---
name: performance-growth-operator
description: Analyze performance across ecommerce, lead generation and SaaS using explicit business economics, attribution evidence, forecast scenarios and creative briefs. Use for monthly growth reviews, profitability diagnosis, budget scenarios or studio planning. Missing economic inputs limit dependent conclusions, not independent analysis.
metadata:
  version: "1.1.0"
---

# Performance Growth Operator

Provide a complete performance review: business economics, channel and funnel diagnosis,
scenarios, action priorities, creative requirements and a verified report. This is reusable
expertise; the active Workspace supplies the reporting period, targets, formulas and costs.

## OLLIE Runtime integration

Read live Workspace Context and tenant bindings for client facts, resources and permissions.
Consume ask-ollie-authority-core for authority, outloop-access-fallback for access and recovery,
email-lifecycle-core for recipients and communication, and the bootstrapper Runtime Contract
for ownership, scheduling, completion and reflection. This skill supplies business methods,
not a second operator. Select relevant methods, combine other skills and continue independent
work when optional inputs are absent. Never infer access from a named service.
Verify findings against source evidence and read back the requested artifact or authorized change.

## Choose the business model and measurement basis

Use the requested period and workspace timezone; otherwise label the previous completed
calendar month as the working period. Distinguish actual, attributed and recognized revenue.
Compare the current closed period with available prior and seasonal windows, noting attribution
lag and changes in source coverage. Do not silently combine overlapping channel conversions.

Classify ecommerce, lead generation, SaaS or a hybrid from the actual revenue model. Use
[the ecommerce method](references/ecommerce-model.md), [lead-generation method](references/lead-generation-model.md)
or [SaaS method](references/saas-model.md) for the relevant part of the business.

## Resolve economics with evidence

For every input record value, unit, period, source and whether it is observed, derived or
estimated. Use approved client facts first, then derivations from that client's data. An
estimate is a task-scoped scenario input, never an agency-wide approved constant. There are
no built-in margins, close rates, churn rates, currency conversions or zero-cost assumptions.

Run scripts/profitability_router.py with an explicit JSON input using --economics. It returns
missing inputs instead of inventing them. This limits only the dependent calculation: continue
channel analysis, identify the source needed and present labeled scenarios where appropriate.
Read [financial input handling](references/financial-fallbacks.md) before interpreting results.

## Diagnose and plan

Connect delivery, spend, qualified demand, revenue and contribution. Separate tracking failure
from poor performance and a creative issue from an offer or conversion-path problem.
Read [data interpretation](references/data-source-doctrine.md) and [action planning](references/action-engine.md).
Propose changes with evidence, expected mechanism, measurement and reversal; Runtime decides
which requested actions are authorized. Do not create schedules or duplicate task ownership.

## Forecast and creative work

Use scripts/forecast_scenarios.py with explicit spend and response assumptions for each scenario.
Reconcile the baseline with the economics calculation. Show sensitivity and label projections;
no scenario becomes a guaranteed outcome. Read [forecasting](references/forecasting.md).

Inspect relevant supplied creative and real product references. Use
[creative diagnosis](references/creative-intelligence.md), [video evidence](references/meta-video-intelligence.md)
and [studio briefing](references/studio-brief.md). Produce the requested briefs and assets through
approved capabilities, preserving authoritative identity and exact supplied copy.

## Complete the reporting and handoff

Read [reporting](references/client-vs-internal-reporting.md), [artifact handling](references/brief-artifacts-and-drive.md)
and [studio handoff](references/client-studio-direct-loop.md). Internal and client reports must
agree on material facts and uncertainty. Never conceal estimates to make a client report look
more certain. Use the active host's document skill when relevant, then inspect exported output.
Runtime owns any authorized delivery and its stored-copy verification.

## Local tools

The economics and forecast scripts read one JSON input and return JSON to stdout. They have
no network, credentials or mutation behavior. The HTML validator checks an existing HTML report;
it does not send it. Source safety scans are supporting checks, never complete privacy proof.