Cloud Architecture Guide
DAY_02 / SECTION_03 // DEVOPS
MODULE READY

Value stream & lead time

The value stream is the sequence of activities to deliver on a customer request — the dual flow of information and material. Manufacturing taught us how to measure it. The same metrics transfer cleanly to technology work.

STREAM_01
IDLE
precision_manufacturing

Manufacturing

Where the metric was born
Order received → materials released. Visible, easy to observe. Levers: small batches · low WIP · prevent rework · constantly optimize for global goals.
STREAM_02
ONLINE
terminal

Technology

Same shape, different glossary
Convert a business hypothesis into a technology-enabled service. Less visible — work happens in commits, tickets, branches. Same levers, different vocabulary.
// lead time vs process time

Lead Time is wall-clock — Ticket Created → Work Completed. Process Time is hands-on — Work Started → Work Completed. The gap is queue time, where most of the waste lives.

Lead time vs process time
// from The DevOps Handbook, Gene Kim et al. — Figure 1.1

// ask: in your job, which is bigger — lead time or process time? almost always lead time. the improvement target is the gap.

// the common scenario · months of lead time
Tech value stream with 3-month lead time
// from The DevOps Handbook, after Damon Edwards "DevOps Kaizen" — Figure 1.2

What you typically see when deployment lead times take months:

  • — Nothing works at the end when you merge everyone's changes
  • — Code that no longer builds correctly or passes tests
  • — Days/weeks to figure out who broke what and how
  • — Poor customer outcomes — features ship late or never

// this isn't a personnel problem. it's a system-design problem.

// the devops ideal · minutes of lead time
Tech value stream with minutes lead time
// from The DevOps Handbook — Figure 1.3 · total cycle 25m
enables
  • — Modular architecture
  • — Loosely coupled services
  • — Well-encapsulated components
  • — Automated test + deploy pipelines
breaks
  • — Shared databases everyone writes to
  • — Monoliths deployed atomically
  • — Manual approval gates
  • — Long-lived branches
// flow metrics · five together tell the story
flow velocity

Items completed per period — is value delivery accelerating?

flow efficiency

Active work ÷ total elapsed — where are the queues?

flow time

Stakeholder pulls a unit of value through — is time-to-value shrinking?

flow load

Active + waiting items — demand vs capacity?

flow distribution

Mix of work types (features / defects / risks / debt) — adjusted to current need?

And one more — %C/A · Complete and Accurate %: at each step, what % of the time does the next step receive work that's "usable as-is"? No need to correct, add, or clarify. High %C/A is a leading indicator of low rework.

help Knowledge Check
Question 1/1

A team has fast process time per ticket but lead time is 3 weeks. What does this tell you?

// pick one to verify