Recommendation Engine

Recommendation Engine

A technological solution for online stores that generates the best offer for individual customers

Increase online sales
Smart data analysis
Smart data analysis

The system with the help of ML machine learning analyzes the user behavior (purchases, search, cart) product specifics (filler, substitute, perm alternative) and sets a recommendation for the customer

Customizing the suggestions
Customizing the suggestions

Based on data analysis, the best-personalized suggestions for specific users are generated in live mode, which can be offered in the online store, as well as by email or another form

Increase in customer satisfaction
Increase in customer satisfaction

The system makes it convenient for customers with personalized suggestions and simplifies the purchase process, saves customers time and energy, which increases their level of satisfaction

System Architecture

The system of recommendations consists of two layers – data and application levels.

At the data layer, a warehouse is created in which information from customer data sources is periodically stored/updated. In addition, analytical algorithms are periodically launched and the results are saved in special data warehouse tables.

Special rules are set at the Application layer, based on which the system determines exactly what type of recommendation should be given to the Customer.

Principles of operation of the recommendation
Principles of operation of the recommendation system

In parallel with customer activity on the e-commerce platform, demand is created in the referral system. The recommendations engine generates a set of optimal suggestions according to pre-established rules and algorithms. After that, the suggestions created by the recommendation engine are returned to the customer in automatic mode.

Principles of operation of the recommendation
Curious to know more about the recommendation system?
Curious to know more about the recommendation system?
Stages of creating a recommendation system

In the process of setting up the system, the customer and the team of BDO Digital Data Lab are jointly involved, both technically and analytically. The process consists of the following main stages:

  1. Analysis and preparation of historical data
  2. Selection and development of analytical models
  3. Formation and writing of rules (depending on business tasks)
  4. Integration of the recommendation system with E-commerce
  5. Observation, improvement of algorithms and rules according to the results
 Stages of creating a recommendation
The Head of the Data Laboratory

Irma has held various leadership positions, including as Lean Transformation Coordinator, IT Portfolio Manager, and Agile Coach. Lean Management was introduced in Georgia together with the consulting company Mckinsey & co and a local team, both in the front office and in the back office.

Irma Berdzenishvili

  • IBerdzenishvili@bdo.ge
  • +995 32 254 58 45
  • Linkedin
  • Get acquainted with the trends and news