Set up outcome tracking

A Play can define what success means. Each Run then tracks the evidence for that particular batch. Most starter Plays produce useful work without measuring its effect.

Agree the measurement first

Ask Claude to adapt Measured shortlist, or add measurement while creating a Play. Agree the measurement before anyone acts:

SuccessWhat will count?

ComparisonWhat will you compare?

Window + sampleWhen is there enough evidence?

EvidenceWho records it, and where?

Measured shortlist compares a ranked queue with a random comparison queue from the same pool, with results compared for each seller. It starts as a draft: the observation window and minimum sample must be completed before activation. The queues have different rules; follow the Play's instructions about which accounts get research, drafts and outreach.

Use the result, then track it

Activate the reviewed Play, run a batch, and record the actual handoff or action. Open Runs → your Run → Outcome to see progress and, when enough evidence is available, the verdict.

Agree the rules before acting; use a new draft for future batches if they change. Record outcomes as the team works.