3 Ways to Harness Machine-Generated Data for Marketing

Businesses operating in the ad tech arena are especially impacted by the Big Data explosion. Terabytes, petabytes, and soon-to-be exabytes of data are streaming in from Web logs, email servers, mobile devices, social media networks, and more. And when it comes to this kind of “machine-generated data,” big volumes are only one part of the equation. The velocity at which data is generated and diversity of sources add new levels of complexity to information analysis. Harnessing all this data is essential for direct marketers—it helps businesses more effectively target their promotions and develop increasingly personalized and compelling offers. 

The catch is the “harness the data” part. As the practice of direct marketing evolves, the analytic technologies of yesterday are often not up to the job of extracting useful insight from the massive amounts of information available for mining today. Here are some key considerations for businesses seeking to adapt their data analytics strategies to meet the challenges, and take advantage of the opportunities, presented by machine-generated data:

  • Efficiency is Job One.

Machine-generated data is one of the fastest-growing categories of Big Data. Every minute, every day, devices ranging from smartphones and tablets to web servers and POS supermarket scanners capture and transmit information about what consumers buy, what web pages they “click” on, what mobile technologies they use, and even where they go (think GPS). To analyze all this information, businesses first need to be able to store and process it—an increasingly difficult task given the sheer volume of machine-generated data. Continuing to rely on more servers, more disk storage subsystems, and more IT resources as data volumes continue to grow is simply not sustainable from either a cost or administrative perspective. Marketers need to move beyond traditional row-based relational databases to get the analytic scalability that’s needed. Solutions worth a look include newer types of databases capable of compressing, loading, and analyzing data more efficiently, to distributed processing technologies supported by frameworks like Hadoop. The objective is to enable rich data analysis using fewer resources; otherwise, the insights gained will come at too high a cost.

  • The need for speed.

Whether serving up a location-based coupon or launching a promotion that takes advantage of a trending social media topic, direct marketers need to be able to strike when the iron is hot. This requires an ability to analyze and then act on incoming information in real time—historical reports that return intelligence in weeks, days, or even hours will no longer cut it. But just as insight is needed faster than ever, expanding data volumes can drag down query performance. New approaches are needed to cut through the clutter and extract the truly useful intelligence quickly. Columnar databases, for example, increase query speed by analyzing data only in relevant columns rather than entire rows. Because most queries only involve a subset of the columns in a table, a columnar database has to retrieve much less data to return results than a row database, which must retrieve all the columns for each row. This means that users can analyze much more data, much faster. Other capabilities such as data compression and distributed data loading can also help improve query performance without requiring expensive hardware infrastructure.

  • Flexibility and simplicity.

In a multi-device, multichannel, and incredibly varied marketing environment, the questions that businesses need to ask of their data are constantly changing. Today a marketer may want to understand how a mobile ad is performing across different carrier networks and devices. Tomorrow, insight into responses to a social media campaign might be required. As a result, analysis can’t be constrained by data schemas that limit the number and type of queries that may be performed. This is where many databases fall short. To change an existing query or set up a new one, an enormous amount of manual configuration may be required (such as indexing, data partitioning, etc.), which is time consuming and requires expert IT help. Flexible and simple-to-use solutions specifically built for ad-hoc queries and complex data investigation are important when analytic objectives are front and center. The key is to find tools that don’t require special database management skills or extensive administration to create and change queries. 

Machine-generated data is changing the game for digital marketers. Solutions deployed to leverage this information need to have the right combination of speed, efficiency, and flexibility to deal with the challenges of large volumes of data from multiple sources. Marketers who get it right will be rewarded by a new world of insight that can help them increase the ROI of their multichannel campaigns. 

Don DeLoach is CEO of Infobright.

Related Posts