The fast machine-learners in China

While AR and VR have grabbed the headlines, 2016 was a year where both agencies and brands in China harvested the power of data and marketing automation, opines Calvin Wong, director of research and analytics for Asia, at Golin.

Firms have been “combining traditional mobile data, public datasets and new data points from smart IoT devices to improve the targeting and resonance of their message”.

Big data and its corresponding analytics offer the ability to gain customer insight by peering through vast volumes of information. This is being deployed by both brands and online platforms.

For Mondelez, data analytics enabled it to add a new customer experience to the company’s famous Oreo biscuits, says Ganesh Kashyap, Mondelez’s director of ecommerce in Asia. “Customers expect products that are differentiated. The question is: how do we deliver that differentiation in our snack food and deliver a compelling experience?”

More than 10 percent of all Mondelez biscuits in China are sold online, so Kashyap and his team came up with a special offer that allowed consumers to personalise Oreo packets online. 

Kashyap brought on board three Chinese artists who created distinctive designs for each Oreo packet. The team then set about creating a campaign to build buzz and awareness, combining online shopping behaviour data from ecommerce partner Alibaba with demographic data from its own social media platforms to ensure the effort was “fairly targeted” at millennials. 

“They were consumers who were already shopping in this environment and who we knew would be interested in those artists’ work,” Kashyap says, adding that the result was “a sustainable lift” for the Oreo range, with sales of the personalised options in China up by 40,000 units over three days.

The campaign is only one example of many ways marketers are learning to capitalise on self-generated data and work with China’s internet giants, Baidu, Alibaba and Tencent, who have finally begun to inch open the door to their walled gardens.

These big players are “global standouts” in their use of big data, according to James Chiu, executive creative director at Razorfish China. However, due to the way these big players have siloed themselves off, opportunities have only recently opened up for agencies to enter into strategic partnerships.

Spending on data–related marketing services in China is expected to more than double over the next year, according to a forecast from Analysys International Enfodesk, to reach RMB27.7 billion (US$4.1 billion) in 2018.

“China does not have legacy issues and can therefore leapfrog straight to advanced data analytics and smart technologies to track and anticipate consumer behaviour,” explains Jessie Qian, partner and head of consumer markets, KPMG China, in a report the company released recently.

And at a break-neck pace, too.

Tencent last month launched a new WeChat ad matching system, which a Deutsche Bank analyst note describes it as a “major set of ad innovations”.

Driven by past data accumulated on WeChat official accounts, the new system allows both advertisers to view and choose specific accounts to advertise on, and improves targeting “significantly”, according to Deutsche. WeChat provides granular details such as ad display date, page views and cost, for both advertisers and account owners.

Baidu, meanwhile, has a deep-learning system which has been developing powerful speech recognition for its voice search after ‘listening’ to hundreds of millions of audio queries. The system boasts an impressive 97-percent recognition rate, and was ranked second in MIT’s top-10 list of technology innovations this year, with Baidu the only Chinese brand.

Hua Zeng, GM of key accounts at Baidu, highlights a Suning campaign on Baidu that triggered a half-screen promotion whenever one of 270,000 Olympic-related terms was entered during the recent Rio Games—many of them voice searches in harder-to-decipher Chinese accents and dialects. As for the electronics brand, it recorded 150 million times of brand exposure for itself.

Another example of how data-driven marketing plays out in the field is a location-based campaign Walmart China deployed in 26 cities of various tiers to grow in-store traffic and sales during the Lunar New Year in February. Technology from location-based specialist xAd, targeted shoppers within 2km of Walmart stores by partnering with mobile apps such as WeChat, Weibo, Dianping, Moji and QQ. 

According to Bill Wang, xAd China’s head of sales, this campaign reached 50 million people, generated 10 million mobile engagements and drove nearly 2 million shoppers to Walmart stores, boosting sales by 4.1 percent.

Then in the B2B niche, there is online wholesaler DHgate.com, which has built its business around retaining and understanding the data generated by its own system and third-party service providers.

It uses those data to create in-depth profiles of users — everything from age to preferences, behaviour, history and product interest — to provide buying recommendations.

“We gather a massive amount of data points and assign them to machine-understandable behavioural tags which are then presented to buyers as recommendations,” said Ella Yang, the website’s head of business intelligence.

For sellers, the company provides a specialised data product called Data Wisdom, which helps them understand their buyers better, Yang says. It measures performance by the click-to-conversion rate. “If clicks are low, so is efficiency.”