If you work in an Agile environment and want to learn how to use data and estimates in practical ways to make achievable plans in multiple-team situations, then this workshop is for you.
This one-day workshop teaches practical data probabilistic forecasting techniques (sounds complex; it isn’t) applied to any flavor of Agile software development. The methods practiced will outperform current popular methods used in Agile that fail due to an inability to cope with highly uncertain work items and complex system factors (dependencies and queues that cause low process efficiency).
All techniques are explained by solving real-world problems using exercises based on real projects and often from the attendees themselves. You will learn how to use concrete tools and knowledge that allow software and IT projects to be rapidly forecast using historical data when available. Also, how to use range estimates when there is no relevant historical data.
The Agenda at a glance:
- How much data do we need to forecast?
- Forecasting 101 - thinking models about prediction
- Monte Carlo Forecasting
- Estimating the initial size and scope growth
- Determining the delivery pace for new and existing Teams
- Prioritization - ordering for value delivery
- Risk Management and dealing with system factors
Who this course is for?
Forecasting using data will benefit people in the following roles:
- Product Management
- Product Owners
- Scrum masters
- Agile coaches
More generally, it is perfect for:
- Anyone who currently uses (or teaches) estimates to forecast how long or how big teams need to be to deliver work (portfolio or product management, product owners, scrum masters, coaches).
Learning outcomes
- To learn the pros and cons of different forecasting techniques.
- To learn how to use data to forecast and how much data is needed.
- To learn how to perform Monte Carlo forecasting and why.
- To learn how to estimate work size and delivery pace ranges quickly.
- To learn what system factors play a role in forecasting.
Problems and examples discussed and solved during the workshop include –
- How to forecast what will fit into a fixed delivery time (e.g., sprint, release, quarter)
- How to predict the flow rate through a development and delivery process
- How to forecast how long a project will take using range estimates or historical data
- How to manage dependencies between teams to reliably predict
- How to estimate the remaining defect counts and ways to assess release readiness
- How to determine staff risk impact due to skills or lack of availability
- How to integrate risk management with feature and project forecasts
Detailed Agenda
- Prediction intervals - physical demonstrations of the concept
- Prediction intervals - dice exercise
- Debrief Prediction Intervals
- "When to leave home forecasting exercise"
- Debrief: Good practices we can learn from Google maps
- Morning tea
- Slides: Monte Carlo forecasting basic theory
- Basic Monte Carlo Forecasting Burndown Exercise
- Monte Carlo debrief and take-aways
- Reference Class Forecasting Exercise
- Story Count Monte Carlo Forecast Exercise
- Scope Growth Canvas
- Size estimate debrief
- Lunch
- Slides: Introduce prioritization "zones" and "tie-breakers"
- ER Triage Prioritization Exercise
- Slides on pace
- Exercise: Ramp-up and down factors
- Debrief - gallery walk of the factors and mitigations
- Delivery Pace Canvas
- Amdahl's Law
- Credit card replacement Exercise Introduction
- Most important insight? What is still unclear?
- Break - Afternoon Tea
- Slides: Blockers and clustering
- Exercise: Blocker clustering
- Gallery walk or shout outs: Blockers
- Slides and Demo: How risk factors impact forecasting
- Exercise: Brainstorm risks and categorize them based on impact
- Gallery walk on blockers and risks
- Extra topic and more detail needed
- Takeaways and next steps