Course 8 · Measure · 7 lessons

Measure What Happened

Set measurement windows, record defensible results, and read verdicts honestly, keeping task-linked results apart from workspace-wide movement.

Start the course

Lessons

01Define the measurement window before startingSet the measurement window before work starts, so everyone agrees when the baseline was captured and when the result counts as final.
02Choose a metric that matches the taskMatch each task to an honest metric, or declare it completion-proof-only, so every result you record stays defensible.
03Connect published URLs or provider identifiers where requiredBind the exact published page URL or provider post URL to a done task so integration-backed measurement reads the right thing.
04Record manual results safelyRecord a manual result from the source and window named on the task, so the value holds up as evidence and the next plan can learn from it.
05Read integration-backed resultsTell whether an integration-backed result is pending, final, or failed, and know what to do in each state.
06Separate task-linked metrics from workspace metricsKeep task-linked results apart from workspace-wide metrics, so you never claim the business moved because a single task worked.
07Interpret met, partial, missed, and inconclusive resultsRead the verdict on a final result (exceeded, met, partial, missed, or inconclusive) honestly and decide what it implies for the next plan.

Who it is for

Operators running plan tasks who need to show what the work achieved (to themselves, a client, or a teammate) without overclaiming.

Course checkpoint

Before you move on, make sure you can answer these questions out loud, to a client if needed. What does the measurement window decide, and why must it be set before the work starts? Which tasks get a metric, which are completion-proof-only, and what is the proof for each? What does an integration-backed result need bound to it before it can be measured automatically? What three things must every recorded result name? What does the app show while an integration-backed result is pending, and what does it show when a provider read fails? Given two numbers, which is task-linked, which is workspace-wide, and what does each one prove? What are the five verdicts, and why is a completed task with no result proof never "met"? If you can explain what a result proves and what it does not prove, without reaching for "caused" or "drove", you are ready.

Try it in Signal & Science

Open the "Log a task outcome" panel on the Marketing Plan, open one completed task, save its result with "Add Result", and confirm the task card shows the recorded value.

Open in Signal & Science

Next course

Make the System Learn

Put it to work

Know what to do next.Know whether it worked.

Signal & Science keeps the evidence, the plan, and the results in one place, so each lesson here becomes a working habit in your own workspace.