Sales Forecasting
Introduction
This plugin allows you to generate predictions of future sales quantities, by integrating nopCommerce with Majako's Sales Forecasting API.
Using machine learning and statistical analysis of your order history, the forecasting engine can train a finely tuned prediction model for your store. This allows you to minimise dead stock while simultaneously reducing the risk of out-of-stock orders. By lifting the administrative burden of manually predicting future sales for each product off your shoulders, the plugin lets you use your time on more productive tasks.
Features
Majako's Sales Forecasting plugin connects your nopCommerce store to our cloud-based forecasting API. With a normal product search, the plugin collects order quantities from your order history and sends it to the API, which fits a machine-learning model and predicts per-product sales during a given number of days into the future.Robust and dynamic
The forecasting engine will find the model that best explains your store's historical sales while still being as simple as possible. This guarantees robustness against outliers and anomalies, while at the same time allowing the model to capture and extrapolate trends in sales data.Discounts
Past discounts are taken into account when processing your order history, and you can choose to include discounts for the predicted period. The forecast will base your upcoming sales on the performance of previous discounts. This way, you can be prepared for increased sales during campaigns.CSV export
Forecast results can be saved with a single click to a CSV file, which can be viewed in Microsoft Excel or other spreadsheet software, where you can do more advanced filtering and sorting to get a better overview.Getting started
In order to use this plugin, an API subscription is required. A subscription key can be acquired via a link on the download page. Subscription is free during an initial testing period, after which you will be required to purchase a monthly subscription to continue using the plugin. Once acquired, the subscription key can be input on the plugin configuration page.
The forecasting page can be found under Sales in the admin menu. There, you simply specify the number of days ahead you want to forecast for, and make a normal product search for the products you want to include.
Notes
Generally, the more products (or specifically, order items) are included in the search, the better the results. However, there may be a trade-off with accuracy for products with different sales patterns, as the model will try to generalise as much as possible to prevent overfitting. If you notice a group of products getting worse results, try forecasting for these separately.
Finally, note that a (small) minimum number of products/orders are required to run a forecast. If you get an error, try including more products in your search.