Health scores
Hollow Sentinel health dimensions — growth, operations, customer, experience, and overall store health.
Health scores translate store activity into dimension-level grades you can scan in seconds. They use the health-v4 model from Shopify-native data and your rolling baselines — not from custom reports you build.
Dimensions
| Dimension | What it reflects |
|---|---|
| Growth | Revenue and orders vs 90-day baselines when available; order-volume index during learning |
| Operations | Low-stock count vs baseline; rule fallback when baselines are still forming |
| Customer | Dormant customer retention signals |
| Experience | Theme stability, Scale experience probes, and open experience insights |
| Overall | Weighted composite without double-counting inventory |
Learning period (14 days)
For the first 14 days after install, Hollow Sentinel is in learning mode. The app captures one store snapshot per day (scheduled overnight UTC) and uses those snapshots to establish baselines for revenue, inventory, and experience signals.
| Milestone | What happens |
|---|---|
| Install | OAuth completes; Store timeline records webhooks immediately |
| First snapshot | Health scores and setup briefing can populate from live Shopify data |
| Days 1–13 | Confidence shows as learning; intelligence modules may show Learning status |
| 14 daily snapshots | Learning completes — baselines stabilize and anomaly confidence improves |
Experience probes (Scale)
Scale runs a scheduled HTTP availability check against your primary shop domain (weekly). Results appear on the Health page and in Command center when entitled. This is a lightweight uptime signal — not full storefront RUM or broken-link crawling.
Inventory forecast (Scale+)
Scale includes velocity-based days-of-cover forecasting for low-stock SKUs on Command center and Health. Sentinel projects when variants may stock out using recent unit velocity and your default reorder lead time (configurable in Settings). The stockout_forecast insight opens when cover falls below your lead-time threshold.
Baselines and recompute
Sentinel recomputes baselines on a rolling window from your historical snapshots. Scores stabilize after several days of data. Sudden theme publishes may temporarily affect experience scores until the new state becomes the baseline.
Sentinel vs Engine operating score
Hollow Sentinel health scores focus on passive monitoring and anomaly detection. Arcanum Engine operating score weights installed automation coverage and attributed playbook results. Merchants running both products see complementary scores — Sentinel for watch, Engine for act.
Was this page helpful?