Skip to content

The Return / Report Loop

What it is
The closed-loop control system that makes the roadmap living — direction set, results fed back, gaps closed on a cadence
Why it exists
A plan without a return loop is a wish. This is the machinery that turns targets into decisions.

The loop in one picture

DIRECTION            RETURN              REPORT               DECISION
(Targets §3)   →    (Actuals §4)   →    (MBR §5)       →     (Decision log)
set the plan        book the result     diff & explain       change direction
    ↑                                                            │
    └──────────────────  targets/beliefs updated  ───────────────┘
Every entity roadmap runs this loop. The enterprise scoreboard rolls the five entity north-stars into one pane; the decision log is the memory of every direction change.

The seven steps (how a number becomes a decision)

  1. Set the target in the entity's §3 — with a LEAD/LAG tag, an owner, and RAG thresholds (glossary).
  2. Capture the actual in §4 as it lands — from QB → warehouse → this table. No hand-typed financials.
  3. Compute variance (actual − plan, and % of plan) and assign RAG (🟢/🟡/🔴).
  4. Interpret — the so-what: the root cause of the variance, tied to the leading indicator that predicted it. Not a restatement of the number.
  5. Decide — the now-what: a specific action, owner, due date. A report that produces no decision was theater.
  6. Log the decision in the decision log; if it changes a belief it edits §0/§1/§3, and git records why.
  7. Validate next cycle — the MBR opens by reviewing last cycle's actions before any new business.

The cadence stack — match the meeting to the question

Closed loops fail when everything is monthly. Each rhythm answers a different question → see Cadence.

Rhythm Question Artifact
Weekly On track this week? What's blocked? Scorecard + issues
Monthly (MBR) Plan vs actual — why, and what do we change? Monthly report
Quarterly (QBR) Are the targets right? Re-forecast, re-allocate. Quarterly review
Annual Is the direction still right? V2MOM + north-star reset

RAG discipline

Thresholds are set once in the glossary and do not move. Color is a trigger for a conversation: every 🔴 gets a named owner and a now-what; every 🟡 gets a note in the MBR.

Leading vs lagging

Lag tells you if you won; lead tells you if you're about to. You can only act on leads. Every LAG target on a scoreboard should have a LEAD measure beside it — e.g. rental dollar utilization (LAG) is driven by fleet-ready turnaround hours (LEAD). The glossary tags every metric.

Where numbers come from (the wiring)

QuickBooks (per entity, on SEQ-QB) → GL bridge → PostgreSQL warehouse (core.gl_fact / mart.v_pl) → this loop. Build-count, job-cost, AR, inventory and lead metrics wire now; revenue/margin are gated on the GL-detail pipeline (#55, core.gl_fact). Until a metric's pipeline is live its cell reads — with the source cited.

Pages in this section

  • Scoreboard — the consolidated plan-vs-actual pane (all five entities).
  • Cadence — weekly / monthly / quarterly / annual, who + when + agenda.
  • MBR template — the canonical monthly report managers copy.
  • Metrics glossary — every KPI defined once.