Choose a metric that matches the task

Measuring the wrong thing creates false confidence. A social post measured by "website conversions" will look like a failure; a partner-outreach task measured by "pageviews" is measuring the wrong activity entirely. Some tasks should never carry a metric at all: relational, engagement, research, and setup tasks get completion proof only. Choosing the metric honestly is what keeps every result defensible later.

When you review a task's contract before starting it, and whenever a task card seems to demand a number for work that cannot honestly produce one.

• Know the task type: publishing/content (blog post, social post, ad), outreach/relational (partner emails, intros), research (competitor read, audience scan), or setup (install, connect, configure).
• Know the difference between distribution evidence (reach, views, impressions: the work was seen) and business impact (leads, revenue: the work changed something).

1. 1Read "How we'll know it worked" on the task card and ask: is this task's success a number, or a fact? If it is a number, that number is the metric candidate.
2. 2For publishing tasks, the metric candidate is provider-native: pageviews on the exact published page (GA4), or engagement on the exact post (the social provider's own metrics).
3. 3For relational, engagement, research, and setup tasks, declare the task completion-proof-only. The proof is the work itself (sent emails, a delivered scan, a connected integration), not a metric. Do not invent one.
4. 4Check the metric against the TARGETS section on the Plan page ("No target yet. Set the KPI this plan is driving toward." if empty). The task metric should serve the plan's target, not compete with it.
5. 5Refuse vanity metrics: views alone are distribution, not impact. If the only available number is reach, name it as distribution evidence and say so. Never present it as business impact.
6. 6If the contract names a metric that does not fit, say so before executing (raise it with "Ask about this task", or skip the task with a reason and add one with the right contract) rather than measuring the wrong thing and calling it a result.

Measurable tasks show a measurement contract on the card. Completion-only tasks show no result affordance. That is correct, not a bug. The Plan page's TARGETS section shows the KPI the plan is driving toward, with a "Set a target" button when no target exists yet.