“Big Data” is an ambiguous term. Sometimes it refers to data volume, sometimes to data variety, and sometimes to data velocity. All organizations have data collection challenges on at least one of these fronts if not more. But even where these challenges are being overcome through technology, the next challenge is in advancing the use of that data for marketing beyond the “Descriptive” level of data utilization.
Most digital marketing organizations with “Analytics” functions are really only focused on descriptive performance analytics, and hopefully also to some extent the diagnostics that explain performance. More advanced digital marketing teams have managed to pass along their data to market research and consumer insights teams to have digital data included in consumer research, insights and segmentation work.
The next level of data application is Prescriptive application. Here the results of the prior level of descriptive work are being refined through trial-and-error in establishing rules for the delivery of the digital experience. This is an intermediate stage to the Predictive application of data, where the rules become a decision tree governed by Bayesian models with probabilities defined against desired outcomes.
The final level of data application is the programmatic, where new data about the performance of models is applied as soon as it is generated, and the Bayesian trees become self-pruning.