hirendoshi, Author at Practice Agile

27 Oct 2014

Evidence-based Estimation

An analogy I can think of is… I want my dart to hit the dart board, and not necessarily the bull’s eye…. as it calls for a lot of details which apparently is missing during estimation. However, if my dart doesn’t hit anywhere on the dart board… it’s almost like shooting in the dark; a very disappointing estimation scenario.

A vast chunk of new teams struggle in arriving at a correct project estimate; with most them failing due to a huge variation (with estimated Vs actual)! Very few succeed in this journey by coming closer to the actual in the initial cycles itself.

Project managers in traditional development focus on detailed scope, so that the cost, estimate, and time are accurate and frozen before kick starting a project. How far this is successful and precise is debatable! In agile projects, during release planning the development team arrives at a high-level estimate for each story in the product backlog normally in story point which aids in finer estimation during sprint planning. This helps the team to get started, rather than wait for all the project details to get finalized.

Story point is an arbitrary measurement of a feature’s size relative to other features and not the time needed to complete a feature. For example, by looking at the picture of the whales you cannot determine the exact age or weight of each whale but can compare the size of each with respect to each other. This is a very handy approach as detailed information to estimate the effort may not be available so early. Story points are used to calculate how many user stories a team can complete in a sprint which is termed velocity.

The story point size assumptions are interpreted using an estimation scale, the most common ones include numeric sizing (1 through 10), t-shirt sizes (XS, S, M, L, XL, XXL, XXXL), Fibonacci sequence (1, 2, 3, 5, 8, 13, 21, 34, etc.), etc.

A team to get started with story point estimating must be clear, synchronous, and agree upon the below aspects (I call it the ‘Five subtle rules’ as they are hidden and work at the back of our mind.).

Common reference – Each team member estimating should have a same reference. For example, the size L to measure a story point size should be the same criterion for everyone in the team. Any conflict with respect to the reference index results in huge variation and wrong estimation. Agreeing upon a common reference is very critical.

Collective wisdom – Estimation is a team activity and hence all should participate in this. Having one person do this will be a huge risk as the margin of error will be high! However, a collective decision helps in arriving at a common estimate which in all probability will be accurate.

In addition, during relative estimation the team should also take into account the effort, complexity, and uncertainty for every story before arriving at a number.

Effort – How much effort a story would take to complete, relative to the reference story.

Complexity – How complex is the story with respect to the reference story.

Uncertainty – How much risk/unknowns a story holds relative to the reference story

Even with this common agreement, it may still take a few iterations for the team before arriving at a common estimation value. With estimation being definitely hard, how do we get the best estimates of story size?

Why not take a closer look at “Evidence-based estimation”, learn from experience, become proficient with every calibration, and adopt during the initial project estimation phase to understand its true benefits!

Here is the approach a team should take to proceed with “Evidence-based Estimation”:

Let’s consider a scrum team of 7 members commence a 2-week sprint of 10 working days. With 6 hours as the effort per day, the capacity would approximately be 7 * 10 * 6 = 420 hours.

During the 1st sprint planning meeting, the team begins story estimation from scratch with no story points assigned to stories in the product backlog.
A story is picked from the product backlog followed by task and time breakdown for each. This step continues till the estimated capacity adds to about 400 hours/5 stories (considering the average team capacity is 420 hours).
Through the sprint, the team works towards completing all the planned stories.
Just before the retrospective meeting, the team assembles to story point the completed stories (definitely now with some experience/evidence).
Among the 5 stories, the team picks the one with medium effort, complexity, and uncertainty and assigns a story point (keeping the 5 subtle rules in mind). For example, when we take an estimation scale of 1, 2, 3, 5, 8, 13; this story is assigned a story point of 3 and it becomes the common reference story. A closer look at this exercise clearly indicates that the story point number is emerging out of working experience/evidence and not by mere guess work!
The team continues story point estimation for all the stories in Sprint 1 (per Step 5) and assigns a value lower, higher, or the same when compared to the reference story point.
As the team proceeds from one story to another, the comparative reference points become varied apart from being evident which helps them with better estimation. At this point if the team feels a need to refine the previous estimates, they can. The idea is to get better estimates with experience that are realistic.

The teams may decide to use this estimation during the initial few sprints till they are confident of the process. With experience, the team becomes an expert and are equipped at deriving an almost accurate story point during the release planning (rather than after the completion of a sprint), thereby aiding in better agility and transparency.

