Between the decorating, meal preparation, and visits of relatives, Christmas is one of the most stressful times of the year. Paramount in all our minds, of course, is the dreaded Christmas shopping and all the difficult questions that brings. What are the most popular products? Which one do you get? What will really make that special person’s Christmas extra merry? Gladly, IBM has the answer. Utilising its groundbreaking Watson technology, it has developed an app, the IBM Watson Trends App (also available as a website), that delivers to users clear recommendations on what the must-have products are this festive season. At mporium we love insights, and so we love this! It’s certainly clever, but how does it work and why is it worthwhile?
How does it work?
As a topline overview, the IBM Watson Trends App is a sentiment analyser that assesses tens of millions of online conversations about brands and products every day to work out which are being spoken about most often and in the most glowing terms. This data is then converted into a Trend Score (marked out of 100) and presented to the user as a Top 100 list across three categories (Consumer Electronics, Toys, and Health and Fitness). To add extra context, the app displays the list with capsule summaries and the stories behind those items. Here’s the promotional video:
So what about a more in-depth view? While the basic approach sounds quite straightforward, indeed not all that different to what other companies are already doing, IBM Watson’s going above and beyond. Firstly, it’s not just looking at a handful of platforms. Social networks, blogs, forums, comments, and ratings and reviews elements of websites are all scoured by the Watson technology for “conversations related to purchase decisions” at all stages of the purchase journey: “those who are about to make a purchase, people who are conversing as they make a purchase, and conversations after a purchase decision.” The result is a comprehensive understanding of when people are talking and where they are talking.
IBM Watson doesn’t just evaluate quantity though; it’s also interested in quality. Understanding “the context, meaning and sentiment of tone” of the digital conversations, the technology is not simply acknowledging the conversations exist and adding another score to the tally with every conversation detected, but reading and understanding them. It uses machine learning and natural language processing to work out what’s being said about a given product and adds that sentiment to the overall Trend Score. So, for example, if you’re tweeting your excitement about the new iPhone, while also leaving 2 star reviews on Amazon for a new smart TV, the iPhone Trend Score will nudge up, while the TV’s score will be nudged down.
Why is it useful?
It’s all well and good creating an app that does a lot of cool and exciting things, but it’s a different matter to make one that’s actually useful. Is Watson Trends actually serving a purpose to the people who download it? There’s certainly a lot of information that goes into Watson Trends, and a lot that’s visible to the end user, not all of which may be handy. But that top line recommendation is an utterly key asset to busy consumers seeking inspiration and ideas at this frenetic time of year. Struggling to think of something for your tech-savvy brother? Just tap the app, browse that category, and check out what everyone else is considering and what they’re saying about it.
In that context, Watson Trends essentially becomes a consumer’s Personal Shopper, and that allows IBM to tap into the desire for a curated shopping experience. Product Curation hasn’t gained as much traction in recent years as its close cousin Personalisation, but they’re very similar in terms of what they’re delivering: a bespoke collection of items that helps consumers find what they may want to buy quickly and easily.
Product curation gained significant traction in 2012 when UK department store Harvey Nichols opened a BeautyMART boutique in its store. The pop-up shop featured products curated by two industry insiders and the range covered a vast array of items and brands. The idea was simple: that the promise of a curated list of products would draw people into the shops and persuade them to buy. Consumers wouldn’t have to wander aisles aimlessly trying to find out the best deals, they’d have the expert opinion on what the best products are right in front of them, in the shop.
Naturally, digital curation has followed, with sites such as BeachMint, ShopStyle and ShoeDazzle offering celebrity curations, and others such as Nuji allowing consumers to store items they might like to an account that they can revisit and purchase at a later date. Meanwhile, the Canopy Gift Finder pushes things even further, allowing consumers to answer a handful of simple questions about who they want to buy for and delivering recommendations based upon those answers. With the speed of mobile web browsing making consumers even more eager for efficiency in their shopping habits, curation will only become more significant.
Watson Trends is the next evolution in answering that demand, and if IBM can make it work over the frenetic Christmas period, it has a valuable tool on its hands. By expanding the points of information to include a vast range of people and a vast range of data types, the results the IBM Watson can deliver are far more trustworthy than those from individual curators; they’re almost objective opinion. To be recommended in such a way is worth its weight in gold for a brand, so they too will be watching how the app performs over Christmas, hoping it gains the necessary traction. If it does, another vital sales touchpoint will emerge, and the continued battle for positive online sentiment will become even more important.