By Teresa Leung
Computerworld Hong Kong
February 24, 2010

HONG KONG - The upward trend of business intelligence is here to stay. BI made it to IDC’s top 10 tech priorities and Gartner’s top five for 2010.In Hong Kong, the BI software market is expected to reach US$19.6 million in 2009 and grow a five-year CAGR of 6.4 percent, according to an IDC report published in 2009.

“From results of a survey we did in 2009, 25 percent of respondents from Hong Kong said they will deploy or will increase current deployment of BI/analytics software during the next 12 months,” said Sharon Tan, research manager at IDC.

Innovative use of BI beyond reporting
Local users have also ventured into innovative use of BI, said Jason Tang, head of Practices and Solutions, SAS Institute Hong Kong. “Some of the local firms have gone into what we call sentiment analysis,” said Tang. “Firms analyze external data and texts from blogs, forums, and social networking sites to learn what customers think about their products and services.”

External data provides additional perspectives on targeted campaigns, he added. “Cost reduction was the goal last year, but this year differentiation and revenues generation are crucial, and will be facilitated by successful targeted campaigns,” said Tang.

Real-time decision making has also become possible for BI users, Tang noted. “For instance, customer service reps can immediately do cross-sale and up sale during phone conversation with customers,” said Tang. “BI can also help HR and logistics executives to automate resources allocations and identify available resources real-time.”

Predictive analytics will be hot next, said Linda Young, brand manager, Information Management, Software Group, IBM China/Hong Kong. “Historical data analysis was what BI tools did. This is what we call trend analysis,” said Young. “But predictions are what businesses need today.”

Predictive analytics are rule-based, ‘what-if?’ analyses that help firms estimate inventories, cash flow, turnaround of goods, and profitability, according to Young.

Industry-wise, companies are also using BI for a variety of purposes, including enhancement of green practices, said Young. “One of our customers, Cosco, deploys BI-based simulation to help control its CO2 footprint and operational costs,” said Young.

Other uses include complaint management by governments, credit analysis by banks, portfolio ROI prediction by investment banks, and identifying criminals and terrorists by police, Young added.

Before the credit crunch in 2008, most BI requirements were from retailers, manufacturers, and players in the financial sector, said Young. “But in 2009, inquiries from retail and manufacturing dropped, while most were from insurance and finance,” said Young. “We expect demand (for BI) to become stable in 2010.”

New requirements
“Users want to complete deployment and be able to use the tools within two to four weeks rather than several months to a year,” said Sinko Choy, sales director, EPM/BI, Oracle Hong Kong.

Integration with mobile devices is another capability required. “Because executives are always on-the-go, they need alerts to their mobile phones,” Choy noted.

As BI tools come with more capabilities these days, CIOs are increasingly concerned about users’ data analysis skills. “That’s why you see universities have added BI courses to keep up with the changing workplace,” said Young.

BI for SMBs
“Large organizations tend to have more BI users, therefore they make up a huge proportion of BI users in general,” said Tan from IDC. “But as there are cheaper express versions of BI tools, uptake in the SMB segment is expected while the economy is recovering.”

Free analytics tools like Google Analytics helps drive the adoption of BI among SMBs. According to Young, Google Analytics is an example of advanced analytics.

“Google Analytics or inexpensive SaaS-based BI that don’t require upfront investment in hardware are good ways for SMBs to learn more about the technology,” said Young.

“[But] these inexpensive tools don’t compete with on-premise BI tools,” said Young. “When SMBs grow larger, accumulate more data, and have more stringent data privacy requirements, they will build their own BI platforms.”

While agreeing cloud or SaaS is a tech enabler, Tang from SAS said it won’t drive BI uptake. “Cloud can make implementation much easier, but ROI and other benefits are the drivers,” he noted.

Businesses need to deal with key issues like risk management and compliance, said Tan. “Since BI contains much sensitive data, companies are cautious when it comes to putting them onto the cloud,” she concluded.

Efficiency Unit enhances complaint management with text miner
Business Intelligence tools don’t just work for corporations. According to the 1823 Call Center of the Hong Kong SAR Government’s Efficiency Unit, the use of a text miner that consolidates information and uncovers hidden relationships through statistical analyses helps improve complaint management.

The Efficiency Unit’s implementation of SAS Text Miner started in March 2009 and was completed in September, said WF Yuk, assistant director, Efficiency Unit, Chief Secretary for Administration’s Office, the Hong Kong SAR Government.

Data dictionary
At a cost of HK$3.3 million, the project isn’t just about implementing software, said Yuk. According to him, Chinese unstructured data is hard for text mining because of the structure of Chinese characters.
“For instance, Quarry Bay (鰂魚涌) in Chinese is a location in Hong Kong. But while 鰂 and 魚 together is a type of fish, 涌 means a river,” said Yuk.

“Another example is the category of construction waste,” said Yuk. “We need to let the system know that sand, rubble, concrete and bricks found in urban areas are related and belong to the category of construction waste. It took us a lot of effort to build a data dictionary.”

In 2009, the unit handled more than 2.6 million calls and more than 100,000 text-based messages including e-mails and fax, said Yuk, adding that about 10 percent of the calls are complaints.
Prior to the deployment of text miner and related tools, staffers needed to dig data out from each of the complaint cases to track progress, Yuk noted. “It also took a week to compile reports on key performance indicators such as abandoned call rate, customer satisfaction rate, and first-time resolution rate,” he said.
Now staffers answering calls input incidents reported into the system, said Yuk, adding that calls are recorded so that supervisors can check accuracy.

The text mining system is accessed only by business analysts and call center supervisors.

Timely response
What the text mining system does is to identify the same incidents as reported by different individuals. “For example, a person tells us a location is full of litter while another says there are many unlicensed hawkers,” said Yuk. “Yet a third caller tells us that snacks sold at the same location look dirty.”

“They seem to be three different incidents, but they are one incident at the same location,” said Yuk. “The text mining system is supposed to bring us to attention to this so that we response timely by notifying the responsible department rather than several departments related to hawker management, public hygiene, and food safety–thus delaying necessary actions.”

The text miner also tracks the progress of cases, said Yuk. “The text miner allows us to create performance reports at the click of a mouse through dashboards as all complaints information is consolidated into a single system, said Yuk.

In addition, users can track ‘black spots’ of different types such as locations where the highest number of dead birds is seen, said Yuk. “This helps us proactively prompt departments to take necessary action.”

Possibly Related Posts:

• Inside Apple’s secret plan to kill the cash register
• Kodak says Apple attempting to buy patents ‘on the cheap’
• Monitor Facebook Status Updates While You Use Chrome
• Apple, Samsung CEOs to Negotiate Patent Dispute
• Apple forces new name for London Apple Expo, Microsoft signs up

 Print This Post