Google, Microsoft, Facebook and a slew of other companies continue to build algorithms that predict when, where and how consumers will make online purchases.
The most recent technology comes from Intent Lab, a research unit of Performics and the Northwestern University Medill School of Journalism. The joint venture teamed up with Microsoft to allow marketers to personalize campaigns on Bing Ads.
The Intent Lab, which focuses on ways to better understand human behavior, created an Intent Scoring Algorithm that helps to explain the abstract thinking of consumers as they move closer to making a purchase, conducted by analyzing the words in search queries and the context around them.
The algorithm also reveals that when advertising copy matches what the consumer thinks, the consumer is more likely to click on the ad. When consumers use abstract search terms like "how" or "what," or concrete terms like “best” or “top,” they are more likely to click on an ad written using similar abstract language.
Algorithms have been used to target search ads for years, but now they strongly recognize intent-based actions and keywords, allowing marketers to more accurately target advertisements.
In another example, Facebook submitted a patent with the U.S. Patent & Trademark Office in November that uses an algorithm to predict household demographics based on image data. The system predicts household features like size and demographic composition based on profile photos, which is quite scary.
“The online system applies one or more models trained using deep learning techniques to generate the predictions,” according to the patent.
An algorithm identifies each individual in the photo, as well as analyzing the text, to determine the relationship between the person or the people in the photo. It also analyzes the profile data and tags associated with the photos.