Skip to content
stokrio

Forecasting & planning

Forecasts you can trust, explained product by product

Stokrio tests a dozen forecasting models against each product's own sales history and keeps the one that would have predicted it best — then shows you why.

Get startedSee pricing
Product forecast with weekly demand, 80% range, projected stock and the reorder calculation
The real product, shown with demo data.

How demand forecasting works

Stokrio forecasts weekly demand for every product from its Shopify order history. It classifies each product (smooth, erratic, intermittent, lumpy, new), backtests models such as exponential smoothing, damped trend, Holt-Winters, Croston, SBA and TSB on rolling origins, picks the most accurate by MASE/RMSSE, and treats out-of-stock days as missing demand rather than zero sales.

Model competition per SKU
Exponential smoothing, damped trend, Holt-Winters, seasonal profiles and intermittent-demand models (Croston, SBA, TSB) are backtested on each product. Simpler models win ties.
Stockout-aware
Days a product was out of stock are excluded from demand history, so a sell-out doesn't teach the forecast to buy less.
Seasonality and new products
Yearly patterns are detected per product; new products borrow their category's pattern and rate until they have history.
Prediction intervals
Every forecast shows an 80% range, so you can see how certain it is before you buy.
Events and promotions
Add multipliers, fixed quantities or monthly targets for launches and sales — they flow straight into reorder quantities.
Measured accuracy
Backtest accuracy per product plus realised accuracy of past forecasts against what actually sold, with bias.

Frequently asked questions

How much sales history does Stokrio need?
Forecasts start with a few weeks of sales. Seasonal patterns need about a year; until then a product uses its category's seasonal profile.
Does a stockout make the forecast too low?
No. Days a product was out of stock are treated as missing demand, not zero sales, so forecasts aren't dragged down by sell-outs.
Can I override the forecast?
Yes. Add promotions, launches or monthly targets as adjustments for a product, category or the whole catalogue. Overrides are applied to replenishment immediately.
Which models does Stokrio use?
Naive and seasonal naive benchmarks, moving average, simple exponential smoothing, damped trend, Holt-Winters (additive and multiplicative), seasonal profiles, Croston, SBA and TSB. The best model per product is chosen by backtest.

Works with

Replenishment & reorder planning

Know what to reorder, when and how much. Statistical safety stock, real supplier lead times, MOQ and case-pack rounding, order-by dates and budget-aware purchasing for Shopify brands.

Learn more

Demand & supply plan

A monthly demand plan with forecast, plan and last year side by side, planner overrides and last-year-plus-growth, plus a supply plan with recommended receipts and order-by dates.

Learn more

Alerts & weekly digest

Stockout, reorder, late purchase order and sales spike alerts checked every night against the forecast, with a morning email and a Monday digest of sales, buying and deliveries.

Learn more

Screenshots use illustrative demo data, not customer results.

See Stokrio with your own data

Create a workspace in minutes. Connect Shopify or start with the API.

Get startedTalk to us