MerchantFlowMerchantFlow Docs
Dashboard

Anomaly Alerts for E-commerce

Get proactive alerts when your store deviates from baseline - revenue drops, ROAS shifts, sync failures, and credential expirations - without watching dashboards all day.

Anomaly Alerts

Anomaly alerts are proactive notifications surfaced when MerchantFlow detects that your store's metrics or sync state has deviated meaningfully from baseline - a revenue drop, an unexpected ROAS shift, a failing integration, or an expiring credential. They show up in the in-app Notifications panel and the Daily Email Summary so issues catch your eye before they become problems.

Plan availability: Anomaly alerts are included on the Pro and Plus tiers. Starter workspaces do not receive anomaly alerts. See Pricing Plans.

What Triggers an Alert

The anomaly engine sweeps tenant data on a schedule and fires alerts in these categories:

CategoryExample trigger
RevenueDay-over-day or period-over-period drop beyond the typical variance
ROASA connected ad channel's ROAS drops well below your trailing average
Ad spendSudden spike in spend without a matching revenue lift
Profit marginNet margin compression beyond a threshold
Sync healthA sync job has failed repeatedly, or hasn't run within its expected window
CredentialsAn OAuth credential is approaching expiry or has been revoked by the provider
FulfillmentUnfulfilled order backlog above normal levels
Data qualityCOGS or attribution coverage drops below a threshold for active products

Where Alerts Appear

Anomaly alerts surface in three places, so the same finding catches you whether you're in the app or in your inbox:

  1. In-app notifications dropdown in the dashboard header (bell icon) - see Notifications
  2. Daily AI Email Summary in the Operational Alerts section if you have it enabled - see Email Summary
  3. AI Insights panel on the main dashboard - severe alerts surface as warning or negative insight cards - see AI Insights

How Detection Works

The anomaly engine compares each metric against its trailing baseline (a rolling window of the same period type) and flags deviations that fall outside the expected range. Sync-health and credential alerts are deterministic - they fire whenever a job has failed N times in a row or a credential is N days from expiry. Data-quality alerts fire when coverage thresholds drop below the configured floor.

Detection runs in the background on a schedule. There is no per-tenant configuration UI for thresholds in the current product - the engine ships with sensible defaults tuned to e-commerce variance.

Acting on an Alert

Each alert links to the relevant dashboard page where you can investigate:

  • Revenue/profit alerts → main dashboard with the affected timeframe pre-selected
  • ROAS / ad-spend alerts → the Ads page for the affected channel
  • Sync health alertsSync Status for the failing job
  • Credential alertsSettings > Integrations (/dashboard/settings/integrations) to reconnect

Frequently Asked Questions

Can I configure anomaly thresholds?

Not in the current product. The engine ships with defaults appropriate for typical e-commerce variance. If a category produces too many alerts for your store, contact [email protected].

Why am I not getting alerts?

Most likely reasons:

  1. You're on the Starter tier - anomaly alerts require Pro or Plus. See Pricing Plans.
  2. You haven't accumulated enough history yet - the engine needs a baseline to compare against.
  3. Email alerts are gated by your Daily Email Summary preferences and the Operational Alerts section toggle.

Do alerts replace my SaaS observability tools?

No. Anomaly alerts are about your commerce data, not infrastructure. For platform-side incidents, you'll still see them via the live indicator and integration health UI in the dashboard.

Are alerts sent to all team members?

The in-app notifications panel is per-user. The daily email summary goes to whoever the workspace owner has configured (including additional CC recipients).

Related Guides


Last updated: April 30, 2026