Software Project Estimation Techniques

Presentation from the Daycamp4Developers virtual conference talking about how to estimate for software projects, and how to keep track of how those estimations are working out.

Transcript

1. Software Project Estimations Lorna Mitchell, Day Camp for Developers 6
2. How Late??
3. Is Estimating Important?
4. Is Estimating Important? • One of the two most common causes of runaway projects is unstable requirements
5. Is Estimating Important? • One of the two most common causes of runaway projects is unstable requirements • One of the two most common causes of runaway projects is optimistic estimation (from IEEE: http://lrnja.net/10DRtB4)
6. Is Requirements Gathering Important?
7. Only if you want to build something your client actually wants
8. Project Requirements Requirements are not just about the features. Consider: • resource and/or cost • deadlines • platform constraints
9. Estimations Recipe for good estimations: • Clear, stable requirements • Feedback loop • Ability to change the plan
10. User Stories Capture a single action requirement and it's motives Template: As a <role>, I want <goal/desire> so that <benefit>
11. Example User Stories • As a user, I want to see details of a talk so I can find out about it • As an admin, I want to be able to edit details of a talk so I can make changes • As a speaker I want to be able to claim a talk so it will be listed on my account • As a user, I want to rate and comment on a talk so that I can give feedback to the speaker
12. Ground Rules of Estimation
13. Ground Rules of Estimation • Only a person capable of completing a task can estimate it • If a task is estimated at more than half a day, the task should be split • If a task cannot be estimated, steps should be taken to reduce the unknowns
14. Estimation Techniques
15. Traditional Method
16. Planning Poker Uses the wisdom of crowds to get an accurate estimate.
19. Planning Poker Uses the wisdom of crowds to get an accurate estimate. You will need: • 3+ players and a facilitator • Detailed list of tasks • Deck of Planning Poker cards, or equivalent
20. How To Play Planning Poker • Take each task in turn • Facilitator describes the task • At a signal, players estimate • Estimates are close, use average • If estimates differ, task is too vague • Quickly redefine and re-estimate
21. How To Run A Meeting • If you are leading it, then LEAD • Prepare an agenda • Observe timings • Curtail other discussion
22. What's Missing? Commonly forgotten tasks include: • Data migration • Copy management • Client demo time • Server setup • Handover documentation
23. The Flaw In The Plan: Reality
24. Guide motto: Be Prepared
25. Risk Risks deserve our recognition and respect Try to mitigate risks: • Identify them • Single points of failure: look for alternatives/redundancy • Unknowns: "spike" to reduce the unknown aspect
26. Tracking Estimates
27. Tracking Estimates Each task has an estimate • After working on a task, re-estimate how much work is left in that task • When it's done, estimate goes to zero • Track these per-day to get burndown chart
28. Example Task List Task Owner Mon Tue Wed Thu Fri Buy drinks Kevin 3 3 3 3 0 Make cake Lorna 2 0 0 0 0 Make sandwiches Lorna 6 6 4 3 0 Set table Lorna 2 2 2 2 0 Decorate Cake Lorna 7 7 7 5 2
29. Example Burndown Chart
30. Project Velocity: a freakishly reliable figure showing the disparity between estimated and actual work rates
31. Measuring Project Velocity As well as updating estimates, track time spent per task. For completed tasks, divide estimate/actual to get velocity. Use as a multiplier to correct future estimations
32. Alternative Velocity Measure Agile teams working in sprints can measure story points per sprint
33. How to Deliver a Project on Time • Make excellent estimates • Work out how much time there is • Schedule in a plan • Measure progress • Adjust as required
34. Estimations
35. Estimations • Estimations are important • Accurate estimations are achievable • We continue to estimate and inform during a project
36. Go Forth and Deliver On Time
37. Thankyou! Any questions? Feedback: https://joind.in/8973 Me: http://lornajane.net