IBM Watson Trend app uses big data, machine learning to predict hottest holiday giftsBy Shawn Knight 11 comments
IBM wants to ensure everyone has the opportunity to purchase the gifts their loved ones desire most. To do that, they're using IBM Watson to pinpoint this year's top trends and determine which are likely to sell out first, thus giving savvy shoppers the wherewithal to buy those products first.
The former Jeopardy champion extracts the sentiment from tens of millions of online conversations by skimming 10,000 sources across social media sites, forums, comments, blogs, reviews and ratings. Like finding a needle in a haystack, the app leans on machine learning and natural language processing to highlight patterns and trends that reveal why people are gravitating toward specific products or brands. Insights are then aggregated into distinct trend groups - content, context and sentiment - with each receiving a daily trend score based on impact and momentum.
In the future, IBM plans to add new capabilities to IBM Watson Trend including language and geographic data, further fine-tuning its abilities.
The valuable insight is neatly packaged and presented inside the IBM Watson Trend app where you'll find much more than a simple list of the holiday's hottest products. The app provides a summary of each product trend, a detailed chart plotting its trend score over time, snippets of what people are saying about an item and even a collection of images.
Trends are separated into three categories - consumer electronics, toys and health & fitness. As of writing, IBM says the top trends include amateur photographers looking for entry-level professional cameras, parents looking to buy educational-minded toys for their children and excitement around certain wearable technology.
IBM admits that shoppers could essentially arrive at these same conclusions on their own, provided they had unlimited time and never slept. The company also notes that because it isn't trying to sell anything, the data it provides shoppers isn't skewed by marketing campaigns, advertising dollars and the like.
The IBM Watson Trend app is available as of writing for iOS and on the web.
In related news, IBM is predicting that for the first time ever, consumers will rely more on their mobile devices than desktop computers to seek out deals online. Over the upcoming five-day holiday period (which I assume runs from Thanksgiving through Cyber Monday), IBM anticipates mobile traffic to climb by 57 percent, an increase of 17 percent compared to last year. Purchases made on mobile devices, meanwhile, are predicted to climb by more than 36 percent.
Do you plan on doing any of your holiday shopping via mobile device or will the desktop still reign supreme (or will you venture out into the retail madness)? Let us know your strategy in the comments section below.
Watson image courtesy Clockready, Wikipedia. Shopping photo via Shutterstock