Double jeopardy is one of the most important empirical patterns of consumer brand
purchase behaviour. It asserts that large brands benefit from having more consumers
who are also generally more loyal. Traditional methods for detecting double
jeopardy patterns in consumer purchasing behaviour rely heavily on the availability
of panel data. Although alternative methods have been proposed, these too require
large quantities of data, making them costly to implement for many managers
and researchers. This study proposes a new method for detecting double jeopardy
patterns that requires only small samples of data. Using the instant coffee market in
the US to test this new method, it is shown that repeated discrete choice experiments
can produce proximate measures to those used as inputs to double jeopardy
calculations. This approach gives researchers an economical and easy method to test
whether a market conforms to double jeopardy, allowing them to keep managers
informed about the properties of consumer purchase behaviour in their markets.