Product methodology

How Growomo processes marketing decisions.

Read the documented inputs, processing, AI involvement, approval boundaries, failure states, limitations, and demo-data examples behind Growomo.

Authorized provider data
Connected accounts, selected assets, reporting ranges, and available sync state.
Marketing Data Hub
Keeps source, freshness, connection, and processing context attached to the evidence.
AI Decision Engine
Organizes available data into a recommendation with rationale and uncertainty.
Human review
A marketer approves, rejects, or investigates before any supported protected action.
Experiment Engine
Structures a test, preserves measurement context, and records the reviewed outcome.

Reviewed learning returns to future evidence and decision context.

Three documented systems

Inspect every stage of the decision loop.

Inputs: Authorized provider accounts, assets, reporting ranges, and sync state.
Processing: Connector-specific retrieval with source-aware normalization.
Human boundary: Review freshness, source definitions, and incomplete records.
Limitations: Timing, attribution, and eligible assets vary by provider.

Inputs: Synced evidence, dates, account context, and detected conditions.
Processing: Rules and AI-assisted ranking with rationale and uncertainty.
Human boundary: Approve, reject, or leave a recommendation for investigation.
Limitations: Missing context can make a recommendation incomplete or wrong.

Inputs: Hypothesis, variants, metric, duration assumptions, guardrails, and owner.
Processing: Monitor available evidence without manufacturing certainty.
Human boundary: Review quality and record a continue, stop, investigate, or inconclusive decision.
Limitations: Tests can remain inconclusive and provider metrics can arrive late.

Operating principles

Every data point keeps its provider, freshness, retrieval, and attribution context.

Recommendations preserve rationale, evidence context, limitations, and uncertainty.

People approve, reject, or investigate before any supported protected action.

The methodology surfaces insufficient evidence instead of manufacturing certainty.

Human review boundary

Approve

Accept a supported next step and continue through the protected action path.

Reject

Decline the recommendation and retain the review decision with its context.

Investigate

Request more evidence when source quality, business context, or uncertainty requires it.

Questions and boundaries

Is the Data Hub real time?

No universal real-time claim is made. Growomo shows the latest successfully synced provider data and its available sync state.

Does onboarding selection connect a provider?

No. A provider is connected only after its authorization flow and required account selection complete successfully.

Does Growomo automatically change campaigns?

No blanket automation claim is made. Supported management actions require an explicit user action through a protected workflow.

Are recommendations guaranteed to improve results?

No. Recommendations organize evidence and next actions; they do not guarantee rankings, revenue, ROAS, leads, or conversions.

Does Growomo always declare a winner?

No. A test can remain inconclusive when the evidence, duration, tracking, or sample quality is insufficient.

Can Growomo pause or scale variants automatically?

No blanket automatic-action claim is made. Supported provider actions require an explicit approved user action.