Nearly 55 percent of Big Data projects aren’t completed, according to a survey of IT professionals conducted by Big Data solution provider InfoChimps.
By comparison, “only” 25 percent of IT projects aren’t completed overall, InfoChimps found.
So what’s going on with Big Data that more than half of all projects aren’t completed? It’s inaccurate scope, InfoChimps states in a recent project template, “How to Do a Big Data Project.”
The template is designed to help you beat the odds and succeed. It’s written around four steps that should be basic to all projects:
Like I said, pretty basic, right? But that’s how it often is with technology: We get so excited about figuring out how to make it happen that we forget to answer the who, what, where and why of it.
“We’ve seen companies realize the most significant benefits from Big Data projects when they start with an inventory of business challenges and goals and quickly narrow them down to those expected to provide the highest return,” the report notes.
The paper is 17 pages long, and includes advice, questions and considerations for each of the four steps. If you have several conflicting priorities, see the graph that ranks six typical use cases, by time to implement; business value gain; and whether the project had a high, moderate or low impact.
The highest impact came with customer insight engines, which also had a moderate time to implement and medium business value gain. Customer risk algorithms and enhanced analytics for supply chain had a moderate impact ranking and took a low amount of time to implement, but with less business value gain.
Projects that took longer to implement, but reported a moderate impact were new customer-facing software product (which also placed as having the highest business value gain) and new product delivery channel.
The least impressive option: an executive financial dashboard, which took a long time to implement and yielded a low impact and business value gain.
While we’re on the topic of fleshing out your Big Data project, Philip Howard, a data management research director with Bloor Research, recently wrote a six-part series on issues relating to Big Data that’s worth reading.
For the most part, he sees the governance capabilities required by Big Data as “parallel to those of conventional data.” But, you probably won’t be surprised to learn that the exception is integration, which becomes much more complex with Big Data.
“It will be very rare for big data to exist in splendid isolation,” Howard writes in his introduction piece. “Moreover, different approaches to integration will be needed in different situations and these may change over time. In other words, the integration environment needs to be very flexible.”
He also discusses issues related to:
A less-detailed Big Data summary piece is available for those who like to skip ahead.