Some first-wave big data projects 'written down' says Deloitte
Not enough data a problem for some, while Hadoop integration has proved tricky
Consultancy outfit Deloitte reckons early big data projects have had to be written down because they failed, thanks in part to a “buy it and the benefits will come” mentality.
The source of failure was sometimes difficulty making open source software work and/or integrate with other systems, Deloitte Australia's technology consulting partner Tim Nugent told The Reg. Such failures weren't because the software was of poor quality. Instead, organisations weren't able to make it do meaningful work because they lacked the skills to do so. Integrating big data tools with other systems also proved difficult.
The attempt to develop those skills while also staying abreast of the many changes in the field of big data proved hard for some, Nugent said. Happily, vendors and services providers have since come up to speed and are making it easier for organisations to adopt the likes of Hadoop so they can get big data's enablers working.
Regulation has also made big data projects tough, with Nugent saying “organisations that have pushed the boundaries on personal data use and retention have experienced difficulties in responding to regulatory forces and government scrutiny.”
The business end of big data has also struggled, Nugent said, because those who hoped to consume insights from big data didn't know what they were looking for, or how to turn output into action. Many organisations also assumed that the data they had collected would be sufficient to produce the promised searing insights.
“The orientation now is about bringing data into the businesses,” Nugent said. “In the past it was focussed on integration and consolidation.”
Nugent's opinions are based on observations gleaned in Australia, so perhaps the experience is not universal. If it is, or is in the same postcode as universal, it's bad news for big data advocates because what Nugent has described runs counter to most big data hype, namely that insights are on waiting to be disinterred from the data you already have and will quickly be impactful.
Another of Nugent's insights offer happier news: IT teams, he says, have stood up to be counted and are working happily alongside the business people who want to wield big data. And big data-centric tools like scale-out storage have worked as advertised. ®
Sponsored: Benefits from the lessons learned in HPC