What’s broken in Search: Why retailers need a hard look?

Welcome back! We’ve got something really thought-provoking this time around–is text search the best we can do? The world has changed, there’s something new happening everywhere, but here we are, stuck with plain old keywords-based text search. Doesn’t seem right, eh?

We sat down with Diogo Quintas, our Head of Product (Search) at ViSenze, to talk about exactly this!

Wait…there’s more! In addition to his captivating insights, we’ll also delve into the news that shaking up the AI and fashion worlds! Without further ado, dive in!

Introducing Multi-Search: The fix that retail search badly needs!

Diogo Quintas is the Head of Product (Search) at ViSenze, leading initiatives spanning Visual AI search to MultiSearch technologies. He has a rich background of over 18 years in diverse technology and product development roles in retail.

Over the years, he has honed his skills as a senior product manager at Argos, the largest general merchandise British retailer, and as a product manager at Llamasoft, now part of Coupa, recognized for its leadership in supply chain innovation, among other roles.

Diogo’s expertise lies in understanding customer needs and developing innovative products that meet those demands effectively.

Recently, the ViSenze team had the opportunity to engage with him in a deep dive into the challenges plaguing retail search and discuss the retail industry’s future directions.

Q. 80% of online shoppers leave an online website because they struggle to locate their desired products. What is broken in retail search today?

Most search engines are based on matching keywords. When you input words in a search bar, the engine finds the exact keywords in the database of products.

However, given the diverse ways in which users typically express themselves, there are chances that your products might not be found by most people.

Search engines have techniques like synonyms to ‘normalize’ words with similar meanings to accommodate the richness of language. But, it requires a lot of work and resources to manage it effectively. Many e-commerce businesses do not have the resources to invest in engineers who can finetune the search engine for relevance. It is incredibly hard to make a search engine perform well, and thus most retailers and brands end up making do with a ‘not-so-great’ search.

The second aspect is the shopper’s perspective; as a consumer, you might not find it easy to articulate what you are looking for. Let’s say you find an image – something you’ve seen on Instagram. It’s a beautiful dress that has inspired you, but you can’t easily describe it. This is not limited to fashion – tools, electronics and so on can all be difficult to describe.

As a retailer, how do you help your customers find these products if the tools available are limited to a ‘not-so-great’ text search option? What we’re seeing is retailers exploring new ways to overcome the limitations of old search engine technology and that’s leading them to look at AI solutions like Visual Search, which has been around for a few years now, but more recently the focus has shifted to new AI-based semantic and contextual search technologies that can support multi-modal search and a more natural language search approach.

Q. What trends are you observing in search? How is the digital shopping landscape evolving?

The biggest shift in being inspired, finding, and buying products is that it is occurring more and more outside of traditional channels. The funny thing is, when I was growing up, the website was seen as a non-traditional channel, but now it is considered a traditional one.

New non-traditional channels like social media, (TikTok, Instagram, Telegram, and WhatsApp), are where much of the discovery happens today.

As a brand, it is becoming more challenging to understand discovery and search; you don’t know where the consumer is researching about you and what intervention you need to do in your funnel.

Your ability to deliver personalized results to the consumer is becoming more challenging. More and more, the consumer is coming to transact and not conduct research; that’s a growing business challenge as you miss out on the education and inspiration opportunities that you might have had in the past. When they come to your website/app, they’re much more informed today, and they come with specific asks.

Q. What is multi-modality search, and how can it solve the problems you’ve highlighted?

People are finding products and trends from friends and on social media. A lot of this inspiration is found in the form of images. Whilst Visual Search has been around for a few years now,  Multi-Search is a new search engine type that allows you to express a query through different modes–images, text, video, and so on.

With this, you are matching the image and test representation of the products. What that means for shoppers, is you no longer have to express your search in ways that will match how the retailer’s search engine is tuned. Multi-Search accommodates any word or form of expression, including vague terms or instructive sentences.

From a retailer’s point of view, a Multi-Search solution is easier to maintain; you don’t need to spend time fine-tuning it; it just works. You don’t need to create a lot of metadata for your product, it just works. With Multi-Search, the engine doesn’t rely only on structured data. So, providing a better search for your customers becomes a lot easier and cheaper to maintain. Your customers can express themselves with different input signals–images, text–and you are open to more ways of searching than ever before.

