The importance of being Relevant
As a merchandiser, relevancy is a term that you’ll hear often.
Depending on who you ask, relevancy has many meanings. But everyone will agree that relevancy is centered around tracking and analyzing a shopper’s preferences to show them the products that they are looking for.
Relevancy and Ranking have helped eCommerce stores to anticipate and prioritize a shopper’s needs to provide a seamless retail experience.
Why Unbxd?
When it comes to delivering personalized eCommerce, your store should be equipped to gauge the interest and patterns of a returning shopper.
As an online shopper, the sheer volume of products deters me from navigating via the product category pages when I’m shopping. I prefer searching for the product either by what I intend to buy, by the name of the brand, or the type of product.
But site search isn’t always as effective as I would like it to be. A perfect site search solution should be able to:
- Autocomplete my search query as I’m typing. Maybe even autocorrecting the search query in real-time.
- Suggest products even as I’m typing in my keyword. Maybe even search by synonyms to provide me with products that match my intent.
- Display a search page that isn’t cluttered and allows me to sort by price, popularity or newness.
- Provide me with filters and facets that can help me zero in on the product I need.
- Display high-quality images that correctly represent the product
- Select and compare products and features side-by-side.
- Allow me to preview and add a product to the cart without taking me to the Product Display Page.
- Display search results within 2 seconds and have a responsive design.
In other words, your site search solution should score high on relevancy. I should be able to find the product I’m looking for with ease and simplicity.
Unbxd understands this need of shoppers to be able to find what they are looking for. That’s why we construct and deploy AI and ML models that can fine-tune this relevance to three nines. In this blog, we present to you some of the top use-cases which we have solved for our eCommerce customers across the verticals.
Fields: Unbxd recommends a list of fields in the catalog that should be marked as ‘searchable’. This helps businesses, no matter what vertical they are in, a running start to getting the search results right. For instance, identifying the ‘Part number’ for an auto parts store correctly would be as important as getting the ‘color’ correct for a beauty store.
Tokenization: Unbxd splits multi-worded search queries into related product categories and product-related attributes. For instance, ‘Outdoor Tent’ will be tokenized into ‘Outdoor’ (Category) and ‘Tent’ (Product). Similarly, ‘Silk Shower Curtains’ are tokenized into ‘Silk’ (Fabric) and ‘Shower Curtains’ (Product).
Stemming: Unbxd recommends you stem common search keywords and phrases to its root or base unit. For instance, queries like ‘Running shoes’ and ‘Runners shoes’ will provide the same set of search results, if ‘Running’ and ‘Runners’ is stemmed to ‘Run’.
Synonyms: With a growing list of synonyms for different verticals, Unbxd allows your store to hit the road running. Our comprehensive list of synonyms helps your shopper find the product they need no matter how they search for it.
Exclude Terms: As a merchandiser, you’d like to exclude products that may have the same keywords searched for, but isn’t what the shopper intended to see. Unbxd allows you to create a library of keywords or phrases that you don’t want to be displayed interchangeably. For instance, queries like ‘Dress Shirt’ may also display ‘Shirt dress’, which is not the same category of product and does not meet the shopper’s intent. Unbxd allows you to insert ‘Shirt Dress’ as an ‘Exclude Term’ so that results for ‘Dress Shirt’ won’t display ‘Shirt Dress’.
Mandatory Terms: Mandatory terms work on a similar principle of Tokenization and Named Entity Recognition, where your search engine recognizes the most important keyword or phrase in a query and displays results from the relevant category. For instance, queries like ‘Jumpsuit Dresses’ will give results that have the keyword ‘Jumpsuit’ either within the product title or its description.
Products by Popularity: Unbxd allows you to promote products that are popular in the category your shoppers are browsing. Clickstream insights reveal the number of times a product has been clicked on, added to cart, and purchased in the last 30 days. These statistics allow merchandisers to create campaigns that boost your product’s findability and enhance search relevance, which we will talk about in detail in a future blog post.
Ranking Fields: Once fields are assigned as ‘searchable’, Unbxd recommends indexing fields by its ‘weight’ or ‘relevancy’. For instance, the ‘Title’ field is ranked ‘HIGH’ since shoppers often search for a keyword or phrase that exists within the title. Similarly, ‘productType’ or ‘catLevel’ will be ranked ‘NORMAL’, while ‘color’ will be ranked ‘LOW’.
User Behavior: Inbuilt User behavior engine allows Unbxd to track, collect, and analyze how shoppers interact within your store. Data from Clickstream and Ranking Insights helps you plan careful campaigns that boost your product’s findability and revenue.
Apart from this, Unbxd allows you to rank relevant products higher using the Boost/Bury feature within the Unbxd Console.
These are just instances that are most commonly used by the customers who come to us to fix their relevancy. We do much more than just this to ensure that no shopper returns from your website with 0 or irrelevant products found. If your shopper leaves your store because of a poor experience, they’re never coming back. We understand how you want to create a seamless shopping experience for all your shoppers, where searching for the right product doesn’t take moving heaven and earth.
If you’re wondering how you can make your site search system more relevant to your shoppers and profitable to your brand, look no further. Unbxd is your go-to place.
Send us an email and we’d be happy to show you how we can help power your brand’s relevancy engine.
