Cape Town, 19 Oct 2015
Ambitious 'big bang' big data projects are likely to result in the same disappointments seen in the early days of BI, unless South African enterprises learn to use their data more effectively, says EOH MC.
Jan Badenhorst, service delivery executive at EOH MC, says South African enterprises are being caught up in the current big data hype, but that in many cases, they are not ready or likely to benefit from big data implementations at this stage.
"True big data analysis, where you look at in-depth analysis and predictions based on massive volumes of structured and unstructured data, are a long way off for most enterprises in South Africa," says Badenhorst. "The massive volumes of data being processed in other markets is not in circulation in South Africa yet. In addition, most do not have the resources to fully interrogate big data - or indeed the data within their own companies - at this stage."
Badenhorst cautions that rushing into major big data projects with their associated high levels of investments could lead to disappointment. "We saw the same thing happening with business intelligence," he says. "Companies bought into the hype, and then they were disappointed. They invested heavily, and then said 'the data warehouse isn't giving us the answers that we need'. The reason for this was they were not asking the right questions."
Badenhorst says many local organisations are still behind the rest of the world in terms of BI maturity, which he says is a clear sign they are not ready to gain full value from large-scale big data projects. "In many cases, enterprises are simply referring to historical reports for their decision-making. They aren't drilling down into the data and extracting full value from it. Many have yet to reach a point where they are fully utilising platforms that allow them to interrogate data, make predictions and drive business decision-making."
To benefit from a large-scale big data implementation, enterprises would have to ensure all their historical data and current data was available in a parallel data warehouse, that the necessary analytical skills were available to the company, and that the business value would outweigh the cost of a major big data implementation.
"For most local businesses, the business value would not outweigh the investment at this stage," says Badenhorst. "However, this situation will change in a few years. Currently, businesses can start benefiting from better data-driven decision-making by taking incremental steps. For example, they might start by analysing unstructured social media data in a cloud-based parallel data warehouse. Or they could upskill staff to start asking 'what if' questions of their existing data. Businesses can also start preparing for future big data projects by ensuring that new applications are designed, taking into consideration how the data coming out of them will integrate into other enterprise data."
While many may not be ready for big bang big data approaches, Badenhorst believes this is a good time for enterprises to start preparing for future big data projects in years to come. "Companies need to optimise the data they have, secure more in-house data analytics skills, get their big data strategies in place and consider what investments they will make going forward," he says.
Share