 |
If you’ve ever struggled over the timing of a product introduction, or been frustrated when a promotion falls flat, maybe it’s because you don’t really know what customers are thinking. Ying-Chen Lai, an engineering professor at Arizona State University, is attacking this problem with big data. Lai and a team of researchers wanted to know if there are rules for how people decide what interests them. So they crunched the anonymized reading, browsing and shopping data of more than 74,000 individuals who used two popular e-commerce sites and an e-reader service in China. “The common perception is that human interest is really random,” Lai told me. But that isn’t the case. The team uncovered clear patterns in how long individuals stayed interested in movies, music, clothing, toys, science fiction and other categories, as well as in the likelihood that they would return to these interests over time. From these findings, they developed a model for predicting how long a person will stay interested in any single item. Lai thinks the work has implications for marketing, product development, and national defense, among other fields. For example, he says, companies might eventually be able to use the model to inform their plans for developing and marketing a new product. In defense, analysts might use it to predict how long an adversary will retain interest in a particular target. He notes that more study is needed to determine how the model performs when factors such as age or geographic location are taken into account. “Maybe the model we have now is applicable mostly to the Chinese market,” Lai says, because it uses data only from China. “Maybe for different markets, we need to modify it.” What problems could we solve by predicting human behavior more accurately? Author informationThe post Big Data Shows People’s Tastes Aren’t Fickle appeared first on What's Trending. |
 |
|
