The atomic fallacy is the mistake that data points can be used to accurately characterize a person. People are typically more complex than their data points. Hanft pointed to a 2002 Wall Street Journal article about how TiVo labeled people (often incorrectly) based on the shows they recorded with the device.
The interruption fallacy is the mistaken idea that deeply ingrained ideas can be changed by simply inserting intrusive messages. People are annoyed by such messages and they aren't effective, Hanft wrote.
The narrative fallacy represents the idea that segmented ideas can have influence, when they can't, Hanft wrote. People latch onto narrative more than individual ideas and so targeting people on one issue based on a data point is probably futile. It's the reason people say things like “Obama is a socialist” or “Bush is a warmonger.” Those aren't statements about the issues of socialism or war, they are narratives that help people explain to themselves who politicians are as people.
“At the end of the day,” Hanft wrote, “big data can be enormously useful. But its flaw is that it is far more logical, predicable and rational than the people it measures.”
For the full editorial, check out the article on Mashable.com.