29 May 2014

A Collaboration Lesson from the Redwood Trees

The Redwood trees in Muir woods, San Francisco, are some of the tallest and widest trees on planet earth. Some trees are so wide, that it takes 40 grown up adults to make one circle around it! These trees are known to live for thousands of years and each year they grow bigger and bigger. Most tall trees have roots that go deep down to keep them stable. For example, the roots of the palm tree are as deep as the height of the tree. However, research tells us that though the Redwood trees are the tallest trees on earth, their roots are not deep at all. California has gone through a tremendous number of hurricanes, cyclones, windstorms, tornadoes and earthquakes in the past 2000 years. Any normal tree would have been crushed instantly, but not the Redwood trees. They’ve been standing tall and strong ever since. So, what is the secret behind these trees?

It seems the Redwood trees have some special mystical power. The roots grow outwards instead of deep under the ground. And when these roots come in contact with roots of the another Redwood tree, they wrap around each other multiple times and form a strong bond. Each tree shares a bond with another tree through its roots, and eventually, every single Redwood tree is connected to each other. The roots hold on to one another through the harshest of weather, and keep the family of trees standing tall and strong, together.

The roots of older and wiser trees hold on to the roots of trees just beginning to grow. They’re basically telling them —“You’ll grow big and strong. Reach for the sky and we’ll help you get there. You have the strength of hundreds of trees in the forest because we’re all connected. Our strength is shared together, and it grows together.”

If we dig deeper and try to understand self-organization, a principle very much needed for being Agile, we will realize that collaboration and co-operation are the basic attributes needed to self-organize. Similar to the Redwood trees, Agile team members need to collaborate so that the entire team rises higher and higher. This interconnection helps each individual to grow, inspire and motivate each other. It builds trust and transparency, and at the same time a feeling of safety to explore new hypothesis, innovate and produce a quality deliverable to the customer, with a feeling of satisfaction.
The Redwood tree example was shared in one of the spiritual discourses and I found this example so very relevant to some of the Agile values and principles.

24 Mar 2014 - 1

The Power of Anonymous Retrospectives

The Scrum values – Openness, Commitment, Focus, Respect and Courage are the foundation for the behavior and practices in Scrum. It’s difficult for organizations to adhere to these values from the onset of their Agile journey. Agility is about behavior and cultural changes and the values listed can’t be demanded, they have to be earned by creating transparency and is a journey that never ends.

While coaching Agile teams over the years, I have learned that no matter how open and transparent the organization is, there will be some individuals that won’t openly speak up. They are mostly introverts and have a phobia of public speaking. They will limit their interaction to a bare minimum. In some organizations engineers carry the management fear. They feel like they are being observed and anything and everything they say will reflect in their yearly performance appraisal and they clamp up. In either of these situations or situations similar to these, the Anonymous Retrospective helps get the real pulse on the floor. It helps all the participants open up and talk about what they really feel deep within about the organization, the culture, the people, the leadership, the technology, the motivation factor, etc.

So what is Anonymous Retrospective? As the name suggests, all data collected is completely anonymous. The first rule of the Anonymous Retrospective is the data collected has to be truly anonymous and there should be no attempt to tie the same to any individual – either through the language used to describe it or with an handwriting match. What I normally do is I put an empty container in the middle of the room and give each participant a bunch of Post-its. I ask them to jot down their thoughts on what is working well and the areas that need improvement. I emphasize to the participants that this is a Anonymous Retrospective and encourage them to share their thoughts without having an iota of worry. I then ask them to fold the post-its and put it in the empty container. I normally time box this to around 25 to 30 minutes.

Once everyone is done, I get the container with the post-its and give it a good mix. Then I appoint someone neutral to take notes on their laptop and help me with basic categorization. I pick each note and try to read it as verbatim as possible, except for certain cases where there are personal attacks. I read them one at a time in front of the entire room, the person taking the notes captures it, and then I tear the post-it in front of all the participants and put it in the trash can to maintain confidentiality. This is what I mean by “Anonymous”.

Anonymous Retrospective will generate plenty of data which has to be validated for its accuracy. Once all the data points are collected, work with the participants in the room to finalize the categorize the data, and generate information by connecting data in each category together. Following this put an action plan together to address each category. Most likely you will need multiple sessions to do this… but remember you now have some solid facts with which you can incrementally introduce improvements. This is what the power of Anonymous Retrospective is!

I am happy to hear your thoughts.

10 Oct 2013

Tesco.com uses agile development offshore to roll out internationally

This article ties to the ongoing work I have been doing with the International Teams in TESCO as an Agile Coach in Bangalore since January 2011.

https://www.computerweekly.com/news/2240206944/Tescocom-uses-agile-development-offshore-to-roll-out-internationally

Overall, Computer Weekly covered the article well just with a minor correction: “Deliver the Beer” was a first internal release to validate Agile Software Development was the correct approach. Czech Republic had a first fully functional grocery website.

28 Dec 2012

Agile Coaching: Are your retrospectives effective?

Are you in a situation where your team(s) has been practicing Agile for a while and teams are following ceremonies meticulously, but still there are no significant improvements sprint over sprint or release after release? If yes, I have some antidotes that I will share through series of blogs that you can experiment with.

