Thus far, I have covered social media as a source for data, open source as the mechanism for finding software tools, and monetization as the motivation for burning the midnight oil. But how can I have the computing environment for such large data storage and analysis? Cloud technology has offered an attractive proposition to anyone who wants to experiment with marketing analytics and lacks resources. I met a group of students at Stanford University who were conducting a sentiment analytics of Twitter data in the early days of unstructured data analytics of big data. They were dealing with terabytes of data and running sophisticated algorithms to interpret unstructured data and seek Twitter influence with positive and negative sentiments.25 They told me they used Amazon cloud to perform this extensive and valuable analysis on a tight budget without having to buy and build a dedicated infrastructure.

Time-shared computing was popularized by mainframe computers, which were expensive to buy and house. Getting leased access to the mainframe was the best way of using it without incurring massive capital and other fixed costs. Clouds apply the same principle to the modern computing infrastructure. A variety of cloud solutions have emerged over the last decade. Public clouds house data across many corporations or consumers and offer secure access to each. Private clouds house the data within a corporation’s firewalls, but use the cloud infrastructure to reconfigure the environment for each project, thereby reducing the dedicated purchase of computing infrastructure for each project. Cloud technology can be used at different levels of computing infrastructure. An infrastructure cloud offers a computing environment. A storage cloud offers capabilities for storing data as well as capabilities for backup and restore. For example, Symantec and Amazon provide storage clouds for personal computer data backup. An application cloud houses an entire application, such as

As large quantities of public data started to emerge, cloud providers offered ready-to-use solutions for analysis of this data. It is an easy decision to use a public storage for already public data and analyze it for specific queries. Cloud providers offered low entry points and subscription fees to simplify starting costs for analytics. As offerings from cloud-based analytics, such as Coremetrics26, Radian627, and Attensity28 proliferated, marketers began to augment their traditional sources and manual scans of public data using public cloud analytics sources as input. In a study with a large telco, I found a number of subscriptions to competing cloud-based providers, some of them overlapping in capabilities. The price tags were reasonable, and the starting costs were almost negligible.

The pricing models for cloud-based analytics providers have significantly challenged the software and services industry. In a typical capital purchase, software is sold with an upfront fee and an annual maintenance fee, and customization services are also front loaded and may cost as much as or more than the software. For a typical marketing automation program, nearly 50 percent of the cost may need to be incurred in the first year, while most of the benefits may be back loaded. To make these programs viable, most of the large programs were capitalized with a multiyear amortization schedule. The cloud changed the model to a monthly subscription, where the costs are mostly transaction driven and, hence, back loaded, while the benefits may be accelerated through early deployments. In addition, the marketer may choose to cancel the program any time without incurring expensive upfront costs. Information Management reported a story on Adobe management’s decision in August 2011 to dump its lucrative licensing business in favor of a monthly subscription offering called Creative Cloud. The same shift is underway across the technology industry as vendors vie for a bigger slice of the $36.8 billion cloud software market, which will balloon to $67.3 billion by 2016, according to IDC.29

As usage grew, marketers began to raise data security and ownership questions. The market is now maturing rapidly into using a hybrid environment, where public data is being analyzed using public cloud offerings, but its correlation with internal marketing data is conducted in a private cloud. Private clouds provide many of the benefits of public clouds, and yet offer the benefits of data and ownership protection.

< Prev   CONTENTS   Source   Next >