This column is by Rohan Manthani, VP – Products, Streamoid Technologies
Customers shopping online struggle with discovery, then deliberation, and finally transaction; retailers must reduce friction in all of these aspects of the shopping experience.
A perfect storm of technological events is giving birth to new interfaces for interaction with retailers and brands. The widespread use of messaging apps such as Kik, Whatsapp, and Messenger, combined with vastly improved Natural Language Processing (NLP), heralds the rise of the bots.
Studies have found that users are using messaging apps more than any other kind of app. In fact, the last decade has seen omnichannel retailers shift their focus from brick-and-mortar to digital to mobile. All signs indicate that the next step is conversational commerce with personalized, message-based interaction that can guide users at every step of the buying cycle.
Fashion retail faces a set of unique challenges in user behavior. It is difficult for customers to keep up with the latest trends and discover apparel that suits them.
Currently, online stores show hundreds of products with plenty of information but little guidance. In addition to this, users are bombarded with fashion advice in the forms of celebrity outfits in the news, marketing emails from several retailers, and images on social media. The tremendous variety of content hampers a shopper’s purchase confidence and consequently the conversion rates on the retail site.
Adding to the list of problems, fashion discovery is often reduced to rudimentary text-based sorting and searching. How can colorful fashion apparel be condensed into a keyword? Product images, on the other hand, are a much richer source of information. With a visually intelligent chatbot, users can describe the product in everyday language and see relevant results based on the image, even if the retailer did not ascribe the product with correct keywords. Alternatively, a user could simply upload an image from social media and find similar products!
Using product images to enhance discovery utilizes visual parameters and exposes minor nuances that text-based search does not. On top of making search more natural, visual image processing can also recognize the types of images, products, outfits, and models that the customer is interested in. Detailed visual analytics can be used to further improve retail conversions.
In order for a user to find, say, a blue striped top using a brand’s mobile app, she must first download the app and learn its interface to understand its sort filters; however, even after all of this, she might still not be able to find the top she is looking for. The learning hurdle multiplies for each brand’s app that she downloads. A standardized message based chatbot avoids overwhelming the customer by offering a familiar layout and a specific response to her query. In this scenario, she would just send a message to the brand’s Facebook page and get an immediate response with the relevant tops shown in the chat window.
Existing retail websites do not allow for complicated search queries. “Show me dresses with orange vertical stripes.” Users are constrained by the limits in the design of the website, which is difficult to improve on continuously. Users could filter by color and then scroll through a list of products to find what they want. With a visually intelligent chatbot, they would just ask it this same question in everyday language and get an immediate response with the relevant products.
Several obstacles lie between discovery and purchase that comprise the deliberation phase. Here, customers are primarily influenced by impartial guidance from a nonjudgmental source. A bot fills the role of a trusted operator who advises them while shopping (in the form of text, images, video, tweets, or other rich media).
To resolve doubts prior to purchase, sales assistants in shops have a product first approach—they attempt to sell the product without knowing anything about the customer. A personal stylist takes a more refined approach by examining the customer’s existing wardrobe and her personality before making recommendations.
A visually intelligent chatbot replaces the human stylist and takes personalization a step further. It is cheap, fast, on-demand, scalable, and accessible. By understanding a customer’s existing wardrobe or past purchases, using visual intelligence, a fashion bot can evaluate an endless number of outfit variations. It can suggest the best outfits based on real-time fashion trends, paired with a deep understanding of the user’s browsing and shopping habits. More importantly, the bot can act in the background, subtly alerting users with a notification or text message whenever they encounter a product that goes well with their existing wardrobe. The chatbot, a persistent styling assistant, helps customers whenever and wherever they shop. On the other side, it helps brands shorten the deliberation phase and improve conversion rates on their products.
Typically, customer service teams do not have fashion expertise and cannot help their customers with purchases. In retail stores, a customer can speak with a sales assistant for help. Until today, there has been no digital equivalent—no one to help walk customers through the online shop and help them find what they are looking for. Retailers often find shoppers asking the same fashion questions to customer service representatives who are unable to handle style-related requests with responses that fit the brand’s ideology. A chatbot programmed to deliver responses that suit the brand could augment or even replace the customer service agents.
Besides helping customers discover products and make the decision to purchase them, bots can also drive transactions by allowing the purchase to be made directly from within the messaging platform— even controlling Add to Cart and Save to Wishlist functionality.
Having access to any information needed on a single page improves customer retention because it prevents them from skipping between multiple open tabs while losing interest in the initial product. With the chatbot, customers never have to leave the page they are on. They can ask fashion-related questions, search other products with precision, view their past purchases, and see outfit combinations—all from a single chat window on the page.
The single focus and noise-free environment of a chatbot is a great opportunity to promote relevant sales and discounts. With deep, personalized information on what a customer owns and wishes to own, the bot can make a compelling case to drive user purchase. Additionally, the volume of sale-related communication on conventional media is currently very high. Brands can break the clutter of email and social media posts by using a bot to advertise on popular instant messaging platforms, reaching customers directly with a tailor-made message.
The Bot as a Service (BaaS) can offer personalization at scale without compromising on the brand’s identity.
It strengthens relationships between the brand and their customers by transforming unidirectional marketing outreach into two-way dialog with customer input. Retailers can use bots to aid discovery by using visually intelligent search. These same bots help customers and brands during deliberation by improving purchase confidence with a personalized shopping experience. Bots are also able to optimize and drive transactions. The combination of visual fashion intelligence and chatbots personalizes the shopping experience for customers—helping them find what they want and buy it quickly and efficiently. For the retailer, the integration of this chatbot is a painless process that leverages existing channels and data to improve future sales performance.
We are already seeing many bots being successfully introduced for news, weather, friendship, and other industries. Kik’s nascent bot platform recently exchanged 2 billion messages. Bots are here to stay, and absolute personalization, the holy grail of every marketer, is not a distant dream anymore.