Google Injects Machine Learning Into Analytics App
Google said Friday it will provide a new stream of automated insights in its Google Analytics mobile app, allowing marketers to see key data metrics in a matter of minutes.
Previously, it could take marketers several hours to discover the same data points, Google said. The new feature is available today on both Android and iOS.
Other companies offer analytics products to their customers through machine learning -- IBM and its Watson Analytics comes to mind -- but few, if any, offer similar features on a mobile app. This can be especially useful for marketers who are on the go, about to head into a meeting or unfamiliar with analytics products.
"I think what Google is realizing is people tend to exhibit the same behaviors when they pull reports from Analytics," said Kevin Lee, executive chairman and co-founder of Didit, a full-service digital agency that specializes in search. "Some people just want to see a geography or conversion report. They see people are exhibiting repeat behaviors and are streamlining the process for them."
Over time, Google Analytics will become smarter, and display insights more relevant to the marketer more quickly, the search giant says. The company also integrated a share function so marketers can share data points between coworkers or clients easily.
"Google is taking the initiative and wants to let people know that it is going to be like a brain and put things in front of you," Mr. Lee said. "Displaying an analytics report before someone's 3 p.m. meeting can be handy, but it won't always get things perfect initially. It will have to learn."
Both IBM and Google are excited about machine learning, with the latter saying it wants to integrate the technology into nearly everything it does.
"Sometimes machine learning is just a fancy way of saying 'putting repeat behavior in front of the user,'" Mr. Lee said. "It can be like Amazon putting products you've previously bought in front of you when you visit their site."
Google said it will introduce the machine learning feature into its desktop offering in the near future.