One Pretty Good Practice I want to see all our clients adopt is tying customer feeback to future behavior.
This involves keeping customer survey records active, and updating them as the customer makes new buying decisions. For example, when a customer who did a customer survey in the past buys a new product, closes an account, upgrades, etc., the new transaction is added to the old survey record.
That makes the customer survey not just a snapshot of the customer's opinions at one moment in time, but an ever-expanding longitudinal data set tracking how customer opinions affect future behavior.
Once this data set exists, it's straightforward to analyze it to determine strategic drivers like just how much loyalty a bad customer service experience costs, which customers might be ripe for a new purchase, and what early warning signs to look out for in a customer who is considering leaving.
These are things most companies can only guess and estimate. Having this hard data allows making much smarter decisions about where to allocate resources, knowing with precision what customers like to complain about and what really drives their behavior.