Platform
Demand forecasting built for how your planners actually work
Routeharvest goes below the category level. Every SKU. Every location. Every week. Delivered to the planner's inbox before the reorder window closes.
What the platform does
Per-SKU ML models
One model per SKU-location pair. Seasonality, promotions, and lead-time variance baked in at the unit level, not smoothed out at the category average.
Six-week alerts
Planner inbox shows: "SKU-042 at Cincinnati DC hits zero June 19. Reorder by June 3." Not a dashboard to check: a directive that arrives on its own.
Same-day integration
Reads your ERP transaction history and WMS location data via direct connector. No CSV exports, no data engineering team, no three-month onboarding.
Dead inventory surface
Surfaces slow-moving SKUs by location before they tie up cash for a full quarter. Overstock patterns flagged alongside stockout risks in the same weekly plan.
Forecasting Engine
How the per-SKU model is built
For each SKU-location pair in your catalog, Routeharvest trains an independent time-series model on that pair's transaction history. The model accounts for seasonal demand curves, promotion lift windows, supplier lead-time variance, and regional distribution patterns.
The result is a forecast with a probability-weighted confidence interval, not a single-point estimate. Planner alerts include the high-risk scenario, not the median, so reorder decisions have buffer built in.
- Trained on your actual transaction history, not industry benchmarks
- Confidence bands widen in high-variance conditions (promotions, season breaks)
- Model retrains weekly on fresh ERP data
- Minimum 12 weeks of history required for a new SKU to get its own model
SKU Intelligence
The numbers that matter at the picking floor
Reporting
Forecasts that arrive without being asked for
Most forecasting tools require a planner to open a dashboard, navigate to a SKU, and manually assess risk. Routeharvest flips that model. The weekly plan is assembled automatically and pushed to the planner as a prioritized action list.
Urgent reorder windows appear at the top. Dead inventory accumulating in a regional DC appears mid-list with cost impact. The planner reviews, approves, and gets on with the rest of Monday.
- Weekly plan email with SKU-level action items
- Slack alert when a stockout window shortens unexpectedly
- CSV export for ERP reorder entry
- API endpoint for custom downstream workflows (Growth plan and above)
API Access
Pull forecast data into your own systems
Read-access API included in Growth and Scale plans. Full read/write on Scale. RESTful JSON endpoints with API key authentication.
GET /v1/forecasts?sku=SKU-042&location=cincinnati-dc&horizon=42
Authorization: Bearer rhvt_live_xxxx...
{
"sku_id": "SKU-042",
"location": "cincinnati-dc",
"horizon_days": 42,
"stockout_date": "2026-06-19",
"reorder_by": "2026-06-03",
"confidence": 0.87,
"units_on_hand": 214,
"forecast_p50": 31.4,
"forecast_p90": 44.1
}Ready to see your first six-week forecast?
Connect your ERP and get a live stockout forecast within 48 hours. No long onboarding, no dedicated data team required.
Get Early Access