Forecasting with confidence

Forecasting has always been a subject of interest to all project management teams. Team have used multiple methods (from detailed MS Project scheduling to Earned Value Management) to try and estimate when a project will finish. People spend a lot of effort to come with one precise hard date knowing fully well how their past predictions have failed, how their assumptions have been invalidated, etc.

If one were to discuss further with a Project Manager, they will understand that there is a level of confidence that a Project Manager associates with this date.  Perhaps, he is very confident (90% - it is rarely 100%)! It could be even as low as 60-70% depending on the number of issues and risks in the project. In other words, an end date is always associated with a probability, which we in our management talk and presentations, ignore or fail to highlight. In reality, a Project Manager would like to say - I am 70% confident that can deliver this scope of work within the next 3 months but 90% confident that I can deliver this in the next 4 months.

Kanban systems predict the future by extrapolating current throughput on the present backlog. Most Kanban tools use the CFD for the same. This gives you the gap between the rate at which you need to finish your work versus the rate at which you are finishing work at present. Like MS Project, it also gives one date when this backlog will finish, assuming current throughput.

Fact is that future is not definitive! Project risks change, requirements change, team resources change... the list goes on! So, just like this inadequate for traditional tools, it is inadequate to limit ourselves to this analysis in the Kanban Method also.

This session, using Kanban Tool, will demonstrate:

  1. The use of probabilistic theory to have a confidence % associated with a particular delivery date
  2. Once you know the gap between the current throughtput and the desired throughput, what resourcing changes can be done to bridge this gap. While it is understood that resourcing is not the only option to bridge the gap, it is perhaps the most commonly used method. 

This approach has been vetted by David Anderson and has been presented in the LSSC 2013 conference by one of my colleagues.

 
7 favorite thumb_down thumb_up 12 comments visibility_off  Remove from Watchlist visibility  Add to Watchlist
 

Outline/structure of the Session

  1. Explain the background the business problem that we are trying to solve
  2. Do a demonstration of how a simulation approach can be used to forecast

Learning Outcome

The use of simulation models in the Kanban method to forecast a project end date with confidence levels.

Target Audience

Delivery Managers, CXOs

schedule Submitted 3 years ago

Comments Subscribe to Comments

comment Comment on this Proposal
  • AgileSattva Consulting LLP
    By AgileSattva Consulting LLP  ~  1 year ago
    reply Reply

    Hi Sudipta,

     As we understand this could be applicable only if the requirement in any form in the backlog (could be stories, or small features) should be of the same sizes? So it comes down to sizing the stories!

    would you cover the above aspects of it, and also you may get questions related to Throughput (that you are proposing) vs Velocity. Would you talk about it as well?

    Deepak

    • Sudipta Lahiri
      By Sudipta Lahiri  ~  3 years ago
      reply Reply

      Hi Dhananjay:

       

      In a Kanban system, one does T-shirt sizing. So, one is able to get data like: A medium size user story takes between 10-14 days. A small size user story takes less than 2 days. Similarly, for other work types, like defects.

      The simulation model uses this base data and applies a distribution pattern to get a confidence factor on a given end date.

      Therefore, work items (not just requirements) don't need to be of the same size; they can be of different size and can be of different work types (user stories, defects, technical stories, etc.).

       

      Regards

      Sudipta.

      • Ram Srinivasan
        By Ram Srinivasan  ~  3 years ago
        reply Reply

        In a Kanban system, one does T-shirt sizing.

        I seriously question this. 

        A medium size user story takes between 10-14 days. A small size user story takes less than 2 days. 

        Really? Where did you get this data from? 

        • Sudipta Lahiri
          By Sudipta Lahiri  ~  3 years ago
          reply Reply

          I have enclosed an earlier Webinar that I did on this topi in March. Since then, our simulation models and the UI have had 2 major ugrades.

          The audio is a little broken in a few places. However, you will be able to understand the gist of the session.

        • Sudipta Lahiri
          By Sudipta Lahiri  ~  3 years ago
          reply Reply

          Hi Ram,

          I am only giving an example...a Kanban system will be able to give data like this. It will obviously depend from project to project. This isn't an industry standard of any sort. 

          Hope that clarifies... :-)

          Regards

          Sudipta.

