Klevu - Product Discovery Platform

Product Discovery Platform: Search, Merchandising and Recommendations


Millions of shoppers use Klevu to discover products on their favorite ecommerce sites. Klevu's proprietary technology increases conversion, reduces bounce rates and drives loyalty for more than 3,000 leading global brands, including Puma, Fred Perry, Paul Smith, Avon, Stüssy, Pfaltzgraff, and Native. Klevu is an AI Search and Discovery Platform that leverages AI, Natural Language Processing and User Behavior Analytics to elevate the search experience, and automatically re-merchandise category listing pages and product recommendations. Retailers that use Klevu's full Product Discovery Platform provide unparalleled customer experience, and achieve 37% more revenue per web session than those that don't.

Description of the Integration


The integration of Klevu with your commercetools store involves the following steps:

  1. Setting up an API client in commercetools

  2. Signing up a Klevu account

  3. Frontend integration to start using Klevu

Our article Integration steps for commercetools explains step-by-step instructions to complete the above steps.

Data mapping between commercetools and Klevu

If you are wondering what data Klevu reads from your commerce tools store, please look at our article on Data Mapping between commercetools and Klevu. The article also lists frequently asked questions by our customers.


Finally, please make sure you have read our guide Pre-integration Checklist to get the maximum out of our Klevu solution.

How Klevu AI Works

At its core, Klevu is a search relevancy engine trained on a data set of nearly 10 billion search queries covering 40+ ecommerce industry domains, 30+ languages and many examples of a variety of search query types as suggested by the Baymard Institute (see https://baymard.com/blog/ecommerce-search-query-types). As shown in the diagram above, internally, Klevu enhances product catalogs by performing morphological normalisations as well as semantic enrichment to add contextually relevant data. This helps in reducing the occurrences of no result found (i.e. bounce rate) as well as reducing the time from search to purchase.

The relevancy promise, one of Klevu's USPs, is fulfilled by the real-time semantic analysis of user queries to identify the main product nouns (that the consumers may be after) and the respective context to bring the most relevant products higher up in the search results. Extending the definition of relevancy here, Klevu's ranking algorithm also takes into account product-based trend, users' individual preferences (i.e. personalization) and merchants' business targets (i.e. increasing average order values and overall revenue).

Unlike most “garbage-in-garbage-out” keyword-based search technologies that require a lot of manual set up, Klevu is both an automated information enhancement system as well as a semantic search engine that only demands a bit of patience from the merchants while the engine learns itself from the consumer shopping patterns.