Why Catalog Search is Critical to B2B eCommerce

In B2B ecommerce, catalog search could be considered a primary tool for customer interaction — and buyers expect high-level relevance, suggestions, multiple languages, faceting, and near-instantaneous responses. All of the above can be meticulously attained via the best b2b ecommerce platform.

Unlike B2C, in B2B ecommerce, users often look for items containing unstructured information. These can be manuals, drawings, assembly diagrams, photos, original and non-original (similar) spare parts. Since the average cheque in B2B is significantly higher than in the consumer market, vendors must do everything possible to provide the best user experience in catalog search, be it a goods or b2b services marketplace.

There are two essential factors in B2B commerce when we talk about product catalog search challenges. The first refers to the fact that corporate buyers (aka users) visit the vendor ecommerce website solely for the purpose of purchasing, and not for a pleasant time like in B2C. Their task is to buy the necessary goods or services for the company and do this as quickly as possible. 

Therefore, the vendor must value the time of the buyers, who are also paid by their employer. In other words, searching through the product catalog should work as quickly and as accurately as possible. But these two functions, accuracy and search speed, do not go well together in terms of the design and operation of IT systems. 

The second factor is that in most cases, the buyer must be authorized to access the catalog. That is, the seller knows the user's personal data and the company for which the user is buying. Accordingly, the purchase history of this user and of the entire company is known. Such information helps both the personalization of offers and the relevance of search results. 

Historically, “searching” a directory meant retrieving data from a structured data repository. With the advancement of search technology, today's sellers have at their disposal both open source Elasticsearch and cloud AI-powered cognitive search technologies.

Virto Commerce catalog search module supports Elasticsearch/Elastic Cloud and Azure Search engines.  Document-based full-text search and text analysis for products and categories are available for users. The search logic is extendable and customizable with third party modules. 

In this first article, we discuss why Elasticsearch is used for one of the most complex problems in ecommerce, namely search. The second article is dedicated to Azure Search service as the best on the market example of AI-powered cognitive search for B2B ecommerce portals.

There are 6 essential search engine requirements that have been successfully implemented in Elasticsearch:

The first and most common aspect of the best search engine is fast queries. Elasticsearch combines inverted indexes and multiple levels of caching and several other optimizations abled to fast data processing.

The next demanded feature is being able to search the data through the query language.  Elasticsearch has its own highly enhanced rich querying capabilities. It gives the ability to write complex queries without much work.

The third Elasticsearch item is having an API first support. When developers have different frameworks to work with and they want the integration with their search engines to be seamless, Elasticsearch with its API first engine is the best choice from day 1. Developers can accomplish a lot of complex tasks with a simple API call.

The next item is advanced controls which help developers to control the search experience they present. This can be anywhere between showing specific items as suggestions, adding weights to certain products and even tuning relevancy. 

In systems with a large amount of data, multiple copies of Elasticsearch can be combined into a cluster. Elasticsearch provides a full set of APIs that allow advanced controls. Scaling is an important topic, it's a combination of the data you want to make searchable versus the number of queries fired through it, also versus the number of users and accesses. As one of these factors scales out, you'd like for the search engine to scale out linearly. Elasticsearch was designed from day 1 as a distributed system. According to the concept of clustering nodes, the sharing in distributing the workload can scale almost linearly with your user’s workload.

Things like multilingual support, completion suggestions and even analyzers to precisely control what the text looks like are enabled in Elasticsearch. There is a whole set of out-of-the-box frameworks that help developers to build a search experience more easily.

Some other common reasons why developers like Elasticsearch and pick it to serve ecommerce catalog search: 

1. When a user searches for a product and while typing, Elasticsearch immediately pops up a list of suggested items. This is powered by a feature called completeness in Elasticsearch.  
2. The second feature is based on the concept of facets and grouping results based on categories. Elasticsearch powers this by a feature called aggregation, available again out-of-the-box.  
3. The third most important feature is the product search result displayed first and is usually achieved through an API call. Typically, the sequence in which the products are displayed is of great importance to the ecommerce portal owner.  

For more information on how catalog search works in one of the examples of ecommerce asp - Virto Commerce B2B ecommerce platform and details on search architecture, please visit the Search Fundamentals section of the platform documentation.