Handling the Information Overload in Marketing With Big Data

November 5, 2015
367 Views

With sophisticated tools to capture every little movement and activity of customers – both online and offline, data is no longer a problem when it comes to understanding customer behavior. Effective marketing strategy depends on how well this data can be translated into meaningful information that can be put up on your marketing dashboard. A recent poll among CMOs of a number of large organizations showed that over 70 percent of those surveyed felt ill-equipped to deal with the data coming at them from all directions.

With sophisticated tools to capture every little movement and activity of customers – both online and offline, data is no longer a problem when it comes to understanding customer behavior. Effective marketing strategy depends on how well this data can be translated into meaningful information that can be put up on your marketing dashboard. A recent poll among CMOs of a number of large organizations showed that over 70 percent of those surveyed felt ill-equipped to deal with the data coming at them from all directions.

One issue obviously is the quality of data. With the rise in the amount of data to manage, the first challenge for marketers is to filter this information to retain only those data points that can be used to make meaningful inferences. According to some estimates, only ten percent of raw data is structured and ready for analysis. The rest is all unstructured in the form of free-form text, images, audio and video that needs to be processed for any meaningful interpretation.

A lot of organizations already make use of techniques like automated content categorization, ontology management, sentiment analysis and text mining to extract patterns and structures within unstructured data. According to David Pope, a Principal Solutions Architect at SAS, organizations are moving towards the use of smart filters that can help identify relevant data from the huge swathes of unstructured data. This not only weeds out unuseful data, but also drastically brings down the computing time. In one study of customers receiving coupons at a grocery checkout, the analysis of millions of customers was brought down from 4.5 hours to as less as sixty seconds.

Such filtering and analysis of data is quite useful in a POS setup where every purchase can be tied against specific customer ID and profile. But how can big data be used in the wild? In recent times, digital offline marketing has taken off in a big way where interactive and engaging data is advertised outdoors through billboards, interactive flyers/menu cards, etc. One way marketers are connecting the offline channel with their online big data setup is through secondary online data.

Take the example of an interactive digital billboard on a popular location like the Times Square. How does a marketer identify the number of engaged viewers on such an ad? One way is to advertise targeted hashtags or short codes that are unique to this location and benchmark this against other controlled spaces to measure total viewership. Other ways include tagging data points with location data and using this information in the big data analytics systems.

With businesses today being able to route every single data point into their technology back-end, the need for sophisticated big data systems to analyze and interpret this data has never been higher. The need of the hour is newer cloud-based systems that can bring down the cost of such analysis so that businesses, both big and small, can cost-effectively use their data to shape their online and offline marketing strategies. What do you think?