The concept of Personalized Conversational Shopping is to bring that great in-store experience online. Shopping that is driven by understanding the shopper AND the store. The user experience is like conversing with an efficient, intelligent sales associate.
When you walk into a store, if you are a long-term customer, the shopping associate may know you, your shopping history — what you’ve looked at and what you’ve purchased. In most cases, the shopping associate relies on understanding what you are looking for, asking questions about preferences, making suggestions. And a good shopping associate does that very well.
The personalized intelligent virtual assistant, though, always has access to all the store’s data and can use that to personalize the shopping experience for each customer.
The Data Is Here. We Just Have to Use It.
Everyone talks about being “data driven”. Easy to say. And easy to talk about some data metrics and some marketing actions taken. But the data we have collected on each of our shoppers can do much more. We can personalize the shopping experience to each user in our online store.
Consider the knowledge that the human shopping associate has. An understanding of the store and the layout, the inventory, and the language that the brands use to describe the items. And an understanding of the shopper, what is trending, and the language the shopper uses to describe what she is looking for. This is the knowledge that the shopping associate uses to understand the shopper’s intent, match it to the most relevant items in the store, and engage the shopper in a conversation leading to purchase.
The strength of AI-driven machines is the ability to use huge amounts of data. The intelligent, AI-driven shopping assistant is able to use the same data as the human shopping associate — and much more of it.
Using Data to Drive the Conversation
To understand the shopper, first consider the data we have coming into the conversation. The shopper’s browsing and purchase history. Favorite brands, colors, materials, etc. to drive sorting and filtering? Does the shopper prefer sale items, new items, designer items? Customer lifetime value (CLV) to drive loyalty promotions?
Once the interaction starts, it is a conversation. Each utterance must be considered in the context of the entire conversation. If the shopper is looking at a set of items, is the new utterance additional filtering of that set to a brand? Or asking to see the blue items of the category rather than red? Or has the shopper changed intent and is moving onto a different category — from blazers to blouses? Just like the human store associate, the intelligent shopping assistant has to understand the conversation and respond accordingly.
But understanding the shopper is not sufficient. The intelligent shopping assistant has to understand the store as well. What are the product categories and facets, filters and filter values? What items are in each and the features and attributes of each item. And, since the shopper often uses words to describe items and features that differ from those in the catalog, the assistant must understand the shopper’s natural language as well as store language and map between the two. Product popularity, ratings, reviews & promotions also.
In summary, it is the data that a shopping assistant uses that makes it able to respond intelligently and personalize the interaction to each shopper.
Persona? What is a Persona?
“Persona” is defined as the aspect of someone’s character that is presented to or perceived by others. Simply put, the persona of the intelligent shopping assistant is how your shopper perceives the machine. How smart is it? Does it understand my questions? Can it help me find what I’m looking for? Does it suggest items that I really want?
And, perhaps most important to the retailer, is it consistent with my brand? The image I have carefully crafted for my store. It is embodied in the training given to the human store associates. And, for a retailer who cares about its brand image, the shopping assistant’s persona is a critical aspect of the shopping experience.
Plug and Play? Or Custom Branded?
Several vendors are promoting plug-and-play voice shopping bots. The strong advantage is that a system can be up and running quickly.
However, there are obvious disadvantages to plug-and-play offerings. Integration with much of the store data and tuning of the natural language semantics to the store’s catalog and shoppers is skipped. And the persona is generic — not that of the retailer’s brand image.
A further complication is the language of shopping is not generic and, therefore, a generic knowledge base cannot be applied to all stores or categories. We find that our e-commerce clients each employ unique language to describe their products and services and that the knowledge of that language and the language of their customers must form part of the knowledge base of every intelligent virtual assistant. To train the knowledge base, a combination of supervised and unsupervised machine learning using each retailer’s data is required.
Without this training, the intelligent personalized conversational shopping assistant will have reduce understanding of your shoppers — not very intelligent, not very personalized, not very conversational.
Of course, custom-branding the experience and training the knowledge bases for each retailer comes at a price — it takes longer and is more costly to develop. But, the bottom line is that customizing the shopping assistant to the brand image, store and customer data is what differentiates a generic shopping experience from a great, branded one.
Interested? Next Steps?
The current crisis is having a huge impact on shopping behavior. Those shoppers who had embraced e-commerce increased their volume. And many who were reluctant to use online shopping have now jumped in.
Investing in the shopping experience will be a strong differentiator of those retailers who convert the surge of new shoppers into long-term customers. Andrew Carlisle Managing director and digital lead (retail), Accenture noted, “Once the immediate crisis passes, some shoppers will return to bricks and mortar for some segments. But the fact is, digital platforms are set to become a major driver of growth for all retailers, large and small. This was true before the pandemic, of course. It has now been accelerated to a degree few could have predicted. Adapting to the post-pandemic digital reality is an imperative no retailer can ignore.”
In an early April Accenture survey, consumers were bullish on their increasing use of voice-enabled digital assistants and recommendation apps as part of the shopping experience. Now is the time to start prototyping and trialing new technologies such as personalized conversational shopping to understand what your customers want and the potential impact on your business.
Vioby Personalized Conversational Shopping
Vioby has been working with retailers to develop branded personalized conversational shopping that deliver great shopping experiences for their customers. The intelligent interactive conversations with shoppers use voice and text modalities and integrate our client’s retail data. We leverage our long experience in enterprise-scale voice-interactive systems coupled with our e-commerce marketing automation technology. Our video of voice shopping shows how a shopper can browse large retailer catalogs to find and purchase the items they want.
Are you a retailer or agency working with retailers? We’d be glad to explore partnering with you and your team to define, create and deploy a great brand-centric voice shopping experience for your customers.
Mike Krasner is a co-founder of Vioby, a Boston-area developer of AI-based marketing automation tools for e-commerce retailers and their agencies. Leveraging their extensive background in creating and deploying enterprise-scale voice-interactive systems, Vioby is now developing voice shopping for retailers and brands.