kentoh - Fotolia

For data analytics leaders in Asia Pacific, success comes from willingness to fail

Adopting data analytics technologies is something business and IT professionals in the APAC region will have to master if IDC predictions are right

Analytics has long been more than a dalliance for Philippine Long Distance Telephone Company (PLDT), that country's national telecommunications provider and second-largest company, with more than 75 million fixed and wireless subscribers.

In recent years PLDT has implemented a big data analytics environment, delivering invaluable behavioural insights for its capacity planning. It has also spun off a business group, Talas Data Intelligence, which is central to the company’s strategy and was this year folded into the ePLDT digital infrastructure and ICT business.

This year's launch of ePLDT big data platform as a service (BdaaS) – a commercial service that uses PLDT's expertise in big data analytics – marked a dramatic phase in the evolution of analytics in the company, which began three years ago with the goal of building a shared big data resource across the group.

“It makes sense from a shared services perspective to own the data from the start, and to leverage it internally,” said ePLDT associate vice-president MM Yunque. “We're trying to put together a single view of the customer at enterprise scale.”

Although the company had previously operated a “data lake” into which the previous 12 months' raw data was dumped on a regular basis, its analytical capabilities had been relatively limited in the past. The ePLDT programme began building on a Hortonworks analytics platform – based on the open source Hadoop big data platform – that emerged after a three-month implementation to provide the foundation for PLDT's data-driven operational mode.

Since June 2015, PLDT has been ingesting de-identified data streams detailing information including network reporting and subscribers' calling activities, comprising more than 6TB of new data per day. An aggressive recruitment campaign – which lured machine-learning experts, including data scientists from MIT and the University of Southern California, and a Princeton astrophysicist – has fostered data-analysis innovation that has changed the way PLDT does capacity planning.

Intelligent guesswork helps fill in the holes in subscribers' behaviour; for example, if a subscriber is connected to a particular base station in Manila at 1pm and their phone reconnects in Davao, 960km away, at 2:45pm, it's a safe bet that they've just caught a flight. That information may have implications on PLDT's capacity planning in and around airports – as well as in issues such as placement of advertisements, for which PLDT has partnered with a range of third parties who gain anonymous access to the big data insights.

“Location and movement data provide a lot of data for site planning,” Yunque said. “By partnering the scientists with the business – the brand leaders and people who have looked at the data for years – the scientists become insight artists, who can create meaning out of the information. There are a lot of insights that can be gleaned from these types of analysis.”

Climbing the learning curve

PLDT's all-in commitment to big data reflects an aspirational goal for many companies, who have readily embraced the idea of big data as key to taking their businesses to a new level of reactivity and strategic planning.

Often, it's seen as a saviour of sorts for internal business managers who are still struggling to get the insight they want, despite years of broad investment in ERP and business-intelligence platforms. In a recent SAP-CFO Research global survey of 1544 finance executives, for example, 75% said their companies' success would depend on making effective use of big data – but 79% said doing so would require developing analytics capabilities they didn't have access to today.

Just 17% of Asia-Pacific executives believed their businesses had advanced analytics tools, with 50% lacking the confidence to say that their finance functions offered adequate analysis. And 35% of the finance executives presaged growing influence in IT activities – driven by a trend that, the report found, will force them to become far more technologically self-sufficient and stronger “information analysers”.

The strength of this transformation has driven high forecasts on big data analytics. IDC, for one, currently has the APAC market for big data technology and services to grow 32% a year, which will see it triple from $US1.2bn in 2014 to $US3.6bn by 2018.

Yet making the most of big data analytics requires far more than technology: as PLDT, Australian energy giant Woodside and others have found, it requires an investment in data scientists who are empowered to not only scour enterprise data, but to push the insights that data produces into the organisation in new ways.

Take an R&D approach

Bill Franks, chief analytics officer with data analytics giant Teradata, sees the process as being best approached as a new form of R&D – with the same sort of intellectual and financial latitude that is granted to scientists in those environments. “R&D is an accepted practice,” he said, “and that's how you have to judge analytics.”

“Build a team dedicated to discovery that has to come up with a couple of discoveries every quarter that have paid for their existence,” he added. “Nobody worries about the failures they've had along the way, as long as they've documented their failures and how they can support what you're trying to do.”

Read more about data analytics in Apac

“I like to think of it as managed chaos, but it's really targeted risk-taking: it's about not allowing analysts to go after any question they decide might be theoretically relevant because it's fun, but about sitting with the business team and coming up with ideas that are business-relevant.”

PLDT has adopted a similar approach: despite the brains it assembled to drive its analytics efforts, the continuous involvement of the business serves as “guardrails” to focus their effort on business-relevant outcomes, Yunque said.

Interestingly, despite the clear value of data scientists, many companies are still trying to do analytics using their existing talent. A recent Teradata Data Analysis Index survey, for example, found that 88% of companies planned to use analytics but 86% were not planning to hire data scientists.

These results were particularly notable given that many of the biggest obstacles cited by respondents – siloed datasets (41%), management buy-in (39%) and the ability to demonstrate a return on investment (ROI) (32%) – are best addressed by a joint IT-business effort. Furthermore, 71% of respondents said budget for big data projects is not owned by IT. A total of 63% said data analytics requests come predominantly from line-of-business departments.

Given that the business recognises the value of analytics and IT knows it's going to be called on to deliver it, it would seem to be a no-brainer for organisations to have a strategy for analytics. Yet many are still struggling.

Operational performance reporting

Ultimately, the fulfilment of analytics' potential seems to be best achieved when the analytics function is subsumed deep into everyday business operations, as at global logistics giant DHL. There, analytics has been evolving for more than a decade and took a quantum leap in 2010 when, global head of business intelligence Nikolaus Walkowsky said it became clear that internal silos were impeding a global customer view.

“Operations was trying to understand what the sales department was doing and sales was trying to understand what the customer service operation was doing, and it came to this point where we wanted an integrated view of all the business,” he said.

This begat a concerted effort to boost the business-relevant use of analytics that began with operational performance reporting – DHL collects data from around 30 “checkpoints” for each of the 180 million shipments it handles annually – and uses Teradata analytics tools to correlate that behaviour against operational and customer service indicators as well as costing data from the general ledger.

This has helped analytics dovetail with business functions such as sales performance, logistics optimisation, customer loyalty and proactive identification of churn risk factors. “We collect as much data, in near real time, as possible,” Walkowsky explains.

“All this data allows us to keep the quality of our services high, and it's driving all our decisions. Because we now have the facts, we can find unprofitable customers and finally start doing something about them. We are a completely customer-driven company and analytics helps us to live an insanely customer-centric culture.”

Read more on Business intelligence and analytics