not! The team should be ready to embrace change as they get better insights into customer needs and potential solutions. Lets dispassionately accept new stories, reject old ones that are no longer useful and update ones that need to be updated. Once again, an active Product owner and Team partnership can make a big difference. It is not as important to stick to the initial plan as it is to fulfill needs of the customers.
Related reading:
- Agile estimating and planning, Mike Cohn
- The NoEstimates Hashtag, Woody Zuill
- The Scrum Guide
Comments
Post a Comment