AI-Supported Sales Forecastsfor Markant

Retailers and manufacturers need accurate sales predictions to plan, increase sales, reduce waste, and avoid empty shelves. Markant has the mammoth task of streamlining these forecasts for 14,000 suppliers and 150 distribution partners. Each of these companies had a treasure trove of data that no one could access. Why not use these unused data sources, analyze them together, and share the findings with individuals?

For such analysis, speed is crucial as there are only five days between collecting data and the forecast deadline. Additionally, Markant’s partners are allowed to only see their own data. The solution also must be scalable and flexible enough to integrate non-harmonized data from all retail partners.

MAPP aggregates all retailer sales forecasts—anonymized and harmonized across all retailers. Processed data lets suppliers plan their production schedules and serves as an early warning system for order peaks and troughs.

The predictions must be as granular as possible, broken down by item, distribution center, and store. They must also consider seasonal and event-driven sales fluctuations. MAPP optimizes distribution and warehousing, transportation routes, production planning, ordering, and increases customer loyalty.

With MAPP, Markant provides its partners with sales forecasts on a weekly basis, six months in advance. The system receives, validates, classifies, and compares 100 GB of new data for each prediction with the old data. MAPP then also links all semantically related data.

MAPP calculates 1.6 billion forecasts in half a business day. The quality of the cross-dealer forecast is significantly better than even the best forecast from a single supplier. All companies involved benefit from a much better planning basis for production, transport, storage, and sales. Dashboards, email alerts, and IDE interfaces allow quick handling of distribution or special cases.