  • Ram Srinivasan
    By Ram Srinivasan  ~  3 years ago
    reply Reply

    Hi Sudipta,

    What is unique about your topic/proposal that is not in David's Kanban book/ other sources? And why would you clasify it as an advanced topic? If this has been presented elsewhere, can you please share the presentation with us?

    Thanks,

    Ram

    • Sudipta Lahiri
      By Sudipta Lahiri  ~  3 years ago
      reply Reply

      Hi Ram,

      I am not aware if David's latest book covers this but at least the earlier ones did not. He has been talking about applying pricing differentials based on probabilistic simulation for the last 12-18 months. So, for example, if someone knows that I have a 80% chance of delivering in 4 weeks and a 60% chance of delivering in 2 weeks, how much premium would a customer be willing to pay if he really needs it in 2 weeks? 

      The only other person, literally, who has talked about this is Troy Magennis from Focussed Objective. He has close association with David. Leankit has used Troy's engine and integrated with their product.

      Since David has been an advisor to us, we have taken his inputs to build the simulation models. However, we have applied it a bit differently... and in a different area (because we feel that market is not yet ready for differential pricing).

      We are using this same thinking and model to do what-if analysis and help the Project Manager with a confidence % with which he can get back to business. So, he can make a statement like this: If I get 2 additional developers of this profile, my confidence of delivering to your 2 weeks schedule will go up from 60% to 80%. This is just an example but the power of this is HUGE.

      On this being an advanced topic - in most project teams that I am consulting to in India, people are just stuggling with the basics. Defining Agile Requirements, getting Cycle time to take a decreasing trend! For them, this is really beyond what they can chew now. Also, in order to do this, you must be able to interpret CFDs well. That leads you to think of what is the next What-If scenario that you should be simulating for. Once again, that needs a reasonable level of experience.

      On earlier presentations of this topic:

      1. I have talked on this topic when David visited India in Nov 2012. Video recordings are not available from this session.

      2. Mahesh Singh, our co-founder, has presented about this in LKNA 2013 in Chicago. I believe those are paid recordings but I will see if I can get access to them.

      3. I have done a webinar on the same and will give reference to it in the submission. This was an old webinar when we made our first release of the simulation engine. Since then, it has become a lot more sophisticated. We plan to do repeat webinars of the same in near future.

       

      Regards

      Sudipta.

  • Joel Tosi
    By Joel Tosi  ~  3 years ago
    reply Reply

    Hi Sudipta,

      In your other comment you mentioned offshoring.  Is this more tied to offshoring where someone somewhere else hands over to development some work and wants to know when it is done?

     

    Your explanation also confused me slightly.  You mention that Kanban is not enough, then you say in this session you will use the Kanban tool to show it.  Are you showing how Kanban addresses this uncertainty of delivery or are you showing something else?

     

    Best,

    Joel

    • Sudipta Lahiri
      By Sudipta Lahiri  ~  3 years ago
      reply Reply

      Hi Joel,

       

      I did add to the confusion! Pls ignore that comment.

       

      I was responding to Sachin's comment on whether Kanban emphaises on estimation.... and my response was that is doesn't. 

       

      Regards

      Sudipta.

  • Sachin goel
    By Sachin goel  ~  3 years ago
    reply Reply

    Hi - in my view, all end users will like to know how much the product will cost and when will this get delivered. Its apparent in Service Industry as they have to bid for it. In other functional organisations, internal IT teams also have to present a high level budget and scope for approval to spend.  Thanks - Sachin

  • Sachin goel
    By Sachin goel  ~  3 years ago
    reply Reply

    Hi Sudipta - are there plans / emphasis on the estimation techniques evolution in line with Agile practices. Apologies for my limited knowledge of Kanban pls. Thanks - Sachin

    • Sudipta Lahiri
      By Sudipta Lahiri  ~  3 years ago
      reply Reply

      Hi Sachin,

       

      Lean and Kanban guys generally like people to not spend time on detailed estimation. It is considered as waster because estimates are generally dead pretty soon once the project starts. Nevertheless, in the offshore services industry, estimation is part of the game. In most projects, the customer asks for an estimate and the Sales team has to give one. Of course, the lighter the estimation model, the better it is.

       

      Regards

      Sudipta.


