LIVE FROM CEATEC, TOKYO: The most important reason for taking a big data initiative is to create shared value for your organisation.
That was the message from Tomonori Tomura, CEO of the Japan Management Research Institute, speaking yesterday at a conference session on the topic of ‘The dark side of big data and the force of the data Jedi.’
To ensure big data creates value, said Tomura, you first need to have your own definition and own reasons for doing it. “You need a clear strategy and have to be careful not to drift away from your mission.”
With this in mind, a company can then determine how much to spend collecting and analysing data and prevent an initiative from being a wasted effort.
When deciding when to start doing a big data project, he said that too often companies, particularly Japanese ones, make a decision to avoid risk as they don’t want to be too early. “But the danger really is being too late rather than too early.”
Tomura detailed five big data traps for companies to watch out for.
The first is the belief that doing big data will increase your profits. Some firms, he said, believe that the ability to process huge amounts of data means they can be successful and win in business.
He cautioned that more big data by itself is not the answer. “You can’t copy the activities that everyone else is following. You need to allow for differentiation based on your market conditions.”
The big question is how to take advantage of big data to create a competitive edge? “If your focus is too general, you no longer have an advantage. It could be a competitive disadvantage.”
The second pitfall is a tendency to think that big data is a magic wand, much like IT is too often thought of by non-IT professionals as a black box that can accomplish anything.
Tomura, who is a certified big data strategist, said big data analytics can’t show blue ocean direction. “You need to have the big picture in mind and to have a good understanding of what is happening out in the field.”
The third is that many executives think that big data can predict the future. Management may think that big data predictions will be accurate, but he urged data professionals to advise them to avoid this thinking because analytics can only go so far.
The fourth common trap is to think you are wrong and big data results are correct. Again, he said executives should have their ears close to the ground for a feel of what’s really happening in the market. “Are the big data results truly appropriate?”
He noted that, particular with social media, there can be a lot of false signals that appear to be a problem, when in fact there are “wolves making a lot of noise, but when approached they become sheep”.
Companies shouldn’t blindly follow what big data analytics tells you to do, he said, warning the audience to watch out for the eco-chamber effect, which is when an expected outcome is reinforced throughout the process.
The final trap he said is that many companies don’t verify the results of big data exercises. There is the danger that people may pollute the data to end up with a favourable result or that they allow their own biases to impact the results.
Tomura said companies need to have the processes in place, such as an early warning system, to prevent internal actions that lead to the misuse of data. Few companies take the time to consider the background of those collecting and managing the data, he said.
He forecast that in four to five years the majority of data analytics will be implemented through standard, off-the-shelf software packages like other enterprise software programmes, and may be part of a bigger subcontracted project.
This means data professionals will need to have a strategic view and not follow the crowd to continue to offer value.