One of the Agile principle is, “At regular intervals, the team reflects on how to become more effective, then tunes and adjusts its behavior accordingly”. This in a nutshell is Continuous Improvement (CI) and one of the ceremonies that assists you in implementing CI is Retrospectives.

Start with Retrospectives: Ask these questions.

Are the retrospectives effective?
Are the team members open & honest?
Is there a good flow and exchange of information that is fact based that the team can relate to?
Is enough flavor added to each retrospective to ensure that they don’t become monotonous?
Is the facilitator neutral?
Did the team put the action plan for the improvement areas after root causing the problems?
Is the team taking at least one improvement idea that they are in total control of instead of relying on parties outside their team?
Is someone within the team held accountable to ensure the improvements are put in action?
Did the team reflect back on the improvements implemented in the retrospective that follows?

My observation as a Agile Coach has been that teams are generally very enthusiastic to begin new work as soon as current work is completed and they cut corners or miss on Retrospective entirely thereby missing on a important Agile Principle – “Inspect and Adapt”

27 Dec 2012

Tesco’s Success Story with Agile Adoption

Over the past 2 years I have been helping Tesco’s Dotcom International Grocery Home Shopping (IGHS) group in the capacity of Agile Coach to build their eCommerce Platform. Tesco Dotcom’s challenge was to take the world’s largest grocery website international to multiple countries outside UK as quickly as possible and be the market leader.

As the saying goes, “The proof is in the Pudding” …. By using Agile as a Software Development Methodology with a combination of Scrum, XP, Kanban and lean principles of choice, Tesco was able to launch their Dotcom operation to new countries regularly and is currently live in 5 countries within the span of just over 2.5 years. This group with several Agile teams distributed across 2 geographies was able to bag 4 major awards within the organization, including the “Tesco – IT project Cup” of the year.

It is my privilege and honor to be part of a journey with this passionate team that was constantly hungry to take the Agile adoption from one level to another tirelessly through continuous Inspection and Adaption and both the passion & Hunger continues….

03 Oct 2010

Diminishing ROI with NO Automation in Agile development

My interaction with various clients that are in process of transitioning to Agile often ask this question – Is it ok to skip automated Unit and Acceptance testing and only perform manual testing? The reservation normally arises due to old school of thought where manual testing was tolerated and where CLIs and API needed for testing are normally an after thought. Sometimes it is also because QA engineers are manual testers and have no needed programming skills. And at times people tell me Automation takes time off from active feature development and is an overhead.

No matter what is the reasoning, I always preach to these teams, “Automation is a rule and not an exception” and I try to justify my statement using the example below.

The diagram below, “Agile with No Automation” helps illustrate the impact of skipping automation totally, causing diminishing ROI. The goal of each sprint should be a running software that is fully tested, integrated and potentially ready to ship. To meet this goal all features have to be regression tested at end of each sprint.

For the purpose of this example let’s assume Scrum team is able to do the same amount of work each Sprint and each feature is of same size.

In Sprint 1, the scrum team completed 3 new features without any automated tests. All testing is done manually to meet the sprint goal.
In Sprint 2, the team is only able to complete 2 features as the time needed to develop the third feature has to be used to perform manual regression testing of 3 features developed in Sprint 1.
In Sprint 3, the team will be able to complete only one new feature as the time available for remaining features will be used for manual regression testing of 3 features from Sprint 1 and 2 features of Sprint 2 in addition to performing manual testing the new feature delivered in the current sprint, i.e. sprint 3.

As you see in the above example, every new sprint team is spinning wasted cycles doing manual regression testing the old features which could have been totally eliminated if all tests are automated.

Now, let’s consider this scenario where all tests Unit and Acceptance tests are automated each sprint with illustration in the diagram below, “Agile with Automation”.

In sprint 1, the scrum team completed 3 features and all test are automated. When the team says done, they are indeed done – everything coded, tested, integrated and working software.
In the following sprint, i.e. sprint 2, the scrum team will be able to complete additional 3 new features as regression testing is automated. Remember, it is machine time as opposed to laborious human time spent running manual regression tests.
In sprint 3, the team develops 3 additional new features without worrying about regression testing the old features.

Automation gives developers and quality assurance engineers enough confidence to make the necessary changes in the code base thereby generating a potentially shippable product every sprint.

The above example shows that ROI without automation is 6 features and in the same time frame the ROI is 9 features with automation.

Remember, to reap benefits of Agile Software development, “Automation is the rule and not an exception”.

31 Aug 2010 - 9

Template of task breakdown for a user story

In a few Agile/Scrum training and coaching engagements that I have been involved with, I shared the template of ‘task breakdown for a user story‘ and found that teams find it very useful. This task list also defines ‘done’ criteria for the story. I am sharing the same template with everyone in this forum for your reference.