  • Naresh Jain
    Naresh Jain
    schedule 3 years ago
    Sold Out!
    45 mins
    Demonstration
    Intermediate

    "Release Early, Release Often" is a proven mantra and many companies have taken this one step further by releasing products to real users with every commit a.k.a Continuous Deployment (CD).

    Over the years, I've built many web/infrastructure products, where we've effectively practiced CD. However at Edventure Labs, when we started building iPad games, we realized there was no easy was to practice CD, esp. given the fact that Apple review takes a few days.

    Our main question was: As mobile app developers, how should we architect/design our apps for CD?

    We were a young startup, learning new behavior about our users (kids aged 5-8) everyday. We could not afford any delay in releasing latest, greatest features to our users. To solve this problem, I believe we've built an innovative solution to enable any mobile app developer to achieve CD.

    If you are building real products, which have platform/3rd-party dependencies and you want to practice CD, this session is for you.

  • Liked Mike Burrows
    keyboard_arrow_down

    Kanban through its Values: An Agenda for Scale

    Mike Burrows
    Mike Burrows
    schedule 3 years ago
    Sold Out!
    45 mins
    Talk
    Intermediate

    Introducing the Kanban method through a 3-layered value system - a familiar core that stimulates and drives change, a middle layer that is about direction and alignment, and a protective outer layer of discipline and working agreements.

    This humane, values-centric model aligns Kanban with the concept of the Learning Organisation and suggests ways to seek resonances with other methods. It has some practical benefits too: it can help us engage more effectively with the organisation as it currently is; it encourages us to self-reflect on our effectiveness as agents of change; it provides a convenient framework for the capture of stories.

  • Liked Sudipta Lahiri
    keyboard_arrow_down

    Capacity Planning for Dynamic Teams

    Sudipta Lahiri
    Sudipta Lahiri
    schedule 3 years ago
    Sold Out!
    20 mins
    Experience Report
    Intermediate

    Fixed price (and fixed scope) projects dominate the offshore industry. These projects have offshore/onsite teams. They often have large team size (over 100s of people in one team).

    Agile thinking uses team velocity/ throughput and uses that to project an end date (Kanban system) or how much scope can be accomplished in a given time duration (number of sprints in SCRUM). They assume a stable team. However, this is not applicable for projects. They experience resource and productivity ramp-up issues. Often, resources keep changing as new projects come in. Projects do not have past velocity or throughput data. Extrapolating historical data from other similar projects, though possible, is inaccurate for multiple reasons.

    This talk is based on our experience of working with such project teams. They want to adopt agile methods. We show how they can adopt the Kanban Method and yet do: A) Initial Capacity Planning B) Assess the impact of scope creep to the project end date.

    The session assumes a basic understanding of the Kanban method.

  • Vinod Sankaranarayanan
    Vinod Sankaranarayanan
    Pravin Kumar Thakur
    Pravin Kumar Thakur
    schedule 3 years ago
    Sold Out!
    20 mins
    Experience Report
    Beginner

    After more than 5 years of supporting the thetrainline.com platform, ThoughtWorks worked with The Trainline teams to transfer knowledge and context  back to the Trainline Teams.

    This methodology was co-created by ThoughtWorks and Trainline as a healthy sustainable and mature way to transfer knowledge. The transition itself was about a year long and involved multiple agile concepts around remote pairing, program MVP and above all, continuous delivery and non-disruption to business through the process.

    This presentation would take the audience through the experiences and learnings of the process. This session is co-presented by ThoughtWorks and Trainline (vendor and customer) and will provide an insight across multiple spectrums of delivery and business.

  • Liked Ellen Grove
    keyboard_arrow_down

    Build Your Dreams: User Requirements Gathering with LEGO Serious Play

    Ellen Grove
    Ellen Grove
    schedule 3 years ago
    Sold Out!
    90 mins
    Workshop
    Beginner

    Let your hands be the search engine for your brain! LEGO® Serious Play® is a powerful thinking, communicating and problem solving technique that can help you and your team do serious work through structured play activities using a popular and playful 3D modeling toy. Through a facilitated process of building models that, storytelling and reflection, every person at the table is engaged and actively participating in the discussion, whether the topic is individual aspirations, team relationships, developing a new product or solving a wicked organizational problem. Everyone builds and everyone tells their story – all participants have equal opportunity to put their own points of view on the table, unlocking new perspectives and exposing the answers that are already in the room.  LEGO Serious Play has been used successfully for team-building and problem solving in a variety of organizations, from NASA to RBC to academic settings and public utilities.  

