Digital transformation of the e-commerce industry.



"Personalization" – the word for 2022


Over the last year, our shopping habits have changed a lot. We started to visit e-commerce websites more often and make bigger purchases. Our behaviors and needs have also changed – 80% of people declare that they are more likely to buy in a shop that cares about personalizing experiences (source: Instapage). E-commerce websites must respond to these new habits if they want to meet our expectations. And the sooner and more accurately they do it, the better – the competition never sleeps, and the online customer can go to another shop much more easily.


In order to maintain the engagement of the current consumers and to gain new users and their loyalty, it is necessary (no surprise here) to skillfully use the potential of the available tools. Google cloud offers such a range of ready-made services.


Recommendations AI – an intelligent recommendation engine from Google


For years, Google has been taking care to provide users with content that matches their intent and needs. The learning mechanisms and getting to know the user in Ads, algorithms for suggesting videos on YouTube or creating lists of results in the Google Search engine – these technologies that the giant already has at its disposal have been modified and made available to the owners of e-commerce websites as the Recommendations AI cloud service.

Recommendations AI is a self-learning prediction engine that combines information on product availability, promotions or special offers with data on user behavior – current and historical campaigns. The service continuously analyses and serves new product suggestions in a few seconds in response to changes in the product range or user behavior.


Google service and its own recommendation engine


The recommendation engines developed so far often focus on physical dependencies and are limited to suggesting products from the same category or based on the choices of other customers. The advantage of Recommendations AI is access to historical data, the ability to immediately analyze it and read intent. The service can catch such nuances as longer views of a specific subpage or hovering the mouse over the graphics of a given product (let's remember that the service is part of the Google ecosystem – it uses, among others, Google Analytics or Google Tag Manager and is able to correlate this data well).


Digital transformation with a Google Cloud partner


FOTC, as an official Google Cloud partner, has been supporting private and public organizations from Central and Eastern Europe in the effective use of Google cloud services since 2014.


We have 70 experts on board, including certified engineers and architects, as well as security and network specialists. We work with over 2,000 companies, including the largest Polish brands in the retail industry (Arena.pl, Erli.pl, Selsey). In 2020, Google Cloud granted us the Work Transformation specialization, confirming a sufficiently high documented level of knowledge, competence and experience in implementing Google cloud services.


We provide our clients with free funds for testing Google cloud services (including the recommendation engine) and ongoing support at every stage of their journey to the cloud. Our knowledge and skills allow us to fully implement the Recommendations AI service among e-commerce websites within a few weeks.


What is the service implementation like with the support of FOTC?


The implementation of the service has three steps:

  • data entry

  • creating a model

  • indicating the location of displaying recommendations

In the first step, a Google Cloud partner imports data about the shop assortment and user behavior to the service. If the e-commerce website already uses other Google tools (Google Tag Manager or Google Merchant Center), the partner integrates the engine with these services, so that the data is downloaded automatically. Otherwise, it is necessary to export data and adapt it to the requirements of the cloud service.


The next step is to choose the type of the AI model and adjust it to the needs of the website. The client can choose from three types of recommendations:

  • Other products you may like,

  • Often bought together,

  • Suggestions for you.

FOTC also supports the client in the selection of the model KPIs. It can be click-through rate, conversion rate or revenue per session.

The last step is to indicate where the product recommendations should be displayed on the website. It can be the home page, a product page, a popup after adding an item to the cart, the cart itself or mailing campaigns.

The Google Cloud partner accompanies you at every stage of implementation and optimization. It configures the service, monitors the training process and periodic fine-tuning of the models, and supports the client in the ongoing use of Recommendations AI and the Google Cloud Platform infrastructure.


Find out more during Wioletta's talk


We cordially invite you to the talk "Harness the power of AI to sell more in 2022 ", during which Wioletta Lamot-Majewska will present Recommendations AI, provide examples of the use of the service and tell you what steps need to be taken to achieve ROI for real personalization in 2022.