Implicit response refers to the various methods that measure the body's involuntary reaction to stimulus. It can include techniques like Facial Coding, Biometrics and Reaction Time (though not all of these are wholly implicit). The power of these techniques is that they gauge unconscious physiological reality, unedited by the conscious mind.
A famous demonstration in Cal-Tech involved putting people in fMRI machines and demonstrating how different parts of the brain would ' light up', depending on whether they were told a wine was cheap or expensive. Though widely used in quantitative research (Millward Brown has a contract with Affectiva for facial coding) these technologies are still in their infancy.
There are academic debates around what is being measured and what the measures actually mean. There are also practical difficulties with some of the approaches, they can be quite intrusive, small technical variables can invalidate findings, they can be expensive, they can involve very small sample sizes, they do lack diagnostic precision (it tells you that there was a response but not why).
For all that they do provide insight at a level that is not otherwise accessible. Best practice at the moment is to deploy techniques that include implicit response, alongside other more traditional approaches.