liveBook · Manning

Control chart

As explained in Build High Value Software Systems

A statistical process-control chart that plots process data over time against a center line and control limits to distinguish common-cause variation from unusual variation.

A control chart plots a process metric over time against boundaries for its normal variation. For software delivery, it helps distinguish routine differences between work items from unusually slow or fast items worth investigating. That matters when one delayed ticket might otherwise distort a team’s view of its typical flow. section 9.2.1

Three lines guide the reading:

  • The center line shows where the process normally runs, using an average or median. section 9.2.1
  • The upper control limit marks the high end of normal variation. An item above it is a signal to investigate, not a reason to blame its owner. section 9.2.1
  • The lower control limit marks the low end. An unusually fast item may reveal a useful pattern to repeat. section 9.2.1

Find the wait, not just the outlier

If a cycle-time point breaches the upper limit, break that item’s time into phases. An eight-day wait for review points to a different problem than eight days spent coding. Compare lead-time and cycle-time charts too: volatile lead time alongside stable cycle time suggests the waiting occurs before work starts. section 9.2.2

Figure 9.4 A lead time control chart. Ticket 9 breaches the upper control limit—that's a signal worth investigating, not a number to punish. The remaining tickets fall within the natural variation of the process.

The lead-time chart shows one ticket above the upper control limit while the others remain within normal variation. section 9.2.1

Use a breach as a question: was the item blocked by a dependency, delayed in review, or poorly scoped? Avoid changing the process in response to every point inside the limits. The chart identifies where to look; investigating the work reveals what to change. section 9.2.1

How Build High Value Software Systems covers it

How deep it goes
Defines three chart lines and interprets breaches, then applies the approach to delivery metrics without specifying how to calculate control limits.
Examples
Illustrates a ticket above the upper limit and an outlier cycle time dominated by an eight-day review wait.
What it assumes
Introduces chart lines from scratch while assuming work-item durations can be obtained from issue-tracker timestamps.
What you can do after
Diagnose/debug with it by selecting unusual work items for investigation and separating execution delays from pre-start queues.
How it connects
Connects lead time, cycle time, queue time, and review waits to common-cause and special-cause variation.

Also explained in

Compare how 2 books explain control chart →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build High Value Software Systems. The section links point to the parts of the book it draws on.