1. They ask consumers/customers their views on things  with which they have no experience, typically new products. This gives results that are rarely accurate, as there is no context against which potential consumers/customers can judge the relative performance of the product or service wth what they already buy, or what they may no longer buy.
  2. They ask people what they buy, (which is worth knowing) instead of why they buy. Why consumers  make the purchase decisions they do is the key data.

Retailers make this mistake all the time, simply because they now have so much information from the scan data about what is bought, down to when, and with what else it is bought, that they confuse this “what was bought” data with the more important why it was bought.