    This presentation provides a hands-on introduction to LEGO Serious Play, so that you can experience firsthand how using LEGO to do real work unleashes creativity and enables meaningful conversations in a very short time. We will explore how to use this playful technique to collaboratively elicit information about user requirements and strategic design issues using the open source User Requirements with Lego methodology developed by a team at the University of Lugano, Switzerland.  This approach is particularly suited to Agile teams that want to get team members and stakeholders sharing their different perspectives on common goals in an open and light-weight manner.

  • Liked Nitin Ramrakhyani
    keyboard_arrow_down

    Lean Roots to Grow, Wings to fly!

    Nitin Ramrakhyani
    Nitin Ramrakhyani
    schedule 3 years ago
    Sold Out!
    45 mins
    Talk
    Beginner

    A lot has been said about Kanban and how these can be implemented in Software development, but the learning remains superficial till we go deep down to its roots to understand the core underlying practices and principles and why/how these practices evolved over a period of time. Infact the roots of most of the Agile methods can be traced back to Lean/Toyota Production Systems, a set of practices and techniques used by Toyota to build great set of cars with limited amount of resources. Even though building software is much different than building a car, there are many lessons and practices that can be learnt and applied nonetheless.

    In this interactive and visual talk, we'll take a virtual trip to Japan and learn some of the best practices/concepts that originated at Toyota for building "world-class" cars and see how each of these can be applied to software development. Learning about the roots of Lean should help the attendees in sowing the seeds of Lean improvement in their organizations and would help in building better software and improving the efficiency of the software delivery lifecycle.

  • Liked Tathagat Varma
    keyboard_arrow_down

    Designing agile feedbacks for agile learning - an experience report

    Tathagat Varma
    Tathagat Varma
    schedule 3 years ago
    Sold Out!
    20 mins
    Experience Report
    Beginner

    Feedback is perhaps the most important aspect of the overall agile lifecycle. If the feedback is too wide and shallow, it won't give enough actionable feedback. If it is too narrow and deep, it might fail to register feedback outside its focus area. So, how does one go about designing feedbacks that enable agile learning. We call them agile feedbacks.

    In this brief session, we will share an experience from designing agile feedbacks for agile trainings and workshops. The objective was to get most critical feedback in shortest amount of time to enable quick action planning. We created feedback that took a maximum of 5 minutes and enabled the most important learning in both, focussed as well as open-ended manner that allowed us to focus on the most critical items. We employed elements of Design Thinking and Rapid Iterative Testing and Evaluation (RITE) to improve the process and quality of feedback themselves. We will also be touching up these concepts and how effective they were.

  • Liked Naresh Jain
    keyboard_arrow_down

    SAMPLE PROPOSAL - Product Discovery Workshop

    Naresh Jain
    Naresh Jain
    schedule 3 years ago
    Sold Out!
    90 mins
    Tutorial
    Beginner

    Many product companies struggle with a big challenge: how to identify a Minimal Viable Product that will let them quickly validate their product hypothesis?

    Teams that share the product vision and agree on priorities for features are able to move faster and more effectively.

    During this workshop, we’ll take a hypothetical product and coach you on how to effectively come up with an evolutionary roadmap for your product.

    This 90 mins workshop teaches you how to collaborate on the vision of the product and create a Product Backlog, a User Story map and a pragmatic Release Plan.

    This is a sample proposal to demonstrate how your proposal can look on this submission system.

  • Liked Tarang Baxi
    keyboard_arrow_down

    1000 Words - Illustrating Project Challenges with Visuals

    Tarang Baxi
    Tarang Baxi
    schedule 3 years ago
    Sold Out!
    90 mins
    Workshop
    Beginner

    A project can face varied challenges through its life, foreseen and otherwise - runaway scope, high defect volumes, depressed velocity, and many more. Addressing many of these first requires recognition of the problem and then action from one or more sets of project stakeholders. Telling the story with simple visuals can be a very powerful way to articulate a challenge (the what), the potential root causes (the why) and the options available to fix it (the now-what). Teams typically already track a lot of data related to throughput, quality, scope and cost. Creative use of this data combined with simple, hand-crafted visuals can be much more effective than hundreds of bullet points. In this hands-on workshop, you get to exercise your visual thinking and visual communication skills. We introduce some simple visual thinking techniques like Look-See-Imagine-Show, and then let you apply them in a project simulation, so that you can practice hand-rolling simple visuals that speak volumes (no fancy tools needed!).

