Our inventory demand forecasting feature leverages historical sales data and advanced statistical models to predict item sales for the next 6–12 months. Designed to help you make more informed inventory and purchasing decisions, the forecasts are generated at the per-SKU-per-channel level and updated weekly at end of day Monday, EST to reflect the latest data. We also allow you to enter manual forecasts for any SKU/channel combination either to compare against the Settle forecasts or to cover channels that Settle does not have historical data for.
This feature is currently in limited beta availability. If you're interested in getting on the waitlist, reach out to your Customer Success Rep.
Key features
Inputs to the forecast
Sales data: Historical sales per SKU by channel. Settle obtains this information by combining the orders data received when you connect your Shopify, Amazon, Walmart eCommerce or other sales channels and or when you connect your WMS.
Timestamps: The timing of sales transactions.
Forecast outputs
Aggregated views: Weekly and monthly predictions with the ability to toggle between views.
Weekly demand forecasts are for the coming 26 weeks
Monthly demand forecasts are for the coming 12 months
Channel-level insights: Breakdown of forecasts by sales channel for precise planning.
Machine learning models
We process all sales data received through three machine learning (ML) models to determine the most accurate approach:
Prophet
SARIMA
Autoregressive Models (linear regression)
These models will be tested against a holdout dataset (the last three months of historical data) to select the most reliable model for each brand. The forecast you receive back in the Settle App is the output of the model that best predicted the most recent three months of historical data from the prior 9 months of data.
Data requirements
Minimum data requirements:
At least one year of historical sales data as recorded via at least 20 weeks in the prior 52 weeks with > 1 unit sold.
Active sales in the last 30 days.
Prediction frequency
Forecasts are updated weekly at end of day Monday, EST, as new data becomes available.
Models are reviewed and recalibrated quarterly to ensure accuracy.
User interface
Accessing forecasts
Demand forecasts are available from within the left side navigation bar, Inventory section, Forecasts tab
Admin and Procurement roles can enable demand forecasting via the settings section in the Settle app.
Forecasts are displayed in an intuitive interface with weekly and monthly breakdowns for each SKU.
Searching and filtering
Users can search for forecasts by SKU name, internal SKU, or sales SKU.
Autocomplete functionality for sales channel searches.
Default sorting is by SKU name, with no manual sorting required.
Detailed breakdown
Clicking on a SKU provides a drawer view with detailed projections.
Default view is monthly, but users can switch to weekly breakdowns.
Manual forecasting
Users can add manual forecasts for SKUs, which won’t impact Settle's forecasts or charts.
Manual forecasts are pre-filled from existing channel totals or left blank if unavailable.
Users can view manual forecasts alongside Settle forecasts in the table.
Usage limits and upselling
Launch: Does not include demand forecasting.
Accelerate: Limited to forecasting on 25 SKUs. Users can reveal their preferred SKUs for forecasting. Upon selecting the 26th SKU, an upsell modal will prompt them to upgrade their plan.
Elevate: Inventory forecasting for all SKUs.
If you're interested in accessing this report and want to review pricing plans, please reach out to your Settle Customer Representative for assistance.
The forecasts serve as a starting point to guide your decision-making, with the understanding that you know your business better than any statistical model. Adjustments can always be made to reflect real-world considerations.
👉 Read more about WMS Integration in Settle.
👉 Read more about Shopify Integration in Settle.