People used to make those decisions based on gut feel and instinct. They would decide based on past practices or processes—tribal knowledge I call it—that have been developed over time.
But there are organisations that can and want to do the big data analysis because they understand their customers' behaviours and propensity to spend. Or to know what's going on in the manufacturing processes, and to be able to act more quickly.
We're working with Western Digital in Malaysia, which has a long, expensive manufacturing process. Understandably, any downtime will likely to cost millions of dollars. They need fast data on manufacturing to detect that if something goes wrong, the sooner they can analyse it, the sooner they can make a decision upon it. We work with them to figure out such things: they collected the data, they got this historical analysis, and they flow in their realtime data into that historical model so that they can detect sooner when things are going down. For them, being able to shorten the detection and resolution process by a few days is multimillion dollar benefit. It's not just about real-time offerings, it's about giving information faster to make decisions.
You talk to customers who already know what they want. What about customers who want to be like their competition?
Inevitably, it's a journey. You got to look for some business value, and you look for pain points that can be solved, as a first step. Large companies will have large infrastructure, but we're not going to boil the ocean and solve all those problems all at once. We work with them on a strategy to shift from getting the data after the fact, incrementally improving things so that the decision-making process gets better and better.
For example, we work with some regional banks who have different rules and regulations in different systems in different countries. Their problem is, they don't really get analysis of that information until a week later or more. That's a hard problem to solve—lots of systems, lots of processes, and a lot of information. Now if we could get them a little bit of that information a little bit faster, would that improve their decision-making process? Of course, because they previously decided based on guesswork, and if they could get 20 percent of that information in real-time, they've got a little bit better decision-making process. So what you'd do is incrementally improve things and so it gets better and better.
How should companies embark on their analytics journey?
The important thing is to focus on value, and not necessarily the plethora of new technologies that are out there in the market. Don't focus on the big data ; focus on value. The value may not be size of the data; it may be on getting new types of data, or on velocity which is getting data faster. I always counsel my customers, don't do big data project; do a value from data project and focus on it that way. And all of a sudden, it's a different mind shift.
Sign up for CIO Asia eNewsletters.