liveBook · Manning

Work in progress

As explained in Build High Value Software Systems

Work in progress is the quantity of work that has been started but not yet completed in a process.

A team can be busy merging code, waiting for reviews, and watching deployments while little reaches users. Work in progress (WIP) is the number of items started but not complete. In software, it is often invisible: open branches and pull requests, features in QA, and code that is finished locally but not deployed. section 6.0 section 6.3

WIP is inventory: effort already invested without a completed outcome. As it waits, feedback is delayed and the work can drift from requirements, conflict with other changes, or require rework. Generating more code does not clear a review or deployment queue. A pull approach starts work in response to demand and available capacity rather than pushing more unfinished work downstream. section 6.0 section 1.1.1

Count what is actually in flight

Three open branches, two pull requests under review, and one feature in QA make six WIP items. If an authentication refactor stalls at security review, starting an export feature and then debugging an unrelated search bug keeps adding active work. By Friday, all three may be partly done and none shipped. A limit makes the choice visible: finish an item before taking on another. section 6.3

Listing 2.2 Inventory waste
function checkout(user: User) {
  if (process.env.NEW_CHECKOUT === "true") {
    // v2 path: partially rolled out, still missing edge cases
    return newCheckout(user);
  }
 
  // v1 path: still required for most users
  return oldCheckout(user);
}
 
// months later: both paths still exist, tests and fixes must consider both
 

A temporary feature flag leaves old and new checkout paths in place for months. The unfinished rollout makes every fix and test account for both paths. section 2.4.2

Use a WIP limit as a decision aid, not a target for keeping everyone busy:

  • Count work across the delivery path, including items waiting for review, validation, or deployment. section 6.0
  • When capacity is full, ask what can be completed or unblocked before starting something else. section 6.3
  • Define completion clearly: remove abandoned code paths rather than leaving them behind after requirements change. section 1.1.4

Personal WIP extends beyond tickets. Promised reviews, mentoring, learning, and design work also consume attention, even when a team board omits them. A personal Kanban board, notes, or a simple app can make those commitments visible. They are legitimate work; counting them helps expose overload and protect time to finish. section 14.4

How Build High Value Software Systems covers it

How deep it goes
Defines WIP as a count, works through a limiting scenario, and extends it to personal commitments without prescribing one universal limit.
Examples
Counts branches, pull requests, and QA work; follows a blocked refactor; and shows a feature flag preserving two checkout paths.
What it assumes
Assumes familiarity with branches, pull requests, tickets, reviews, and deployment while defining WIP from scratch.
What you can do after
Practice/perform WIP limiting by counting active items, finishing work before exceeding capacity, and making personal commitments visible.
How it connects
Connects pull systems and capacity to inventory, value streams, and flow.

Also explained in

Compare how 3 books explain work in progress →

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.