April 2, 2013

Self-service BI, the Next Big Thing or Disaster Waiting to Happen?

The last day, i join a poll in it-toolbox 'Self-Service Business Intelligence: Good or Bad?' My answer: A win-win for IT and business professionals. Then will Self-Service BI be the Next Big Thing or Disaster Waiting to Happen?
Here is an article, Self-service BI a 'win-win' for IT and biz, by zdnet editor Jamie Yap, who express the same opinion with me. 
Considering the challenge of handling and then analyzing increasing volumes and varieties of data, self-service business intelligence (BI) is a "win-win" proposition for companies, allowing business end users to quickly get relevant insights on their own, and freeing up IT to focus on the related critical matters beneath the user interface such as data governance.
Yet, this "liberalization" with self-service BI would just be a myth, unless end users are truly empowered in reality to analyze data and create analytic reports by themselves without turning to IT, said JY Pook, Asia-Pacific vice president at Tableau Software.

Only then will the IT middlemen not be diverted to generating reports, and can fully concentrate on value-added tasks at the backend that optimize self-service BI, such as ensuring data governance, security, scalability and network performance, the Singapore-based executive said in an interview Wednesday.

Freeing up IT resources is a significant benefit, considering the tide of big data means companies don't just need better analytics capabilities, but also better storage and database management to cope and curate the data, he explained.

Spot new patterns through visualization
According to Pook, data visualization is one of the key factors to help non-IT end users interact with data easily and effectively on their own.

Tableau's software tools, he said, leverage the fact that when data is presented in a variety of ways quickly, as decided by the user, new patterns get spotted. It causes users to form new insights and new questions that need answering, and this freestyle, iterative process continues in real time as the "drill down" analysis goes on.

That users, based on their own needs and discretion, can "see the data move in front of them" as they slice and dice it, contrasts with traditional enterprise BI, said Elissa Fink, CMO at Tableau Software, who sat in at the same interview.
Pook agreed, noting that self-service BI addresses a fundamental problem of traditional BI. The latter cannot keep up with big data as well as the fast-changing business demands and competitive landscape of data.

Traditional BI tools required specialists to do the job and were not easy to use for non-IT end users to use. So they had to send requests to IT for data analysis, waiting between days to weeks for IT to come back with a static report that may not accurately reflect the constantly changing business environment, he explained.

Simple, but not oversimplifying
While self-service BI is used by non-IT users, it does not mean the resulting data analyses will be oversimplified and hence paint a potentially inaccurate picture, the two executives emphasized.

Fink highlighted that self-service BI is meant to be intuitive, and the more experience users have to explore and work with data, the more refined and sophisticated their queries and insights become over time.

Preventing or protecting end users from facing the high volumes of data is doing them a "great disservice", she said. "Why can't someone at the frontline get insights? Let them do it [and learn]."

Pook pointed out it is usually the IT departments, not end users, that need more convincing to implement self-service BI, since they would naturally be concerned with issues like security, governance and compliance.

He added that raising awareness and educating companies of the possibilities of self-service BI is important. Some vendors might describe their product as such, but if it ultimately still required IT to write reports, it does not fulfill the promise of self-service BI at all.

No comments:

Post a Comment