Search Relevance Controls for Merchandisers

In the age of artificial intelligence and machine learning retailers are increasingly looking for opportunities to automate various functions of their business. However, merchandising is one area where retailers rely on experienced merchandisers for creating delightful experiences for their shoppers. A merchandiser blends their understanding of shoppers and business priorities to come up with creative approaches to achieve the business objectives while inspiring the users to visit your store again. 

 

Unbxd offers a state of the art merchandising product that enables merchandisers to create a wide variety of merchandising campaigns to drive business growth. You can read more about how to effectively leverage our merchandising tool in this blog

 

Although the merchandisers have plenty of tools at their disposal to run effective merchandising campaigns, they have limited control over how the search engine interprets the shopper’s search query. Ability to control and optimize the interpretation of a search query is a common request we have received from merchandising teams of our customers. That is why our product is centred on simplifying how a search engine interprets a shopper’s search query and gives merchandisers the much needed control. 

What to expect? 

The current product and its upcoming releases would allow merchandisers to define different assets/dictionaries to control the interpretation of a search query. Although, our data science team has developed in-house algorithms which use the shoppers behaviour data to constantly improve the interpretation of the search queries and generate these assets/dictionaries automatically. These algorithms require massive amounts of data to work effectively and generate results with high accuracy. While we are working towards improving the algorithms, there was a need to expose controls that allowed merchandisers to fix the issues identified by them. The feature would be extremely useful for small retailers who do not have the sufficient traffic to leverage the AI algorithms mining these assets automatically. The merchandisers would be able to define following assets/dictionaries : 

Concepts

“Concepts” represents the terms in the search query which are considered extremely important by the shoppers. Hence, the products in search results must definitely include the concepts present in the search query.  Concepts can be thought of as a precision instrument which informs the search engine to give higher importance to precision over recall (i.e. number of products in search results). 

 

Let’s take an example to look at how concepts help in improving the precision of search. Let’s say a shopper who uses an Xbox console, visits an online game retailer to search for “Lego Batman Game for Xbox”. If the retailer doesn’t have the Lego Batman games in the stock they may choose to show other Lego games in response to the query to ensure that the customer finds related Lego games which are in stock. However, while performing this expansion the retailer has to be mindful that they are not showing PS4 Lego games to a customer who is searching for Xbox games. In such a situation, the retailer can define “Xbox” as a concept to ensure that customers searching for Xbox games only see Xbox games in the result set.

 

When to define concepts ? 

A merchandiser can use their understanding of shoppers to define concepts effectively. For example, if the shoppers of a fashion retailer are highly brand conscious the merchandiser can include all the brands in the catalog as concepts. This would ensure that when a shopper is searching for a specific brand, they only see the products from that brand only. 

 

Tip : Defining all the independent product types in the catalog as Concepts has generally worked well for most of our customers in the past. 

Phrases

As defined in Wikipedia “A phrase is any group of words, often carrying a special idiomatic meaning”. Phrases ( allows merchandisers to define multi-word terms as a phrase which is treated as a single entity by the search tokenizers.

 

Let’s take the case of a fashion retailer to check how phrases help in improving the relevance. Let’s say the fashion retailer carries “Shirt Dress”. When a shopper searches for a “Shirt Dress” they are interested in the shirt dress for women and would not want to see Shirts or other types of dresses in response to the query. Phrases allow merchandisers to define multi-term phrases which are specific to their business use-cases.

 

 

Stemming (Overrides)

Stemming is the process of reducing a term to its root form. Stemming allows a search engine to algorithmically handle similar terms without the need of explicitly defining synonyms. Although stemming is useful in the majority of cases, there are situations where stemming terms to their root forms may not be ideal for the business. For example, “painting” is derived from the term “paint” and the stemming algorithm would reduce the term “painting” to paint. This might be the expected behaviour for a hardware retailer however it won’t be the ideal behaviour for an Arts retailer selling decorative pieces (including painting). Stemming allows merchandisers to define the stemmed form of the term to override algorithmic stemming for certain terms. 

Stopwords

Stop words are a pool of terms that help in understanding the relationship between the terms in search queries but do not add any value while determining the rank of products in response to a query (Ex. at, to, and, for, etc). The stopwords section (in Unbxd Console) allows retailers to specify their business use-case specific stopwords.

How does it help Merchandisers ? 

Merchandisers had the option of defining merchandising campaigns to fix relevance for queries which were mis-interpreted by the search engine. However, the approach was not ideal because :

  • Merchandisers can only do it for the head queries which received significant volumes. The approach was not scalable for long tailed searches

  • Merchandising options such as landing pages, redirects override the default search behaviour and won’t be able to learn and improve from the shopper’s behaviour on the website 

 

To overcome these challenges, we decided to give more controls to the merchandisers so that they can control how the search engine interprets a search query. The new controls enable the merchandisers to override the search engines default behaviour for handling the business specific use-cases effectively. While our data science team is developing algorithms to solve the problem at scale with limited training data, these controls would go a long way in addressing some of the key challenges faced by merchandisers.

 

In the last one year, we have invested heavily on democratizing the controls of the search and enabling our customers to take complete control over their website search powered by Unbxd. Our product is steered in the same direction, simplifying merchandising controls further for eCommerce businesses of all sizes. We have a lot of interesting updates planned in the upcoming months so stay tuned to receive updates.  



More To Explore

Merchandising

Search Relevance Controls for Merchandisers

In the age of artificial intelligence and machine learning retailers are increasingly looking for opportunities to automate various functions of their business. However, merchandising is