Matibet observed that the choice of architecture depended on what outcomes were expected. A cloud-based analytic platform may not be relevant for all sectors such as certain manufacturing operations where moving production data back and forth to the cloud may impede performance.
"There are other options. For example, if you don't have time to pull in the relevant data, you could do it on the fly in what we call 'event streaming' where you can process the stream of data and make decisions in real-time. How much real-time analytics is required will determine the kind of architecture you want to adopt. It's very likely you might have a mix."
Lack of data science skill sets
Recognising this is an industry-wide problem, participants noted the ideal analytics practitioner must have good business acumen and be IT savvy. But many acknowledged it was difficult to find the right candidates with this mix of skills.
Some proposed placing ownership of data science initiatives with the business unit. This would ensure there was accountability and alignment with the line-of-business that would directly benefit from the project. Additional technology experts could then resource the operational aspects to meet expected outcomes.
For others, this was a leadership issue. If the business is to adopt analytics and drive innovation, it must start with the CEO. Only then will a culture of risk-taking and innovation power transformation of the organisation.
The ROI Question
Can organisations set a return on investment with analytics? For many participants, information technology is still a cost to the organisation but is slowly evolving as a business enabler.
Matibet suggested that not too long ago, many organisations could not justify investing in CRM software. Why was it necessary to spend on what was essentially a custom database? Yet today, no one questions the need for CRM in enterprises.
"In many ways, analytics is at the same stage CRM software was 15 years ago. Or even Internet access. What was the return on investment when they first appeared on the scene?" he asked. "But with analytics today, you don't have to embark on a 12-month project to see results. You could start small and demonstrate the value quickly. Or fail fast and build upon what you've learnt for the future with a minimal investment."
Phua recommended that companies embarking on analytics should copy best practices from other organisations. "There is a large body of knowledge available to any company interested in analytics. These are proof-of-concepts that have been tested in actual business environments and is the easiest way to ensure you get positive outcomes from your investment."
Sign up for CIO Asia eNewsletters.