Dynamic Field: Ranking Engine

Main articles:

• Electronic commerce
• Machine Learning

2021: Launch the engine for personalizing product listing pages

On April 1, 2021, the company Dynamic Yield announced the launch of Russian an engine for personalizing product listing pages for the e-Commerce base. Deep Learning

Pilot testing of the tool showed a 29% increase in revenue per user (RPU). This tool for April 2021 is already available not only on the world, but also on the Russian market.

For this solution, a self-learning commodity sorting engine (Ranking Engine) is used, which in real time rebuilds the pages of product listings on the basis of many indicators, creating an individual personalized experience for each user.

Dynamic Field found that before making a final purchase, about 80% of users view the product listing pages first. That is why it is very important to personalize them, ensuring relevant product issuance for each user individually based on their behavior on the site and a variety of other criteria. However, most listings still remain static and are displayed the same for all users, and merchandisers set up product issuance based only on hypotheses and assumptions in the hope of achieving higher conversion.

The Ranking Engine from Dynamic Field is designed to automatically optimize the sorting of products on the pages of product listings and uses a complex model of deep self-learning. The listing of goods is formed in real time and in a special way for each user at each point in time, using data about his interests, predicting his actions based on comparison with the behavior of other users, as well as taking into account the balances of goods, margin, current promotions and many other factors.

The e-Commerce market is overwhelmed with commodity offerings, especially during the COVID-19 pandemic. And this means that consumers should be able to quickly find the goods they need, "says Liad Agmon, CEO of Dynamic Field. - By using Dynamic Field's Ranking Engine to personalize product listing pages, brands can rearrange some of their most visited pages according to each visitor's interests and needs at each time, increasing the rates of transition to product cards, addition to the shopping cart and, most importantly, average revenue per user (ARPU).

{{quote'A large number of items pile up pages of grocery listings, complicating the search process. With our solution to personalize these pages with Dynamic Field, we have made them easy and understandable to each user at the key point of purchase. Now the displayed products are of greater interest and are suitable for each individual user, and it has become much easier to find them, "added Ekta Chopra, Chief Digital Officer, e.l.f. Cosmetics. }}

Gradually spreading among other customers of the company, this solution is part of the Dynamic Field platform AdaptML system, consisting of Deep Learning algorithms. Using the data of each individual user, the platform creates a personalized digital experience for them at each particular moment in time.

The personalization of listings for a long time remained the hottest request from our customers - the effect of the implementation of such a case is obvious, however, personalization tools present on the market for April 2021 did not offer such functionality. Deep Learning algorithms, taking into account literally hundreds of factors in addition to the obvious set of personal preferences of the user, bring the logic of sorting the listing to a new level - imagine that an individual selection of goods was formed for you by a professional consultant, and the use of technology server API allows you to sort listings from thousands of goods in milliseconds. We are already launching this module with a number of Russian customers, and we are waiting for the results, "commented Andrei Tyshchenko, Regional Director,,, Russia CIS Vostochnaya, Europe Dynamic Field.