Follow Datanami:
February 4, 2015

Survey Finds Uneven Success With Big Data Rollouts

Despite heavy investments in big data deployments, a survey finds that few respondents have either shifted their operations to production or are satisfied with their big data initiatives.

The key to success, the survey found, is establishing a centralized organizational structure when rolling out their big data and analytics units.

The troubling findings are contained in a recent survey released by Capgemini Consulting, which found that only 27 percent of respondents considered their big data initiatives to be “successful” and a mere 8 percent were “fully satisfied” with their implementations.

Moreover, of the 226 executives surveyed in Europe, North America and the Asian-Pacific region, only 13 percent said they had moved their big data implementation to full-scale production. “Success rates for organizations with an analytics business unit are nearly 2.5 times those that have ad-hoc, isolated teams,” the survey found.

Doubts about current big data deployments are increasing at the same time spending is forecast to soar from $13 billion in 2013 to a whopping $114 billion by 2018. These investments are being driven by the widely held belief among executives that big data technology will disrupt their industry in the next three years, the survey found.

While only 13 percent of respondents said their company was in full-scale production, defined as predictive insights that are extensively integrated into business operations, the majority—35 percent—said they had entered “partial production.” That category was defined as integration of predictive insights into some business operations.

Twenty-nine percent said they were in a “proof of concept” phase in which they were using big data techniques for selected use cases. Nineteen percent said they had budgeted for big data projects and had identified focus areas while 5 percent had no current plans to invest in big data initiatives.

Despite heavy investments in big data technologies, the survey concluded that the biggest stumbling blocks to successful implementation include: “scattered silos of data;” ineffective coordination of analytics initiatives; the lack of a clear business case for big data investments; and dependence on legacy systems to process and analyze big data.

In terms of data processing and management, only 36 percent of respondents said they are using cloud-based big data and analytics platforms. Only 31 percent said they were leveraging open source big data and analytics tools.

Another problem often cited by enterprises was a general lack of big data and analytics skills. One-quarter of respondents cited a big data skills gap as a key challenge in implementing big data.

The survey also sought to determine why some organizations succeed in rolling out effective big data projects. “Organizations that have adopted a centralized structure for the big data and analytics units report higher levels of success then their peers who have ad hoc or decentralized teams,” the study concluded.

Capgemini Consulting said it conducted its survey of senior big data executives in November 2014.

Recent items:

Deployment and Active Management of Hadoop in the Cloud

Sharing Infrastructure: Can Hadoop Play Well With Others?