RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

Artificial Intelligence Optimizes Promotions for Henkel

Customers: Henkel Russia (Henkel Rus)

Contractors: Rubbles
Product: Rubbles SOPP
На базе: Artificial intelligence (AI, Artificial intelligence, AI)

Project date: 2019/12  - 2020/07

2021: Increase the accuracy of planning promotions by 10%

The company Henkel has implemented AI a solution-company, the Russian IT Rubbles which allows to improve the accuracy of planning and increase sales of detergents produced under the company's brands. Rubbles announced this on April 6, 2021. Developed using artificial intelligence technologies, the system began to work in July 2020, and in 9 months it led to an increase in the accuracy of planning promotions by 10%. The model also allows you to increase the calculation speed of the forecast and avoid errors related to the human factor.

A solution based on artificial intelligence allows company employees to evaluate the effectiveness of various promotional scenarios and choose the optimal one in terms of goals. The employee of the manufacturing company determines the list of products, the start time, duration, mechanics and the type of support for the future promotion. For example, washing powder will be sold at a discount of 20% for three weeks, in parallel will go advertising the promotion on TV. The obtained data are included in the AI-based forecasting system, which calculates the demand forecast and estimates the effectiveness of promos for each product. The introduction of artificial intelligence solutions is part of the development of Henkel digital technologies, which are one of the key strategic priorities of the company.

The Rubbles model includes features with a description of the product, a retail chain in whose stores promos are planned (matrix width, number of outlets, regional representation, width of the assortment matrix, etc.), stock mechanics, as well as signs describing sales behavior, and price perception of customers (price elasticity, promo frequency, historical depth of discounts, etc.). For the effective operation of the system, it is necessary to accumulate historical data on the effectiveness of promotions for a period of 1.5 years.

Шаблон:Quote 'author = said Alexander Zarakovsky, head of product analytics at Rubbles.

Шаблон:Quote 'author = noted Petr Nodel, Head of Sales Information Support at Henkel, Laundry & Home Care.