Forecast every SKU,
across every branch
Demand forecasting software for distributors running tens of thousands of SKUs across dozens of locations. It forecasts each one from its own consumption, picks the best method per item, and tells you honestly how sure it is.
| Item | Per Day |
|---|---|
| Amoxicillin 500mgAntibioticssmooth | 44 |
| Surgical Gloves (M)Consumablesintermittent | 2 |
| Insulin Glargine PenCold Chainlumpy | 14 |
| Paracetamol 650mgAnalgesicssmooth | 180 |
| Cotton Roll 500gConsumablessmooth | 68 |
You stock 40,000 SKUs. You forecast 1,000 of them.
Across dozens of branches, the honest answer is that most of the catalog runs on gut feel and last month's number. Three things quietly cost you stock every month.
The long tail is invisible
Your team forecasts the top 100 SKUs in a spreadsheet and eyeballs the other tens of thousands. The slow, intermittent, and lumpy items — most of the catalog — get no forecast at all.
Stockouts lie to your history
A month you were out of stock reads as low demand, so next month's forecast is dragged down — and you under-order the very items that were selling.
One number, no honesty
A single forecast with no confidence is a guess in a suit. You can't tell which numbers to trust and which to double-check before you commit stock.
One engine. The right method for every item.
Every method runs on your history — a rolling average, your own weighted average, and the ARIA statistical and machine-learning models. The one that measurably wins on a walk-forward backtest is the one you see. A smooth, high-volume item and a lumpy, once-a-quarter item shouldn't be forecast the same way — so they aren't.
Smooth → Steady, high-volume demand — the machine-learning model captures the trend and seasonality and wins the backtest.
This is what your team sees
The whole catalog on one screen — sortable by volume so the Pareto head surfaces first. Click any SKU for the full analytics: consumption versus forecast, every model compared, the signals behind the number, and the regional demand split.
| Item | Category | Wt. Average | Rolling Avg | ARIA 1.0 | ARIA 2.0 | ARIA 3.0 | Per Day | Total |
|---|---|---|---|---|---|---|---|---|
| Amoxicillin 500mg CapsuleAMX-500-Csmooth | Antibiotics | 1,280 58% | 1,240 51% | 1,298 64% | 1,305 71% | 1,310 82% | 44 | 3,930 cap |
| Surgical Gloves (Medium)GLV-M-100intermittent | Consumables | 74 39% | 90 33% | 66 58% | 62 66% | 71 47% | 2 | 248 box |
| Insulin Glargine Pen 3mlINS-GLR-3lumpy | Cold Chain | 410 63% | 320 44% | 388 55% | 402 60% | 430 52% | 14 | 820 pen |
| Cotton Roll 500gCOT-500smooth | Consumables | 2,020 60% | 2,050 74% | 2,035 66% | 2,040 68% | 2,075 63% | 68 | 6,150 roll |
| Glucose Test Strips (50s)GTS-50erratic | Diagnostics | 500 41% | 500 38% | 480 57% | 512 52% | 468 49% | 16 | 1,440 pack |
| Paracetamol 650mg TabletPCM-650-Tsmooth | Analgesics | 5,280 61% | 5,100 54% | 5,350 67% | 5,380 72% | 5,400 85% | 180 | 16,200 tab |
Live product interface, shown with sample data. A lane that hasn't run shows “—” — never a fabricated number.
What it reads
to get there
Not just last month's number. ARIA learns from the signals a good planner would look at — and thousands they couldn't track by hand.
- 01Demand historyTrend, momentum, months since last sale
- 02Seasonality & calendarMonth-of-year, same-month-last-year
- 03Item attributesCategory, pack size, price tier, lead time
- 04Stock & coverageDays of cover, stockout flags, min/max
- 05Regional relationshipsPer-branch demand shares & imbalance
Built to forecast catalogs of tens of thousands of SKUs
The engine is global — it learns across your whole catalog and scales to millions of SKUs. In a live deployment it forecasts a full distribution catalog, every month:
Accuracy is a measured number, not a promise. Every forecast is walk-forward backtested with MASE and must beat your current method before it ships — and the confidence bands are validated against what actually happened.
Who it's built for
Questions, answered straight
See ARIA forecast your catalog
Bring your consumption data. We'll show you a forecast for every SKU — and the number that proves it beats your baseline.