|M.Sc Student||Zelikovich Idan|
|Subject||On the Gap between Preference and Performance|
|Department||Department of Industrial Engineering and Management||Supervisor||Professor Ido Erev|
|Full Thesis text|
In one of his last interviews, Steve Jobs, the late CEO of Apple was asked: "What market research did you do that led to the iPad?" Jobs gave a short answer: "None, it's not the customer's job to know what they want". The main goal of the current research was to try to clarify what are the conditions and situations under which Jobs' assertion is likely to hold. Also, we try to shed light on the relationship of this assertion to basic study of skill acquisition and decisions from experience. Specifically, we focus on situations where most people are likely to prefer products and environments that their ongoing experience with leads them to inferior results relative both to the possible optimum and to the less preferred options.
A new experimental paradigm was designed in order to examine participants' pattern of behavior in a simplified setting. In the first stage of experiment 1, participants were asked to state their preference between two tools they could use in a multi-trial search task, based on a full description of the incentive structure. In the second stage, participants used the tools. The results indicate that while most participants preferred the tool that offered the option to employ an optimal strategy, their ongoing experience with their selection led their performance to extremely deviate from maximization. The poor performance is captured by participant's tendency to under-explore hard-to-learn-but-efficient strategies (and to over-explore easy-to-learn-but-less-efficient strategies.
Although participants could compute the optimal (or close to optimal) strategy based on the detailed description, their behavior exhibited a tendency to give up too early when exploring for efficient solutions. Reliance on small samples of initial experience when deciding whether to explore, and subsequently underweighting the chance of finding the efficient solution, captures this pattern of behavior.
These results may also highlight a possible solution to the “consumers do not know” problem: elimination of the option to obtain tools that enable the use of easy and inefficient strategies. However, this solution is likely to fail in competitive markets where users can avoid this suggested constraint. For this reason, experiment 2 examines three possible strategies that can be implemented by entities that operate in competitive markets. The results of experiment 2 suggest that if the market of operation is competitive, investing in producing an edge in the easy-to-learn option would generate the best outcomes.