Now reading:
[Case Study]: Implementation Of AI / ML Recommendations In A Women's Clothing Store

[Case Study]: Implementation Of AI / ML Recommendations In A Women's Clothing Store

Avatar Zuzanna Pajorska
15 marca 2021
clock icon 6 min

The AI / ML Recostream recommendation engine was implemented in the ByLola online women's clothing store at the end of October 2020. The implementation itself went very smoothly (the installation took store employees less than 3 minutes). The analysis of key metrics obtained after completion of the implementation showed an increase in conversion and CTR. What is more, the share of recommended products in sales exceeded 25%.

The client also wanted to regain traffic on the pages of unavailable products and increase the content of the shopping cart. We found solutions to these issues and thanks to the recommendation engine based on artificial intelligence technology and machine learning, we significantly improved the store's results in a short period of time.

About the Store

ByLola store online and recommendation engine

ByLola is a Polish clothing brand dedicated to women. The online store offers products made of high-quality fabrics in sizes between 44-68. ByLola's mission is to provide a wide selection of clothes in sizes which cannot be found in the most popular chain stores.

  • Platform: Shoper
  • Number of products: approx. 2,000
  • Store address:

About Recostream

We have been developing an AI recommendation system that allows you to boost sales by 5-10%, while allowing store customers to find products that interest them faster and more accurately. Placing recommendations is possible in a few minutes without the assistance of programmers, and their effectiveness is optimized by artificial intelligence algorithms and machine learning based on the actual customer behavior.

Take Advantage of the 30-day Trial Period

Purpose of the Implementation

Before integrating the Recostream engine with the store, we established 4 main implementation goals:

  • Increase in sales
  • Improvement of the CTR ratio
  • Increase in conversion
  • Recovery of conversions on pages of unavailable products

The Results

The analysis of the store's metrics after 30 days showed an improvement in key indicators. We analyzed independent results based on store statistics collected in Google Analytics Enhanced Ecommerce.

results of product recommendation

Deployment configuration: end of October 2020.

November 2020 until now:another record-breaking conversion rate in store history.

Where We Have Placed the Product Recommendations

The profitability of recommendations largely depends on the location on the store's website. After analyzing the traffic on the store's website, we placed recommendations in five places.

Where We Have Placed the Product Recommendations
  • Product page
  • Pop-up after adding to cart
  • Pop-up on an unavailable product page
  • Shopping cart page
  • Home page

Implemented Solutions

The AI / ML Recostream recommendation engine offers 8 types of recommendations, which differ in the way data about products is filtered and user behavior on the website. Together with the owner of the ByLola store, we decided to introduce all available models in order to maximize profits.

8 implemented models in various store locations:

  • AI Maximized Conversion
  • Most Similar in Category
  • Most Viewed in Category
  • Others Also Viewed in Store
  • Recently Visited in Store
  • Bestsellers in Category
  • Bestsellers in Store
  • Rule-Driven Recommendations
product recommendations
Recommendation model: Others Also Viewed in Store

product recommendation system
Recommendation model: Bestsellers in Store

recommendation engine
Recommendation model: Recently Visited in Store

Installation Process

  1. Registration on the website, where the email address and the URL of the store were provided.
  2. Lines of JavaScript code have been pasted into the shop's website. It took 3 minutes without any programmer assistance.
  3. The Recostream team made a configuration of the appearance and adapted it to the design of the store.
  4. The AI recommendation engine was launched in one business day.
  5. Trial period started: 30 days.
  6. The results were integrated with the Google Analytics account for independent and objective performance evaluation.
  7. Ongoing contact with the customer service manager was maintained.

The Client About Us

The plugin installation itself is a very intuitive and simple process.

The customer panel, where you can see the statistics and set the selected model that is to appear in our store, is clear and transparent.

Support is very fast and reacts in case of any questions and the entire Recostram team is very helpful. Cooperation is pure pleasure. :) The tool continues to develop and adds new useful functionalities.

I recommend it to both beginner stores and more developed ones.

Urszula Lola
The owner of the ByLola store

We Highly Recommend Taking Advantage of the 30-day Trial Period

Register and install the AI / ML recommendation engine in 3 minutes without the help of an IT specialist. We will take care of the rest.

If you have any questions, please contact

You can read more about how the recommendation engine works on our blog.

arrow-up icon