Promotion forecasting

Maximize success of your sales promotions with DSLab's promotion forecasting

Promos are a new norm. This brings a higher level of uncertainty to the forecasting process. With an average of 60% of goods being on offer constantly, demand forecasting for discounted products is more important than ever.

Improve on-shelf availability without keeping excess stock with DSLab's machine learning technology. Get a granular forecast for each item on a sales promotion, on a store level, for 16 weeks ahead.

+ 13%

Lenta, one of the largest retail chains in Russia and the country's second largest hypermarket chain with 14.4 million active loyal customers, automates promotion forecasting with DSLab's technology

READ CASE STUDY

> 600

< 10

in prediction accuracy

features used
to build prediction model

seconds to make predictions for every SKU on a store level

Client story: promotion forecasting

Offer your customers the right product,
at the right time, at the right price

Focus on promo
We take into account promo-specific parameters such as price dynamics, advertising, in-store display, promotions history, and sales of substitutes.
Granular forecasts
Get highly-accurate demand forecasting for each item on a sales promotion on a store level.
Planning ahead
Keep up with marketing initiatives: upload your campaign calendar and get a granular forecast for 16 weeks ahead.
What-if analysis
Maximize the success of your sales promotions by tuning campaign parameters with built-in what-if analysis.
Machine learning
Machine learning technologies at the core of our solution deliver highly-accurate forecasts allowing you to increase on-shelf availability without keeping excess inventory.
Modified algorithm
Forecasting solutions tend to underpredict. We've modified our algorithm to avoid underprediction and lost sales.
Interpretability
On top of our prediction algorithm, we've built a model that interprets the results of the forecast. No more black boxes and unexplainable predictions.
Business metrics
Instead of focusing on accuracy only, we use business metrics - for example, the value of excess inventory - to measure the quality of the forecast.
Control module
To make sure our predictions are relevant, we added an automated control module to the forecast. The module is based on the expertise of retail professionals about certain demand characteristics for all types of goods.
Easy integration
DSLab forecasting solution is easily integrated with standard data sources such as Hadoop or SQL-based data storage, as well as replenishment systems.