I think one of the great problems in marketing academia is that we spend a lot of time thinking about our models and very little time on our data. We have increasing clear views of how things connect up but we don’t really know what it is that we are connecting up. Deciding what objective you think is being pursued is the key part of an investigation to my mind. So how do academic marketers choose their objectives?
Dependent Variables Matter
Academic marketing researchers must decide what outcome to measure. This is the dependent variable (DV). This is the thing that depends on something else. So profit might be your DV and the marketing tactic used the independent variable.
The outcome assessed is surprisingly variable in marketing. Sometimes this is a data issue. the academic only has awareness data so uses awareness as the DV. This is often a reasonable response to an imperfect world. Still, it is easy to forget that that increases in awareness are rarely what the marketer is ultimately trying to achieve. Marketing academics should bear in mind that just because they can measure it doesn’t make it the perfect choice of DV.
DV Mining
More sinister is a problem that arises when some DVs work for what the academic wants to say but others don’t. It is simply wrong when a researcher has access to multiple outcome metrics but only reports those metrics that support their conclusion.
How Do Academic Marketers Choose Their Objectives?
Katsikeas and his colleagues looked at outcome measures in academic marketing research. The metrics are all over the place. Not only were a multitude used but little justification was typically given for the choice.
The authors noted:
…significant problems in how performance outcomes in marketing are commonly conceptualized and operationalized.” ….”researchers have used a range of often ill-defined measures of performance.
Katsikeas et al. 2016, page 1
A major problem is assuming that there is an underlying concept of performance. This is when scholars assume that measuring awareness, retention, accounting profit, or market share are pretty much interchangeable. The problem is that the correlations between the various DVs are typically low. For these measures to be interchangeable they must measure similar things. Low associations mean the metrics must be measuring different things. You can’t just pick which one you fancy and tell the same story.
The upshot is that marketing academics should be careful about what outcome measures they use. I wholeheartedly agree.
For more on academics using the wrong dependent variables see here and here.
Read: Constantine S. Katsikeas, Neil A. Morgan, Leonidas C. Leonidou, and G. Thomas Hult (2016) Assessing Performance Outcomes in Marketing, Journal of Marketing, 80 (1) 1-20.