liveBook · Manning

Lean lens

As explained in Build High Value Software Systems

A perspective for analyzing software delivery in terms of value, flow, pull, quality, and learning.

A delivery slowdown can have many causes. The Lean lens turns a vague complaint into questions about value, flow, pull, quality, and learning. First locate the slow part of the delivery loop: intent to code, code to build and test, build and test to running software, or production feedback back to intent. section 1.2

For a test pipeline that takes 45 minutes, fails intermittently, and needs manual reruns, ask:

  • Value: Which tests prevent defects, and which create false confidence? section 1.2
  • Flow: Are changes tested in small increments or a week-sized batch? section 1.2
  • Pull: Do commits trigger timely tests, or does testing wait until deployment? section 1.2
  • Quality and learning: Are failures addressed and flaky tests investigated, or merely rerun? section 1.2

The answers should change the work, not just label it. Possible moves include parallelizing slow suites, systematically fixing flaky tests, running focused checks on each commit, and making failures block merges rather than deployments. section 1.2

Compare a small change with and without AI

Implement the same small API endpoint or validation rule by hand and with an AI coding assistant. Compare how easily each version can be explained, reviewed, changed, or owned later. Look for assumptions and understanding pushed downstream, even if generation was faster. Use the findings to draft constraints for recurring AI-assisted work. section 3.5 section 3.0

More code or faster production is not necessarily more value. Ask whether work improves delivery or merely adds inventory, rework, or future operating cost—especially when AI makes output cheap to generate. section 2.1 section 3.0

How Build High Value Software Systems covers it

How deep it goes
Develops diagnostic questions and a comparison exercise, then links answers to practical changes rather than prescribing a universal procedure.
Examples
Uses a slow, flaky test pipeline and a handwritten-versus-AI implementation of a small change.
What it assumes
Assumes familiarity with commits, tests, reviews, and small software changes while supplying the Lean questions used to assess them.
What you can do after
Practice/perform a bottleneck diagnosis or implementation comparison, then choose changes that reduce deferred understanding and improve feedback.
How it connects
Links value, flow, pull, quality, and learning to test pipelines, feedback loops, code review, and AI-assisted development.

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.