Views and announcements

Bolstering big data with AI in e-commerce

  • Written by Jordan McDonald,  Staff Machine Learning Engineer at Bazaarvoice.

    Online shopping has seen profound growth since the onset of the COVID-19 Pandemic. This is reflected by the volume of retail sales (attributed to online activity) between 2019 and 2022 in the UK growing by a colossal 58%, according to Statista[1].

    This phenomenon has forced an acceleration of technology strategies for organizations who reside in the e-commerce industry, rewarding those who have the agility to adopt emerging technologies to maintain, or perhaps carve out a competitive advantage.

    Data Is the spark, artificial intelligence is the flame

    E-commerce is a thriving melting pot of data. Data ranging from products, user demographics and social content are just a few that contribute to the evolving landscape of information available. This is data that is ripe for capitalisation through employing a spectrum of artificial intelligence (AI) techniques, which can act as fuel for enhancing product offerings.

    With this premise in mind, this article will explore how innovative organisations are seizing this opportunity.

    Bazaarvoice Inc. primed to disrupt the e-commerce industry with AI

    You only have a few seconds to catch your shoppers’ attention. That means you need to master getting the right content, in the right place, at the right time. Bazaarvoice is the leading provider of product reviews and user-generated content (UGC) solutions, helping brands and retailers tap into wider communities of customers to acquire UGC. 

    One prominent type of UGC is consumer reviews, a source of buying information so widely used, that in fact 79% of shoppers expect a product to have at least six reviews before making an informed purchase[2]. UGC is the insightful, vibrant, trustworthy companion that shoppers are looking for. Bazaarvoice enables brands and retailers to provide shoppers with what they need at every stage of the journey.

    To effectively support our brand and retail partners we need data, figure one highlights the breadth and scale of Bazaarvoice’s network and communities. This is big data at a grand scale, with content that spans images, video, text and beyond. 

    When paired with AI, this data can be applied in several ways, helping bazaarvoice drive increased value for their clients.  

    Contemporary uses of AI in e-commerce

    To appreciate the direction of travel for AI in the industry, it is key to consider the historical benefit AI has generated for companies in the UGC space.

    Sentiment analysis is one of the several techniques used by Bazaarvoice Inc., to enhance our understanding of text based reviews. There are multiple benefits, the main one being that it gives additional transparency to our clients regarding the public perception of their product in the Bazaarvoice network. See figure 2 for an example of how we present this information to our clients, in which we augment the application of sentiment with keywords and intelligently inferred “themes”.

    Content moderation is a key service we offer our clients. Our AI supported content moderation system leverages an ensemble of modeling and scoring tools to detect profanity.

    This is just a taste of the type of AI that is used within the e-commerce industry. While these are highly valuable, how can organisations continue to leverage AI to disrupt the status quo assumptions within the industry?

    Transforming the status quo 

    UGC is an essential ingredient in the e-commerce cocktail. Retailers and brands are very aware of this fact and they work hard to increase both the quality and volume of reviews associated with products on their websites. To do this, brands in the Bazaarvoice network can syndicate content to certain retailers. Syndication is the distribution of a brand’s consumer-generated content, like customer reviews and photos, to the websites of retail partners who sell the brand’s products.

    To illustrate, figure 3 showcases how product ratings and reviews sourced directly from shoppers on a brand's website can be amplified to take advantage of the enormous Bazaarvoice retail network. This is the power of review syndication, one review has the potential to be displayed on numerous client websites.

    Historically, ratings and reviews providers in e-commerce enable review syndication through a process known as “product matching”. In the vast majority of cases this is an automatic process, however for certain client catalogs, manual matching (via humans) is required.

    Manual product matching is inherently a methodical and time-consuming process. Bazaarvoice identified the necessity for innovation in this workflow, overturning the status quo to support our teams of human product matchers.

    The solution introduced an AI model to evaluate the entire product matching backlog. The model leverages sophisticated natural language processing techniques alongside a suite of other data points to either approve a product match or defer the decision to a human. This is a ‘co-bot’ methodology, with the human team remaining in the loop.

    Over time, this innovation will lead to a significant increase in product matches. Ultimately leading to a dramatic impact on the ability of Bazaarvoice to support review syndication.

    Stepping into the time machine

    So far we have delved into the historical and modern ways in which AI is changing the e-commerce industry, but what if we look to the future and speculate on the potential for radical (almost sci-fi!) transformation. The art of the possible includes but is not limited to the examples below.

    Increased prominence of behavioral analytics, applying social media like data points to understand shoppers like never before. This gives the ability to personalise experiences based on usage patterns, personality traits and preferences. 

    Moving beyond sentiment, augmenting review data with emotions such as anger and joy, offering fine grain understanding of shoppers’ experiences and opinions.

    Use a set of images to create a 3D mesh of the product, enabling the shopper to fully explore products rather than being limited to a handful of photos in an image gallery.

    Overhauling how reviews are presented to shoppers with intelligent ordering, powerful search capabilities and even using AI to help the shopper compose high quality reviews. 

    How do you take advantage of this opportunity?

    For Bazaarvoice the answer lies in shaping an alignment of purpose across all of Product Development, making AI an embedded part of our technology toolkit.

    This was initiated through an internal conference known as the “AI Summit”. A two-day event to inspire and propel the organisation towards increased AI adoption. While it’s true that in the past (and present!) Bazaarvoice has embraced the technology, the AI Summit was intended to accelerate the future use of AI within our products.

    The AI Summit was a strategic event to solidify the embrace of AI in Bazaarvoice. Why is this important? We have a belief that our interpretation of the technology tea leaves indicates AI will offer enormous rewards, supporting our ability to delight our customers. 

    Ultimately the pairing of AI within the rapidly growing ecommerce industry acts as a catalyst for innovation, helping Bazaarvoice achieve our business objectives.

    This article first appeared in the Spring 2022 edition of the Sync NI magazine. You can download your FREE copy and sign up to receive future digital editions here.

    About the author

    Aoife is a Sync NI writer with a previous background working in print, online and broadcast media. She has a keen interest in all things tech related. To connect with Aoife feel free to send her an email or connect on LinkedIn.

    Got a news-related tip you’d like to see covered on Sync NI? Email the editorial team for our consideration.

    Sign up now for a FREE weekly newsletter showcasing the latest news, jobs and events in NI’s tech sector.

Share this story