As a rule of thumb, I recommend scrum teams to define a user story that can be completed in 5 Scrum working days by one or more members. Also, as a rule, the task should be granular enough to be completed in 5 hours max (1 Scrum day). Defining granular tasks enforces discipline among team members to develop and deliver small chunks of functionality everyday to mainline source repository instead of delivering a large chunk of code together which is more error prone.

Here is a list of task breakdown for a user story. Please note, this is a reference list and all tasks listed below might not be applicable to each story.

Design for the story – This task is specifically for white boarding or group discussion on how story will be implemented.
Implementation of the story – This task could include defining the interface and methods needed to fulfill the story. There can be multiple tasks for this type.
Write unit tests (TDD) – I encourage and enforce teams to write code using Test Driven Development (TDD). For any coding story this task is mandatory.
Write Acceptance tests – This task is used to write the automated acceptance tests that will help with the story acceptance.
Non functional requirements (NFR’s) – For each story, quality attributes like security, performance, usability, testability, maintainability, extensibility and scalability must be considered.
Code Review – As a rule, every piece of code delivered to mainline repository should be peer reviewed.
Code Re-factor – Beautifying the code by restructuring. This task should be included for each story to avoid a complex mesh of code later.
Mock interfaces – At times, development can take a while to produce real implementations of the interface. In such cases, mock interfaces must be provided to help QA engineers with a head start.
Exploratory testing -Perform Ad-hoc testing for the story. This could include performance testing, scalability testing, etc.
Bug fixes – Depending on the story, appropriate time should be devoted for bug fixes
Bug verification – Depending on the story, appropriate time should be devoted for bug verification
Demo – Internal demo of the story within the team. I like to add this task as it helps Scrum team members to be on the same page. After the daily standup, if a story is complete members can take few minutes and educate others on the implementation of the story.
Update Wiki or document repository with the design and findings of the story
End user documentation – This is for updating the reference guide, Install document, etc, if applicable.

There could be other tasks for building on the CI server, Integration testing, etc that might have to be included based on your environment. If you can think of any additional tasks that I am missing from the above list please share the same so I can update the above template.

29 Jul 2009 - 3

Scrum team organization – Feature teams or Component teams – Part 2

Component Teams (CT) are specialized teams organized around the architecture of the product under development. Examples of such teams are UI / User experience teams, Database design and modeling teams, team of Architects, Security team, etc.

A typical component based scrum team is more or less similar to a feature team i.e. each CT will have 3 to 4 developers, 2 to 3 testers, a Product Owner & Scrum Master with Documentation and Architect representation as needed.

The top node of the diagram “Theme/Epic”depicts customers value add features (Features that customer value). Think of this as some high level requirement that marketing will use in their presentation to advertise a product. This value add feature might not be small enough to finish in one sprint and gets decomposed in to multiple smaller features i.e. Feature 1, Feature 2, Feature 3, and Feature 4. What this essentially means is, once Feature 1 through 4 are done in its entierty, the top level Theme/Epic is “Done”.

The stories for component based teams are fed by these decomposed features Feature 1, Feature 2, Feature 3, and Feature 4. As the diagram shows, Feature 1 is dependent on Component Teams C1, C2 and C3. Similarly, Feature 2 is dependent on component team C1 and C2, Feature 2 is dependendent on CT C1 and C3 and so on. Component teams in general provide functionality and serve multiple Features. Component teams generally don’t generate products that get shipped to the customer. But they develop functionality that is consumed by feature teams and indirectly helps add value to the features being delivered to customers.

If organizations are structured in a way that specialized teams are spread globally then forming component based teams might make sense. Component based teams are great to keep site affinity, work across different timezones and also to keeping team culture intact. But it also comes at an expense of management overhead.

Few challenges with component teams based on my experience

There is overhead of integrating component team’s delivery with the top level features. In the example above, when C1, C2 and C3 finishes the deliveries for Feature 1 they have to be integrated together in Feature 1.
To ensure all scrums are aligned well to deliver end-to-end customer functionality, it will take lot of upfront planning as well as tracking during the entire sprint.
In my experience, with component based teams it’s very difficult to deliver thin slice of end-to-end functionality at end of each iteration because feature team needs time to harden (stabilize code base + write integration tests) their code with component team code delivery.
Since component team serves multiple features teams, negotiation needs to happen between Feature Team and Component teams to avoid starvation. For Ex. In the diagram above Component Team C1 gets requirements from Feature 1, Feature 2, and Feature 3. In the upcoming iteration let’s say C1 can only serve Feature 1. In this case Feature 2 and Feature 3 teams will have to wait until next iteration before their needed functionality is delivered and will be starved.

Thus, for component teams to succeed open communication channel should be established throughout the organization.

Please share your experiences.

Blog