Program health

Improvement playbooks

Concrete levers for weak metrics, linked to the pages where you diagnose each signal.

Questions this page answers

How fast and reliably do we ship to production?
Are engineering goals and benchmarks on target?
Which playbooks apply to our weakest metrics?

Time range

Repository

Team

Showing playbooks for 4 off-target goals. Adjust goals · All playbooks

Cycle time phases

Off target: Cycle time

Cycle time

End-to-end time from first commit on a branch to production.

Measured as: Mean of coding + pickup + review + deploy hours per merged PR. Deploy phase is 0 when no production deploy is linked; PRs without deploy still count in the average.

Common causes

One phase dominates (often pickup or review)
PRs idle across weekends or time zones
Deploy waits after merge
Rework loops after review feedback

Levers that help

Quick win
Open Time distribution — fix the largest phase first.
Team habit
Limit WIP: fewer concurrent PRs per author/reviewer.
Team habit
Merge and deploy daily where quality gates allow.

Diagnose in app:Delivery pipeline Where time goes Metric trends

Off target: Pickup time

Pickup time

Wait from PR ready to first meaningful human review.

Measured as: Hours from PR open (or review requested) to first non-bot review comment or approval.

Common causes

No default reviewers or ownership map
Review load concentrated on one person
PR opened without context or too large to start
Timezone gaps without backup reviewers

Levers that help

Quick win
Assign a reviewer at PR open; use team rotation.
Team habit
Cap PR size; split if diff exceeds team threshold.
Quick win
Daily 15-minute review sweep for open PRs in Attention center.

Diagnose in app:Attention center

Off target: Review time

Review time

Time in active review until merge.

Measured as: Hours from first human review to merge.

Common causes

Large diffs requiring multiple review rounds
Unclear acceptance criteria
Authors slow to respond to feedback
Optional reviewers blocking consensus

Levers that help

Team habit
Use a review checklist for risk areas (auth, migrations, infra).
Quick win
Respond to review comments within one business day.
Team habit
Pair on complex changes instead of async-only mega-PRs.

Diagnose in app:Review quality

Review quality

Off target: PR size

PR size

Typical change size by lines touched.

Measured as: Lines added + deleted on merged PRs (bot-excluded activity).

Common causes

Feature + refactor + formatting in one PR
Generated files or lockfiles bundled with logic
Long-lived branches accumulating diffs

Levers that help

Team habit
Split refactors and behavior changes into separate PRs.
Quick win
Use mechanical commits (format, rename) in isolated merges.
Team habit
Set a team soft cap (e.g. 200 lines) and justify exceptions.

Diagnose in app:

Continue investigating

Follow-up pages that deepen the story for this view

Watchtower DevX

Showing playbooks for 4 off-target goals. Adjust goals · All playbooks

Cycle time phases

Off target: Cycle time

Cycle time

End-to-end time from first commit on a branch to production.

Measured as: Mean of coding + pickup + review + deploy hours per merged PR. Deploy phase is 0 when no production deploy is linked; PRs without deploy still count in the average.

Common causes

One phase dominates (often pickup or review)
PRs idle across weekends or time zones
Deploy waits after merge
Rework loops after review feedback

Levers that help

Quick win
Open Time distribution — fix the largest phase first.
Team habit
Limit WIP: fewer concurrent PRs per author/reviewer.
Team habit
Merge and deploy daily where quality gates allow.

Diagnose in app:Delivery pipeline Where time goes Metric trends

Off target: Pickup time

Pickup time

Wait from PR ready to first meaningful human review.

Measured as: Hours from PR open (or review requested) to first non-bot review comment or approval.

Common causes

No default reviewers or ownership map
Review load concentrated on one person
PR opened without context or too large to start
Timezone gaps without backup reviewers

Levers that help

Quick win
Assign a reviewer at PR open; use team rotation.
Team habit
Cap PR size; split if diff exceeds team threshold.
Quick win
Daily 15-minute review sweep for open PRs in Attention center.

Diagnose in app:Attention center

Off target: Review time

Review time

Time in active review until merge.

Measured as: Hours from first human review to merge.

Common causes

Large diffs requiring multiple review rounds
Unclear acceptance criteria
Authors slow to respond to feedback
Optional reviewers blocking consensus

Levers that help

Team habit
Use a review checklist for risk areas (auth, migrations, infra).
Quick win
Respond to review comments within one business day.
Team habit
Pair on complex changes instead of async-only mega-PRs.

Diagnose in app:Review quality

Review quality

Off target: PR size

PR size

Typical change size by lines touched.

Measured as: Lines added + deleted on merged PRs (bot-excluded activity).

Common causes

Feature + refactor + formatting in one PR
Generated files or lockfiles bundled with logic
Long-lived branches accumulating diffs

Levers that help

Team habit
Split refactors and behavior changes into separate PRs.
Quick win
Use mechanical commits (format, rename) in isolated merges.
Team habit
Set a team soft cap (e.g. 200 lines) and justify exceptions.

Diagnose in app: