Demand Forecasting

Pluto7 demand ML solution uses state of the art Machine Learning models to take full advantage of your data and give you the most accurate forecasts. Integrating seamlessly with your demand, supply, inventory, pricing, replenishment, promotion, and value chain planning it keeps your forecasts constantly up to date and your business always ready to grow.

Pluto7 demand ML forecasting software makes your data work for you.

Demand ML is a combination of outstanding computational power, intelligent algorithms and machine learning allows for optimized demand forecasting that automatically factors in the impact of all known external events, such as holidays, seasons, trends and even weather forecasts, as well as planned changes in your assortment, pricing, store-space allocation, stock policy and promotional activities.

Don’t just forecast demand, shape it

Demand doesn’t occur in a vacuum. External factors, like weather, have an impact and your demand forecasting needs to take account of them, accurately. Yet, many demand changes are self-determined. By closing gaps in how you forecast and reflect the effects of your own actions, such as planned promotions in all other facets of your business planning, you can significantly increase forecast accuracy and, even more importantly, ensure your plans are executed coherently.

Highlighted features of Pluto7’s demand forecasting Algorithm



Promotions matter in retail, yet they are notoriously hard to forecast as so many things can affect the results. Our demand forecasting software uses multi-variate regression to analyze a wide range of factors, such as timing, product and campaign type, marketing effort, in-store display, and pricing strategy to identify the best forecasting approach for each promotion, item and sales channel or store. The resulting promotion forecasts are routinely tens of percentage points more accurate than those using simpler approaches based on average historical uplifts.



In retail, thousands of new items are introduced every year, and you need to be able to generate good forecasts for all of them – for each store or sales channel. Manual forecasting methods and approaches based on manually setting reference items for new products are rarely adequate, so RELEX has built an auto-reference model that automatically identifies the best available reference product for each new item, based on attributes such as price point or brand. To further improve forecast accuracy, a ramp-up profile modelled based on past introductions can be applied. Read our Measuring Forecast Accuracy: The Complete Guide where we explain the facets of forecasting more extensively.