New Technology Attempts to Change Christmas Shopping
As Christmas shopping season approaches, IBM launched a new application, IBM Watson Trend, which predicts and recommends compatible gifts based on demand and popularity. The goal of the app is to not only predict what the trends are, but also show the reason behind it. Watson predicts these trends by going through millions of social media pages, blogs, forums, and ratings. Its results, for example, show that there would be an increase in demand for Minecraft Lego Sets, Samsung TVs, Apple Watch, Nikon digital single-lens reflex (DSLR) camera, and women’s running shoes that are bought for its design rather than its performance. Interestingly so, IBM is not trying to use the information to promote the “rich get richer” effect. Instead, they hope that by showing consumers the various different trends, each person will choose a product that best fits the person they are buying the gift for rather than just buying the product on the “top 10 things to buy for Christmas 2015” list.
The research conducted by IBM’s app, IBM Watson, is fundamentally built on the idea of the “rich-get-richer” rule. Also known as preferential attachment, the idea is that links are formed “preferentially” to pages that have high popularity. In other words, the more well known a product is, the more likely it would appear on the internet via a webpage or in conversation, and the more likely a consumer would think that it is popular (assuming the things said about the product are positive). This natural mechanism, which generates power laws, makes sense intuitively, because when there’s decision-making in the presence of cascades, people have a tendency to copy the decisions of people before them. In other words, if everyone is saying they are buying a Minecraft Lego Set for their 12-year-old boy, then you, as a parent, will most likely buy it too. After scanning the Internet, the app would compile a list of items and comments and ultimately recommend a certain group of items to the app user. However, what complicates this seemingly simple system, are recommendation systems and search tools. They are designed to produce a higher-order feedback effect because it will steer people to process their available options, causing a reduction of the rich-get-richer effect, which is the exact purpose of this app. Similar to many popular recommendation systems, such as the ones used by Amazon and Netflix, IBM Watson tries to expose people to items that are not generally popular but rather match the consumer’s own preferences in order to help consumers explore the wide niche of products that exist in the market. It’s interesting to see that information system models learned in class can be applicable to real world situations and heavily impact the complex social system and human thinking.