Lessons Learned in Project Metrics - Are Your Metrics Dumb or Smart?
Collecting and reporting effective metrics can be a tricky business. Einstein captured it well when he noted "Not everything that can be counted counts, and not everything that counts can be counted".
Software projects have a history of measuring irrelevant and even counter-productive progress tracking metrics. The "Hawthorne Effect" should teach us that we will influence what we measure, yet companies continue to overtly track things like hours worked and lines of code written, unaware that they send the message of valuing long hours over results, and discourage simplifications and healthy refactoring. Quite often the metrics we want to track are intangible and subjective and so people tend to shy away from them.
More fundamentally, why are we even tracking these metrics? Is it to report on what has already occurred or help steer our future course? Often an imperfect view of the future is more useful than a perfect view of the past. In the real world, rear-view mirrors are much smaller than windshields for good reason, yet the accuracy of hind-sight and our attraction to certainty often creates too much of an emphasis on lagging, already occurred measurements compared to leading metrics. So we get fancy graphs of project spend and defect rates, but no better insights into what we should be doing differently in order to meet our goal.
In this presentation I will review many common project metrics and explain why they are largely misguided and counter productive. An alternative set of "Design Factory" metrics will be presented that are "simple and relevant to the true project goal", these metrics leverage the Hawthorne effect and focus on leading metrics to support smarter decision making.