Outcome Measurement
Compare AI observations from before and after each deployment inside a bounded time window to attribute visibility change to the intervention.
Private beta — available to agency partners
This capability is live and working for agencies in GeoScript's private beta program. Learn about agency access →
What Outcome Measurement does
Outcome Measurement is GeoScript's before-and-after attribution system. After a deployment is applied, GeoScript compares AI visibility observations from the pre-deployment period to observations from the post-deployment period within a bounded, half-open time window. The comparison produces a structured measurement record that attributes change — or the absence of change — to the specific deployment. This is the evidence base that reports and Ask GeoScript use to answer "did this work?"
How it works
Bounded, half-open time window
The measurement window is defined at the moment a deployment is applied. Pre-deployment observations are those recorded before the deployment event, within a configured lookback period. Post-deployment observations accumulate in a half-open window — it starts at the deployment timestamp and extends forward as new monitoring cycles complete. This design means measurement improves over time as more post-deployment data arrives.
Attribution to the deployment
Each measurement record is linked to a specific deployment. The measurement pipeline does not compare arbitrary periods — it compares the observation history for the exact client and engine set that the deployment targeted. This scoping prevents spurious cross-deployment comparisons.
Current depth limitation
The measurement pipeline is WORKING — the infrastructure for capturing, comparing, and storing before/after records is real. However, the underlying observations are limited by the current monitoring coverage: three AI engines are live, and observation depth depends on how long monitoring has been active. Evidence depth is therefore thin for recently activated clients. GeoScript reports this honestly — sparse data is not hidden, and the measurement record reflects what was actually observed.
What Outcome Measurement does not do
Clear boundaries increase trust. These limitations are documented explicitly, not buried in footnotes.
Outcome Measurement does not claim perfect causal attribution — the measurement records before-and-after differences; confounding factors (engine updates, seasonal query behavior) are real and are not controlled for.
GeoScript does not guarantee that a deployment will produce a measurable positive change in AI visibility — measurement records what happened, not what was intended.
The measurement window does not re-observe historical data retroactively — only observations collected by GeoScript's monitoring pipeline are included.
Outcome Measurement does not measure search engine rankings, paid advertising, or social media visibility — only AI engine observations.
Evidence depth is currently limited by the 3 live engine integrations — Gemini, Google AI Overview, and Copilot observations are not yet available.
Related
Other features
AI Visibility Monitoring
Private beta · 3 engines liveRun a client-specific prompt universe against multiple AI engines on a schedule, preserving each raw response as an immutable observation.
Automated Deployment
Private betaRisk-aware approval, signed packaging, preflight, activation, independent verification, and rollback for every deployed intervention.
Reports
Private betaImmutable, period-scoped snapshots summarizing visibility change, deployments, and outcomes — with a shareable stable-token URL.
Ask GeoScript
Private betaA natural-language interface that answers questions grounded strictly in a client's own observations, snapshots, and reports — never speculation.
GeoScript is in private beta
Agency partners get access to the full capability set — acquisition, monitoring, website intelligence, optimization, deployment, and reporting — for $299/month per active client.