• May 19 2023 | 6 mins read

Multisearch: Ecommerce Search Enrichment Using Existing Visual Search Tools

For savvy, experienced online shoppers, the visual search function isn’t an unfamiliar tool. They simply upload an image of an attractive product, and the e-commerce store’s search system will present similar products – if not the same one – right in front of their eyes. Visual search has become a practical helping hand for customers who do not have the best words to describe the items they have in mind. Similarly, it is an indispensable tool for ecommerce search enrichment targeted towards online shoppers who may not be proficient in the language of online retail sites. 

The importance of a visual search tool also lies in the portion of consumers who rely on it. Research reveals that 78% of consumers claim to use search bars for products – this percentage stands in stark contrast with the use of navigation menus (claimed 49% of consumers), filter features (37%), and homepage recommendations (30%) during online shopping experiences. User-friendly visual search tools not only create hassle-free, enjoyable experiences for shoppers, but they also prevent issues of search abandonment – that is, they help ensure that consumers don’t give up their product search due to the website yielding few relevant results with word search alone. 

But search functions are constantly evolving. A simple visual search might not always satisfy time-poor shoppers anymore, and an upgrade seems overdue. For instance, a woman may lock eyes on a passerby’s mauve, raglan sleeve, silk top and crave to know what it’s called, how much work it requires to maintain, and where to purchase it. That is a lot of questions popping up in this potential consumer’s head. To dig up all the answers, she would need to first do a visual search of the top, and then type in her question on fabric care on a search engine, and then search for sites that sell that specific top, and then zoom into delivery information. For this woman, the tedious process may drive her away from purchasing the top, especially if she’s caught in the middle of a busy street.

Ecommerce Search Enrichment Using Multisearch

Frustrating situations like these are why Multisearch exists. It’s a compelling search tool, and ViSenze is hard at work now, developing it for you.

At present, Multisearch has been released as a relatively new tool on Google Lens, and it has now expanded to include at least 70 languages in its database. It allows users to combine images and textual phrases to search for very specific items and/or information in mind. This is a leap from traditional technology, with which learners can only do either a visual or textual search at one given moment. It allows users to quickly search for items that they cannot precisely describe and even better, look for varying attributes in relation to those items. 

By way of illustration, if a consumer adores a very specific pencil skirt design but despises its summerey, bright green shade, then she may screenshot it and insert the query ‘brown’ in search of a more seasonal look. In contrast, if she were to do a traditional search using the words ‘brown pencil skirt’, she may not be presented with the exact design in mind. The same tool can be used for matching items. For a busy student looking to buy additional furniture pieces for his dorm room, he could simply upload a photo of his bed and request for, let’s say, a ‘matching bedside table’, thus saving himself the effort of going through endless catalogs with time that he does not have. 

While ViSenze’s Multisearch function will mirror that of Google’s, our tool would be tailored to fit business use. In other words, while Google’s Multisearch system narrows down results from the broad, limitless world wide web for the consumer, at ViSenze, our Multisearch function enables users to search for answers within a retailer’s diverse e-commerce catalog. 

The envisioned outcome of ViSenze’s Multisearch tool is threefold. Firstly, we aim to help customers pinpoint the specific item(s) they need with utmost efficiency, identifying information of interest such as size, color, style, cut, model, along with any other visual or functional attribute. That is, this function will be designed for the high-intent (and perhaps fussy) customers who have a very clear picture of what they seek. This ‘all-in-one’, image-plus-text search strategy can realistically lead to more customer satisfaction (less time wasted!) and by extension, more product discovery as well. 

Furthermore, with far-more-specific queries on Multisearch, our ViSenze system keeps a close eye on what each and every visitor prefers, helping businesses to market goods with highly advanced personalization (e.g., last time, this customer seemed to like our wine products but was particularly curious about the origin – let’s market according to that). This way, there is no need to make generic guesses about what each customer likes, or what specific information they seek. Finally – and perhaps most crucially – at ViSenze we garner insights from your visitors’ activities more broadly, thus equipping you with valuable knowledge of how your line of products can be re-designed, re-developed, or re-marketed.

While Multisearch isn’t available yet on ViSenze at the time of writing, there are many reasons to believe that it would be a powerful add-on to the existing array of AI products that we have. Until then, our industry-leading Smart Search, with a strategic blend of visual product discovery and marketing solutions (‘Complete the look’, ‘Shop the look’, and ‘Visually similar’), can also enable a more efficient, more enjoyable shopping experience for your customers. Learn more about why these AI solutions matter with our blog collection, or simply request a demo with us to test how well our services can help your retail business.