- Market environment like distribution, advertising etc.
- "Aspirational" responses that cannot be followed through
- Commitment to competitive brands
Of course models like BASES and Novaction make a living modelling purchase intent. In my experience the stated purchase intention plays a small role in the actual prediction. In fact a lot of the predictive ability of these models does not come from survey data at all - simply the economics of distribution.
Jan Hofmyr of Synovate, creator of the Conversion Model and all around market research guru wrote an excellent critique of the predictability in market research. His conclusion is that asking purchase intent is about as useful in predicting behaviour as asking nothing. His solution is to couple brand equity with 'barriers' - such as price, distribution etc.
I have always had a problem with asking consumers to recall these rational 'barriers'.
I think that this excellent piece on the neuroscience of buying gives us a better insight into why purchase intention questions do not work. Essentially the act of buying involves conflict between areas of the brain associated with dopamine/reward (nucleus accumbens) and that associated with fear (insula). These are both emotional responses. Reason plays a mediating role if any at all. You buy when the "gain" emotions win over the fear of loss. The problem with purchase intent is that the brain does not sufficiently experience the emotions associated with fear of loss. It cannot. Hence the overstatement. The fear of loss is primarily activated at the point of purchase. It is very hard to simulate. And a survey certainly doesn't do it.