Mobile technology has brought profound changes to the way people consume digital goods, with life cycles of products and services getting shorter.

As people are constantly exposed to a slew of new information, their tastes change every minute, making it harder to predict their future spending patterns.

This is why companies today need the assistance of artificial intelligence that can help them analyze a massive amount of data in real time and implement advertising and marketing strategies customized to individual demands, said Lee Chang-su, vice president of publisher analytics and insights at Tapjoy, a global digital advertising platform.

“Through real-time machine learning or AI, we can forecast and differentiate between payers and nonpayers, or shoppers and window shoppers,” Lee said in a recent interview with The Korea Herald.

“With this categorized data, we can advertise products and services according to the forecast and find ‘potential whales’ or big spenders within a group of window shoppers.”

Given that the wallets of PC and mobile consumers are totally different with the latter less willing to pay for digital goods in a rapidly changing society, it has become important for companies to drop the old ways of thinking whereby they develop a single product and service that can appeal to the masses.

In the world of mobile games where it has become harder to retain loyal customers, 50 percent of sales come from only 0.45 percent of all Korean users on average, with the majority paying for digital items on the first day of gaming. After three days, it will be hard to see users making purchases. In the PC-based online business, about 40 percent of sales come from 2 percent of users.

Whales are hard to spot, but Lee, also the chief executive of 5Rocks, a Seoul-based data analytic solutions provider, said companies should not ignore nonpayers and simply hope that they go away.

Of 90 percent of nonpayers playing mobile games, 7 percent on average have been found to become spenders, according to data by Tapjoy, which has a forecast accuracy rate of over 80 percent.

“If the data showed that some people will spend more in the next few days, then companies should expose them to ads featuring new additional products and services, while nonpayers see different ads that introduce them to services and try them out,” Lee said.

“The important thing is that we need to have the capability to forecast who will potentially spend more in the future, not who spent this and that much in the past.”

Lee reiterated that forecasting through data analyses has become more important than ever in the era of big data, and using AI to help guide businesses will be necessary in the coming age of the Internet of Things.

Otherwise, companies may have to close down their mobile businesses sooner than expected if they still follow old ways.

Amazon is one of the few digital companies to have successfully developed and used real-time machines learning algorithms and offer customized services, Lee noted.

“The mobile business is harder than the PC-based business. Thousands of people can come and go in droves instantly in the mobile world,” the 5Rocks CEO said.

“We will see a lot more data heading toward us, and this means there is going to be more noise as things and objects get more connected to the Internet. AI can help lower the noise and raise the signal for a wide range of businesses.”

Lee, who graduated from the Korea Advanced Institute of Science and Technology with a master’s degree in computer science and wrote his thesis on data analytics and forecasting, founded 5Rocks with Chester Roh, a renowned serial entrepreneur in Korea’s start-up world. Roh’s start-up Tatter and Company, a software company for blog platforms, was acquired by Google in 2008.

Tapjoy acquired Lee’s first tech start-up 5Rocks last August.

“I feel great about my first deal with Tapjoy,” said Lee.

“We have become busier as Tapjoy and 5Rocks are preparing to introduce a consolidated data forecasting and marketing platform service at the Game Developers Conference in San Francisco in March next year.”

By Park Hyong-ki (hkp@heraldcorp.com)