Could it be that the 278 million orders that Alibaba received for its "Double Eleven" special offer campaign on November 11 were filled by the just 1,000 or so employees of the online retail platforms Taobao and Tmall? To tell the truth, China's Internet giant Alibaba relies on an even higher number of "data robots" to ensure smooth logistics and quick response to customer needs.

Alibaba finds the answers to many pressing questions through big data algorithms. How should the Double Eleven web pages for more than 40,000 brands be arranged? How should the website's 1 billion products be ranked? Who looks at which product? Which person saw which advertisement? When and where will the next spike in sales occur? These algorithms are based on a gigantic amount of data – more than 100 Pb – mined by Alababa so far. In terms of volume, the user data in Alibaba's treasure trove are the equivalent of 100 million HD movies or 58 billion electronic books. 

"Today's Alibaba is essentially a digital company," Alibaba founder Jack Ma has said. In the past three years, Ma, who has always been accurate in his strategic thinking, readjusted the Alibaba Group's direction, moving away from a business-to-business and service platform toward digital applications. Ma's closest assistant in this strategic reorientation has been big data veteran Che Pinjue.

Big Data Means "Knowing Early"

While many readers might not be familiar with the term big data, its applications have already penetrated our lives for some time. Stock markets, where numbers rule, are a classical example for the use of predictive algorithms. Technical analysis indicators such as the daily, weekly or monthly K-line, Bollinger Bands or BIAS formulas all forecast the future based on past stock movements. Today, companies and individuals use digital analysis of many more areas such as people's behavior, experiences and lifestyles to predict trends.

"In the era of big data, the core feature is 'using data to find opportunities.' You can often hear entrepreneurs say 'If only I'd known earlier.' Big data uses limited data to predict the unknown part; it predicts the 'full picture' from the 'partial picture,' enabling companies to know early and strengthening their ability to predict their luck in the future," remarks Che.

The other reason why companies are so keen on mining big data is that it is the only way to understand the data of competitors.

In the past, companies used to dig up data in-house; they analyzed them and then made decisions because their own corporate information could be easily and reliably obtained. In contrast, the data of industry competitors could only be gathered and assessed indirectly. When several variables are in play at the same time, nothing works but big data analysis.

For instance, online retailer A concludes from his own data that a 30-percent rebate on weekends will result in a 10-percent increase in transactions. Yet retailer A wants to understand why retailer B has been able to boost sales by 20 percent during the same period. Can this be attributed to the fact that retailer B posts more ads? Is it because his advertising copy creates a leisurely vacation atmosphere? Or is it that consumers, who often use mobile devices with small screens on the weekend, are more likely to place orders with retailer B because his product images are easier to navigate?

"A good question has its answer inside," notes Che with Zen-like equanimity. It is because the goal of the question does not lie in its answer but rather in discovering a new way of thinking. Asking a good question helps create a clear train of thought and focuses the direction of one's search in the vast sea of digital data by "shooting first, targeting later."

Stepping Back to Ask the Right Questions

Che cites an example: An organic farm analyzes its soil temperature data and fertilizer application records. What should be taken into consideration is how the use of these figures could be refined in order to boost sales. Che suggests that the farm should go beyond its own data and place another focus of its data gathering on the market side. The farm could predict future demand and make relevant crop planting suggestions if it compared the amount of unsold food that is discarded in the cities every day with the respective wholesale and retail prices.

"Why should you take the much longer route? Because you need to observe [a situation] from many different angles to avoid blind spots. You need to practice asking questions. The most important thing is maintaining curiosity," Che remarks.

As a pioneer in big data research, Che is keenly aware that digital innovation is a corporate must-have no matter whether companies accept it or not. "Datamation is a life-and-death issue for any enterprise. However, companies also often complain that good talent is hard to come by. Probably no one has imagined that the biggest obstacle to the datamation of originally conventional companies is a new way of thinking about data," Che points out.

Do Taiwanese companies stand a chance of catching up with the big data trend? Tony Chen, founder of customer relationship management (CRM) system brand MIGO Corp., observes that, "the first problem that Taiwanese enterprises encounter is that they do not have any data." Since only a handful of Taiwanese companies possess a large amount of data, the vast majority is forced to look for outside data or to integrate in-house data with outside data. "Companies do not quite understand the concept of big data either," Chen says.

Fred Chiang, head of big data solution provider Etu, a Systex Group company, says that Taiwanese companies are not familiar with the big data concept and know even less about the application of big data methods. "Currently we are at the stage of practicing 'asking the right question'. To become adept at using big data will still take a very long time," Chiang predicts.

Half Human, Half Ghost

Che, who has spent most of his life in the company of programming codes, does not look like the typical software programmer. During his career, he has held diverse jobs ranging from IT head at Britain-based multinational bank HSBC, product manager at Microsoft's Internet service provider MSN, and chief product officer at Hong Kong online marketplace DHgate.com. Yet as soon as Che begins to speak about data, he becomes a preacher, habitually citing example after example, dispensing well-intended advice and explaining complex issues in an unhurried manner.

Born in Hong Kong and educated in Australia, Britain and the United States, Che worked in Silicon Valley before returning home to join the Chinese Internet industry. Che speaks Mandarin with a Cantonese accent, and uses mainland Chinese colloquial expressions interspersed with English words and phrases. In writing, he switches between traditional and simplified Chinese characters. Che's diverse background makes him a good communicator.

Big data work means researching new lifestyles, including psychological, social and anthropological aspects. Therefore, the teams involved must be very diverse. "There has always been a big gap between the product operations department and the data department. The good thing is that I am half human and half ghost, so I understand everything they say," jokes Che.

Even an easygoing type like Che needs to work on having "two brains". He needs to practice shuttling between the commercialization of data and the datamation of business. Sometimes he needs to ask questions from a macroscopic viewpoint, and sometimes he needs to find logical fallacies from a microscopic perspective, always alternating quickly between zooming in and zooming out. Che sums it up with a laugh: "All things on earth are big data."

Translated from the Chinese by Susanne Ganz