With Retail Search, Google Cloud provides an answer to the dreaded problem of search abandonment for customers of e-commerce sites. Recommendation optimization functions and recognition of products from an image are pushed.
When they don’t immediately find the desired product on an e-commerce platform, customers often abandon their search, a product discovery problem that online retailers commonly face. That’s what Google Cloud’s latest fully managed search tool called Retail Search tries to solve. And for good reason: according to a study conducted jointly by Google and The Harris Poll, this phenomenon, which Google calls search abandonment, could cost American retailers approximately $300 billion each year. This study also found that 94% of US consumers abandoned a shopping session because their search results were irrelevant.
Originally announced last year, Retail Search uses Google’s expertise in indexing, retrieving and ranking search results for the specific product catalog of retail players. “It’s a comprehensive service that brings together all the capabilities of a search engine, and replaces the individual components of ranking, indexing, rating and so on,” said Srikanth Belwadi, chief product officer at Google. Cloud. “There is no need to pre-process data, train or over-tune machine learning models, or manually load balance or provision infrastructure to handle unpredictable traffic spikes. . We do all of this for you automatically,” he added.
Output AI and Vision API Product Search Recommendations
Other features include semantic search that efficiently matches product attributes with website content for quick and relevant product discovery, and optimization of product results based on user interaction. users and ranking models. Retail Search builds on Google Cloud’s existing retail-focused solutions, such as Recommendations AI and Vision API Product Search. Launched in July 2020, Recommendations AI is another managed service aimed at improving product recommendations. As for the Vision API Product Search service, launched in 2019, it allows customers to perform a product search from an image.
According to analysts, Retail Search targets retailers looking to differentiate themselves on the digital offer and want to offer competitive online shopping services. “In retail in particular, it can help companies get their customers to quickly find the product or service they are looking for, or to order a product with very similar characteristics, in order to increase conversion rates, cross-selling and up-selling,” said Hayley Sutherland, senior research analyst at IDC.
A stake in gaining market share
For Tracy Woo, principal analyst at Forrester, Google is looking to gain more market share while having access to more data. “These are machine learning algorithms, so the more data collected, the better the results,” she said. “The integration of Google’s search engine into retailers works both ways: retailers benefit from Google’s higher search capabilities and in return, Google collects all the data obtained from these search results”. Google made it clear that its Retail Search models were exclusively limited to highly isolated customer data and were not used to improve Google Search services.
Finally, according to Constellation Research Principal Analyst Ray Wang, Retail Search is very different from anything offered by competing cloud vendors Amazon Web Services and Microsoft Azure. “Google Cloud provides customers with tools to effectively compete against Amazon. As you know, most retailers won’t use AWS for cloud and Google is taking advantage of this by cross-selling. Customers get the capabilities of Amazon.com, without having to go to Amazon and AWS,” he added. Retail Search is now generally available to Google Cloud customers.