Saying people do not know what they like is clearly shortsighted. In large aggregate there are trends and lessons to be learned from consumer data and it is not

Saying people do not know what they like is clearly shortsighted. In large aggregate there are trends and lessons to be learned from consumer data and it is not all noise, nor is it to be taken with a grain of salt. Analytics is a tool, along side vision, design and innovation. If analytics were simply asking people 'do you like this color better than that' then it would not have much viability. However by segmenting sufficient numbers of people into groups and observing the behavioral differences makes for valuable data. Your argument is similar to the criticisms leveled at Twitter as just a bunch of people tweeting about what they had for lunch. There is enormous semantic data in this 'noise'. For example, while people are tweeting about hamburgers they are not tweeting about an Earthquake in Silicon Valley, or a flood or shooting or.... Sometimes noise means something more than it might given only a cursory glance.