Managing Data-~Driven Change: A Model of Unintended Deviation
Data-driven change in hospitality gaming is desirable because of the opportunities to leverage untapped sources of rich and abundant marketing data. However, change has been difficult to implement as indicated by a lack of widespread adoption. Some have attributed these difficulties to cultural, structural, and other generic factors but these explanations fail to explain the root dynamics of data-driven change.
In this dissertation, it is theorized that data-driven change requires a particular form of social interaction, which are called analytical bonds (AB). The suggestion was that there is a sender of an analytic deliverable and a receiver that makes a decision, and that the sender and
receiver do not hold the same level of formal power. To study these bonds, a broader qualitative design of grounded theory was applied to interview data from industry leaders. The resultant model of unintended deviation (MUD) explained that the difficulty arose from a deviation from the company’s intended path towards data-driven change. This deviation stemmed from survivalistic overprotectionism—an organizational behavior where actors work in a self-interested manner and provide a facade of informed analytics by exuding a belief in
data-driven ideals to top management. This was shown to work in opposition to behaviors that are facilitative of analytical bond growth.
The construct of analytic facilitation (AF) was also introduced and represented the activities and behaviors that builds ABs and different individuals demonstrate AF to varying degrees. An embedded instrument was used to measure for AF. As a pilot study, AF was distinguished from knowledge sharing via a factor analysis, and report usage showed initial promise as a means to measure AF via a regression model. Applying the MUD and considering AF, AB, and existing understandings on silo effects, the theory can help leaders to make informed decisions around analytics that prevent deviation and provide corrective action.