  • Liked Suresh C R
    keyboard_arrow_down

    Using Kanban for synchronization & governance in multi-team Agile programs

    Suresh C R
    Suresh C R
    schedule 3 years ago
    Sold Out!
    45 mins
    Experience Report
    Intermediate

    One of the fundamental reasons why management teams are eager to adopt Agile in large programs these days, is the discomfort that is caused by long gestation periods before anything tangible is delivered to the Business, and the risks associated with it. They hope that the practice of the principle of “Frequent Delivery of valuable working software” that Agile espouses will help them address these risks. However, in the Agile Transformation of such programs, a key challenge that emerges is in synchronising the output of the various teams in the program in order to deliver working software every sprint. This problem gets especially compounded with geographically dispersed teams.

    The following experience report shares the experience in guiding an agile transformation engagement for a program involving geographically dispersed teams of a global BFSI player. The work products went through three stages – creation of the technical workflows required for the desired services by the product owner & SMEs, completion of proofs of concept / technical feasibility of core components of the services, and lastly the actual development work including user interfaces. The development teams were spread across two locations and SME’s were in two different locations. Due to pressures from business, the first priority for the program was to start delivering value by means of usable functionality available to the end users.

     

    Kanban as well as Scrum was used in the transformed process to ensure that smooth and regular flow of functionality was delivered at the end of every sprint. The aspects of Kanban used were Pull based scheduling across the various stages, Making the work Visual and Limiting the number of services which were WIP. How these elements of Kanban drove the Synchronization efforts as well as overall governance betweeen teams, while the teams used Scrum for managing engineering activities within the team, will be detailed in the presentation. How this approach can be extended for better management of MMF’s/ Epics will also be explained.

  • Liked Anand Bagmar
    keyboard_arrow_down

    Build the "right" regression suite using Behavior Driven Testing (BDT)

    Anand Bagmar
    Anand Bagmar
    schedule 3 years ago
    Sold Out!
    90 mins
    Workshop
    Intermediate

    BDT is a way to identify the correct scenarios to build a good and effective (manual & automation) regression suite that validates the Business Goals. We will learn about how this is different from BDD, and do some hands-on exercises in form of workshops to understand the concept better.

  • Liked Manik Choudhary
    keyboard_arrow_down

    "Design Thinking" - Create Innovative Right Products

    Manik Choudhary
    Manik Choudhary
    schedule 3 years ago
    Sold Out!
    45 mins
    Case Study
    Advanced

    Design Thinking is an approach to solving Design/Wicked problems by understanding users’ needs and developing insights to solve those needs. It helps to address the right questions and to create the “Right products”.

    Design Thinking is an iterative innovative approach. It is not meant as a replacement for the organization’s standard development approaches; rather Design Thinking works inside of agile projects, offering frameworks for collaboration and innovation.

    Design Thinking ultimately helps you to address the right question and to create the right solution for it. It Means creating innovative "Right Product" by combining diverse people, creative space and an iterative approach.

  • Liked Dinesh Sharma
    keyboard_arrow_down

    Thinking Environment - Do you have one?

    Dinesh Sharma
    Dinesh Sharma
    schedule 3 years ago
    Sold Out!
    90 mins
    Talk
    Advanced

    Everything we do depends for its quality on thinking we do first. Our thinking depends on the quality of our attention for each other.

    A Thinking environment is the set of ten conditions under which human being can 'think' for themselves - with rigour, imaginaton, courage and grace. A Thinking Environment is natural, but rare. It has been squeezed out of our lives and organisations by inferior ways of treating each other.

    Thinking environment is based on ten behaviours that generate the finest thinking, and have become known as The Ten Components of a Thinking Environment. These components are

    • Attention,
    • Equality,
    • Ease,
    • Appreciation,
    • Encouragement,
    • Feelings,
    • Information,
    • Diversity,
    • Incisive Questions and
    • Place.

    Each Component is powerful individually, but the presence of all ten working together gives this process its transformative impact.