Q. This is interesting. Can you delve a bit deeper into it? What other advantages does the retailer gain?

The retailer can now expose the right part of their catalog to the right consumer. Let’s say I’m a climber and I’m looking to buy climbing shoes. There are all these things: type of rubber, how aggressive the shoe is, etc., basically, different features of these shoes. As a consumer, I know that I’ll use it for bouldering or indoor climbing. With the multimodal search engine, I can express my query as “shoes for bouldering”, and the engine will find me the right shoes for my hobby.

It allows you, as a retailer, to understand the consumer’s ‘job to be done’ or, in other words, what the consumer wants to achieve, as opposed to what product they think they need.

When retailers start tackling personalization/recommendations, they can now do it based on the consumer’s objective or intent, rather than solely basing it on products the consumer is looking at. It becomes about how well you rank on the objective to be met rather than on different ‘shelves/categories’. You can present your website and range by organizing it along those objectives rather than categories.

It is now all about making it relevant to your customer rather than the traditional prescribed way of a website, which is a linear flow, forcing the consumer to go on a prescribed path.

Q. Gen Z-er’s spending power is around $140 billion for online retailers. What’s your advice for retailers who are looking to cater to this growing cohort?

Young shoppers are increasingly transacting in conversational mode through Instagram, Telegram, and WhatsApp. The world is moving towards shopping anywhere through any means that is most convenient.

There is no doubt that retailers need to accommodate these trends. You’ll be left behind quickly if you aren’t moving to where your customers are. Conversational commerce is where people talk to you as a brand; if you aren’t there, you miss out!

Just to dig deeper, the GenZ, and even Millennials, for that matter, are increasingly transacting and finding products in social media and chat groups. If you want to acquire and appeal to these users, you must be in those channels and make the experience consistent across these various channels.

Say, if I find your product on Instagram, it should be easy for me to find it on your website. You should think about the integrated experience as a brand. It is no longer just the case that you can maintain a social media presence that is not thoroughly connected to your website.

For example, if you have an app, you can allow consumers to share content with your app in different ways or from different platforms, and using visual or multi-modal search technologies, help them find relevant products with the content they’ve shared.

These are simple steps to start engaging with these new channels, which is where your focus must be.

Bite-Sized News: Updates from the World of AI & Retail

Walmart and Microsoft Launch Generative AI for Enhanced Shopping

Walmart has introduced new generative AI capabilities to enhance shopping experiences for customers.

Using a combination of Walmart proprietary data and technology and large language models, including those available in Microsoft Azure OpenAI Service, as well as retail-specific models built by Walmart, the new capability can understand the context of a customer’s query and generate personalized responses.

Soon, customers will have a more interactive and conversational experience, get answers to specific questions, and receive product suggestions.

For more details, you can read the full article on Microsoft’s blog: Walmart unveils new generative AI-powered capabilities.

Indian Retailers Embrace Future with 71% Planning Generative AI Adoption in Next Year

A recent EY survey highlights a significant shift in the Indian retail sector, with 71% of retailers planning to adopt Generative AI (Gen AI) within the next 12 months. This move, spurred by the desire to gain new insights and enhance efficiency across the retail value chain, marks a transformative era for the industry.

Currently, investments in AI are projected to jump from US$5 billion to US$31 billion by 2028, reflecting the sector’s optimism towards Gen AI’s impact on business, particularly in customer experience and product innovation.

For more details, you can read the full article on EY’s website: EY Survey on Gen AI Adoption in Indian Retail.

Also, here’s our top takeaways from the survey.

Meet Rufus: Amazon’s AI Assistant Revolutionizing Online Shopping Experience

Amazon introduces Rufus, a generative AI-powered conversational shopping assistant, now in beta for select mobile app users.

Rufus is an expert shopping assistant trained on Amazon’s product catalog and information from across the web to answer customer questions on shopping needs, products, and comparisons, make recommendations based on this context, and facilitate product discovery.

Rufus will progressively roll out to additional U.S. customers in the coming weeks.

For more details, you can read the full article on Amazon’s website: Amazon Rufus.