- July 7 UTC 00:00 – 01:30 (San Francisco: July 6 17:00 – 18:30, New York: July 6 20:00 – 21:30)
Audience: Stakeholders, managers, and members of software development/delivery teams.
Software development is an inherently unpredictable activity. Yet, organizations need a practical way to make realistic forecasts for software delivery in order to schedule marketing announcements and product releases, as well as to manage the costs and resources for building and supporting software products. Besides that, organizations need a quantifiable basis for measuring the effects of process improvement efforts.
Metrics that are associated with particular software development/delivery methods tend to depend on those methods being used properly. Earned Value Management depends on a comprehensive project plan, with dependencies and time estimates for all work packages well-defined. Velocity depends on teams using a time-boxed iterative process model, such as Scrum or Extreme Programming, in the proper way – actually delivering solution increments to production on a steady cadence. When these dependencies are not met, the resulting measurements are unreliable. They are attempts to measure things that are not happening. As such, they are often seen as “made up numbers” that tell us nothing about delivery effectiveness, capacity, or true delivery dates.
Metrics associated with the Lean school of thought tend to have no such dependencies. They aim to measure throughput – the number of “value units” delivered per unit of time; lead time – the total elapsed time from concept to cash of a product; cycle time – the mean time to complete a single work item in a process; and process cycle efficiency – the proportion of lead time in which value is added to the product.
These measures do not depend on how the work is done; they only measure elapsed times and value-add times. Therefore, they are useful when we are not applying our chosen process exactly “right,” and when we are changing our processes to try and achieve better outcomes.
In this webinar, we will explore the kinds of information we can glean from cycle time measurements, and how to use that metric to improve both predictability and effectiveness of software delivery, regardless of the methods we are using.