Google Retail API

Retail API is a suite of solutions from Google that helps you implement personalized search and product recommendations enhanced by machine learning into your e-commerce or mobile app.

The tools work on the basis of state-of-the-art machine learning models and can thus be an interesting option for anyone who is not an expert in this field and has no ambition to create their own model.

What does the Retail API consist of?

1. Recommendations AI 

Recommendations AI allows you to create a personalized product recommendation mechanism without having in-depth knowledge of machine learning. The entire system combines information from your product catalog (e.g. Google Merchant Center) with information about your customers’ behavior.

Recommendations AI includes:

  • Custom models – each model is trained specifically for your data based on sequential machine learning models.
  • Personalized results – all recommendations are tailored to the customer. The algorithm works with each customer’s individual online behavior. 
  • Real-time prediction – each prediction considers all the actions the user has done on the page before (e.g. view, add to cart, purchase, etc…).
  • Automatic model training and refinement – the model is “retrained” every day to account for the most up-to-date data. 
  • Optimization goals – you simply set which goal you primarily want to focus on – click-through rate, conversion, turnover, etc.
  • Omnichannel recommendations – thanks to the API model, you can really optimize the entire customer journey – personalized recommendations in email, mobile apps, …

When implementing Recommendation AI, you can choose from several types of models:

  • You could like – predicts, which other products a customer will interact with and which are most likely to be purchased. Depending on the product categories the customer viewed in a given visit.
  • Frequently bought together – predicts items that are frequently bought together with those that the customer viewed in a given visit.
  • Recommended for you – predicts which other products the customer will interact with and which are most likely to be purchased by the customer. It does not take into account the categories that the customer viewed in that visit.
  • Similar items – predicts the items that are most similar to the product the customer is currently viewing. 

2. Retail Search

Retail Search allows you to provide high quality product search results that can be customized to meet the needs of your business. Use Retail Search to improve product searches on your website and mobile apps.

Retail Search includes:

  • Search Enhancement – Increase the number of relevant results returned for queries that would normally provide fewer results. For example, for searches that use very specific keywords.
  • Relevance thresholding – adjust the precision (relevance of the search results returned) and quantity of results (returning more results for a given query) of a search.
  • Filtering – use search term syntax to filter to refine your site’s search results.
  • Sorting – set the order of your search results according to several priority categories.
  • Results prioritization – control the ordering of search results by prioritizing or de-prioritizing certain types of results.

Why use the Retail API?

The fact is that artificial intelligence is no longer a thing of the distant future. In some areas, it’s even already far surpassing humans. This is especially evident in the online world where we have extreme amounts of data – humans are simply no longer able to take into account all variables. 

The Retail API allows you to take advantage of artificial intelligence and machine learning algorithms without having to be an expert in the field. Initial evaluations of the Retail API directly from Google show significant improvements in search results, click-through rates and turnover.

As one example for all, IKEA’s e-commerce store recorded a 30 % increase in click-through rate and even a 2 % increase in average order value after implementing Recommendations AI. IKEA is currently using Recommendations AI for the majority of searches.

What is our opinion?

As part of an internal hackathon, we tested the use of the Retail API in practice and got a feel for its main features. The Retail API is a great option for businesses that are considering implementing personalisation and personalised search into their site using third-party services, but want to retain maximum control over the whole process.

The Retail API is also suitable for companies whose data is already well anchored in the Google ecosystem – they use Google Analytics to measure traffic, use Google Ads for campaigns, and have their product portfolio loaded in Google Merchant Center. 

Thanks to the possibility to use the integration to the above, piloting the launch can be a matter of a few days. It is possible to import historical user interactions on the web, which can be used to learn models instantly, so there is no need to wait, as with other solutions, e.g. half a year for the model to learn to provide the correct results. The solution offers easy integration via Google Tag Manager and with the commonly available Enhanced Ecommerce layer, the required data can be easily collected by modifying the relevant objects.

Especially for companies that have already entrusted data to Google Cloud, we recommend testing this solution.

How to get started?

In order to start using machine learning models hidden behind the Retail API in your business, you will need a few essential things – including a product catalog (Google Merchant Store or a special feed) and historical information about customer behavior (can be taken from GA4 or prepare your own dataset).

Anyway, the easiest way to get started is to contact Archetix! We will help you with setting up the data collection from your e-shop and the subsequent evaluation. We will recommend the most suitable type of model for your business and ensure integration into your project.

Scroll to Top