Release year: 2018

Author: Mik Kersten

Buy this book

Link to my handwritten notes

Review

Revolutionary. The author gives a true recipe for connecting IT with the business in any organization.

Félix rating:
👍👍

⭐ Star quotes

  • (p. 40) Investments made in areas other than the bottleneck are futile.
  • (p. 49) In the codebase, every change to every line of code should have traceability back to its originating feature or defect. Adding traceability after the fact is difficult and error prone.
  • (p. 61) IT feels like a black box in many organizations due to the mismatch of the data models that IT vs the business use to operate.
  • (p. 61) Watermelon projects: When engineering leads are asked by project managers if they are on track, the answer is “yes” because the question is ambiguous. Projects looked green from the outside, but were red inside, like a watermelon.
  • (p. 75) It is impossible for the business to have a shared understanding of a bottleneck without having a shared understanding of productivity.
  • (p. 78) The four types of flow items:
    • Features
    • Defects
    • Risks
    • Debts
  • (p. 80) The Value Stream Network itself needs to be treated as a product, with its own stable delivery team, and not as a project with a defined end.
  • (p. 97) The five flow metrics:
    • Flow Distribution
    • Flow Velocity
    • Flow Time (includes both active and wait time)
    • Flow Load
    • Flow Efficiency
  • (p. 97) The zero-sum game of flow distribution trade-offs forces the business to make the kinds of trade-offs that the development team lead has to make constantly as unplanned work enters the value stream.
  • (p. 101) Flow value metrics are more suited to track productivity and delivery trends within a value stream than across value streams.
  • (p. 104) While separating flow time from lead time is important for products with large backlogs of customer requests, for products with smaller backlogs the two metrics will be more closely converged.
  • (p. 107) Flow load is the leading indicator of the point at which taking on too many flow items in parallel reduces output.
  • (p. 114) Business result metrics:
    • Value: Benefit to the business provided by the value stream
    • Cost: Cost of the value stream to the business
    • Quality: Product quality produced by value stream as perceived by customer
    • Happiness: Engagement of the staff working on the value stream.
  • (p. 122) The goal of the Flow Framework is to ensure that there is at least one metric supporting each of the business results and that each metric is tracked across the org for each value stream.
  • (p. 159) Software productivity declines as software scales, due to disconnects between the architecture and the value stream. This is the force DevOps fights against.
  • (p. 161) ⭐ Disconnected software value streams are the bottlenecks to softweare productivity at scale. The disconnects are caused by the misapplication of the project-management model.
  • (p. 183) The structure of value streams is much more similar to an airline network than a manufacturing line. The airline network is a better metaphor.
  • (p. 184) In network management, bottlenecks are constraints that you simply reroute around. The flow does not need to stop as it does in a linear process. This is what makes them harder to catch.
  • (p. 190) As long as a developer tool like Jira is not connected to a service-desk tool like ServiceNow, the corresponding lack of flow and feedback between those two silos creates an information bottleneck. This is where the connectivity index comes into play, it’s a measure of the degree to which information can flow end to end across value streams.
  • (p. 197) Different value streams may have different tolerances for flow load.
  • (p. 215) The Flow Framework is a framework for managing softeare delivery that is focused on measuring and optimizing the flow of business value through product-oriented software value streams that are correlated to business results.