Saturday, August 14, 2010

Project Vital Signs

The Thoughtworks Anthology: Essays on Software Technology and Innovation (Pragmatic Programmers)At Agile 2010, Stelios Pantazopoulos gave a presentation on project metrics entitled “Project Vital Signs”. Stelios starts by pointing out that IT has lost trust and credibility with the business due to projects having a poor track record of success, poor visibility and ROI that is hard to quantify. Projects are over budget, late, or fail all together. To restore trust and credibility, we need to show how and why a project is on track throughout the project’s lifecycle.
Stelios, suggests applying the metaphor of medical care and their use of “vital signs” to help form a holistic view of the state of the project. In medicine, doctors go over the patient’s medical history, review lab test results, look at vital signs, and then use their experience to diagnose the condition and recommend a treatment. Stelios defines vital signs as simple, quantitative, near real time metrics conveyed via a chart published to a location and easily referenced by all. There are 5 vital signs that need to be monitored. 4 are based on PMBOKs Scope, Quality, Schedule, and Budget. In addition, Stelios feels that monitoring the team’s overall health is also very important.
The 5 charts are:
  1. Scope Burn Up: Backlog burn up chart (stories in each state vs. time) that tracks schedule and scope to show expected delivery date. Number of stories or story points are be tracked. Stelios mentions that if you have a big enough back log, tracking number of stories worked well for him.
  2. Current State of Delivery: Backlog Scrum board that tracks scope and team and shows real time state of delivery. Tracks the state of each story and who is working on which story.
  3. Budget Burn Down: Burn down chart (total $ vs. time) that tracks schedule and budget to show remaining dollars.
  4. Delivery Quality: Dot Chart (bug category vs. time) that tracks schedule and bugs to show overall quality. Bugs are categorized into 4 categories:
    1. High priority/High severity
    2. High priority/Low severity
    3. Low priority/High severity
    4. Low priority/Low severity 
    At the end of each iteration, each bug is represented as a dot in the corresponding category. Ideally, as the schedule nears release date, most bugs should fall into the low priority categories. Stelios also mentions that alternatively one can track automated test results as opposed to bugs.
  5. Team Dynamics: Chart (state vs. time) that tracks the state of the team and schedule to show the state of team at different stages of delivery. The state is collected in the retrospective by asking team member to have a secret vote on where each feels they are on the Tuckman model of group development (forming, storming, norming, performing). Similar to the delivery quality chart, each team member’s opinion is represented as a dot in the corresponding category.

The 5 project vital signs are combined to create a project dashboard that is placed on the wall for high visibility. Stelios wraps up by going through different sample scenarios and demonstrates how the dashboard provides a holistic view of the project. He then diagnosis the project and performs what if scenarios for different treatments showing the state of the project before and after treatment.

Stelios provided the slide deck for the presentation at