Challenge

A multi brand global leader in the CPG sector faced a challenge to optimize product offering for point of sale in a highly competitive market. With hundreds of products, the client faced a risk of over-contacting their distributors and POS customers with sub-optimal product offers.

Scope

Use of Qualitest’s Machine Learning for retail solution to predict the best product variety for any point of sale. By predicting demand more accurately the company increased the diversity of products per point of sale, decreased inventories and returns and lifted revenues.

  1. The first phase was to gather points of sale by common parameters.
  2. The second phase was applying predictive algorithms that find complementary goods based on items inventory availability.
  3. The third phase was implementing the solution into the company’s production environment. In a short project time of eight weeks the demand prediction solution was ready for roll out.

Benefits

  • Predict and recommend additional products with high chances of purchasing by point of sale.
  • Recommend exact quantity of product’s ordering by POS to reduce POS risk of unused inventory.
  • Achieve high customer satisfaction, effectiveness and efficiency through customization and relevance.
  • Increase profit for POS by indicating the full selling potential at each POS.

Qualitest Solution

The company partnered with Qualitest to use its Machine Learning solution. Final model includes three main components:

  1. Demand prediction per product category
  2. Next best products offer
  3. Optimize items quantity from each product specifically for each point of sale.

For each component a separate prediction model was created, each model was fed with hundreds of commercial and location expletory variables. Our auto Machine Learning tool distilled the crucial features that contributed most and optimized the inner algorithm settings to secure the highest level of models’ performances.

Project Stages

Qualitest used complex data streams from multiple brands and product categories. The main model of demand prediction was built based on Machine Learning models ensemble. The next phase included the use of Market Basket analysis combined with prediction modeling for each product. An optimized quantity for each product was applied. The model components are now part of the company’s production environment.

  • Planning and Data reprocessing by the company data analysts, guided by the Qualitest team, explored the data and created a unified panel of explanatory variables.
  • Running on-prem, the Qualitest solution provides 40 Machine Learning algorithms with automatic finetuning to find the best explanatory variables combination.
  • Delivery of a stable and transparent prediction model ready for deployment in production.
  • Implementation: the Qualitest solution provides a variety of implementation methods and processes constantly changing demand data.

Endnotes

Qualitest’s AI-powered Machine Learning Predictive Analytics solution proved to be fast, easy to use and accurate when applied to complex data. The solution will enable the company to expand sales for each POS and increase operations efficiency while ensuring POS satisfaction.