Value of data in focus at DMA China: Liveblog coverage

Being smarter with data: Just do it

Derek Kwok, head of platforms and audience insights at Google Greater China and Korea led a practical session for brand marketers, intimating that there really is no excuse anymore for not making better use of your data.

He pointed out research showing that leading marketers are simply smarter with data:

• They’re 1.5 times more likely to have an integrated marketing and advertising technology stack.
• Organizations that leverage customer behavioural insight outperform peers by 85% in sales growth and 25% in gross margin.
• Organizations with integrated marketing and advertising stacks are 1.4 times as likely to be using customer-level data to segment and reach individuals, compared with marketers without fully integrated technologies.
• Leading marketers are 50% more likely to increase investments in capabilities like machine learning to predict customer needs.

Derek Kwok, Google Greater China & Korea

Kwok said that while the task of connecting data may seem daunting, it’s better to just start with your priorities. “You don’t need to connect all the data—just start with three areas,” he said, recommending that marketers ask themselves these questions.

1. What are top three data silos to be integrated to get a better view of the customer journey?
2. What new customer insights could we unlock by combining our customer data?
3. How can we connect audience insight to media activation and drive better performance?

In response to the latter, he presented a case study Google did with Sands China, which not surprisingly highlighted many of Google’s tools to help marketers:

• A market analytics dashboard was created using Data Studio to visualise the data.
• They leveraged Accelerated Mobile Pages to double the speed of load time.
• They used data-driven creative to increase return on ad spend by 98 times.
• Finally they set up a predictive model through machine learning that delivered twice the bookings compared to other channels.

Learning to do this, Kwok says, “is not as difficult as you can imagine”. While he stressed the importance of using your team, he said more of data science can be self-taught, using YouTube videos, of course, among other reference sites.

Ad fraud requires business solutions, not just tech

Beyond technical solutions, tackling ad fraud requires a business solution where clients pay more and publishers are pressured to clean up their inventories.

Speaking during a panel discussion, Lawrence Wan, managing director with Amnet, said that although costs will go down when technology becomes more automated and scaled, changing KPIs may push costs back up as advertisers want to focus more on better-quality publishers.

“The cost variables will merely focus less on the solutions to resolve it, and more on the media we are buying and the results you can get from it,” said Wan. He added that invalid traffic remains the key problem in China, although brand safety and viewability came up in discussios earlier this year when the YouTube brand safety controversy erupted.

L-R: Lawrence Wan, Annie Wu, Robert Sawatzky

Annie Wu, business consulting manager at Miaozhen Systems, said the fragmented nature of publisher networks in China has been a huge hurdle, and the accreditation parties have only been able to track through device-ID data, plus geographic data and browser data. “The business solution is…whether to push the publishers to embed the JavaScipt tag, the same as required for publishers overseas,” she said.

Wan pointed out that clients must also be willing to pay more if they want to demand compensation from adtech platforms when traffic and viewability standards do not fall within the range stipulated in contracts.

Meanwhile, Wu said it is wrong to assume that China has a higher occurrence of IVT (invalid traffic); she said it has around the same level as other markets. “In fact, we find that the invalid traffic is coming from Southeast Asia due to cheaper WiFi there and stricter device ownership regulations in China in recent years,” she said.

She pointed out viewability is a new threat for emerging platforms such as OTT, and clients and publishers must be compelled to use viewability as a currency in the media contract.

“Last year, the clients are talking about quality and invalid impressions," she said. "But this year, the clients are talking more about quality and invalid clicks. Traffic is also more expensive.”

The discussion was moderated by Campaign’s head of content, Robert Sawatzky. 

AI knows the audience best

The benefits of an AI-driven news aggregation platform such as Toutiao offer advertisers higher clickthrough rates (CTRs) and lower conversion costs, said Yujie Xu, director of performance-based advertising for the company.

He acknowledged that a media site may not have the highest ad volume, but the platform’s model of customised content and content-focused ads draws the audience rather than commercial ads. “Better user experience leads to higher impressions," he said. "One of the pillars of our stratagies has been managing audience to find the content that matches the suitable audience.”

Having said that, Xu said first-party data is equally important and has the highest impact in driving conversions.

“As a platform, we have a lot of data on our users. While we are using our own data, we want to bring in the first-party data,” said Xu, adding that data consolidation between advertisers and the platform’s audience is a key focus of the platform.

Launched in 2012, Toutiao also runs TopBuzz, an English news app, and has expanded to user-generated content through acquisition of social video app Musical.ly and Flipagram this year.

Spring Festival an emotional rollercoaster for Chinese consumers: Alimama

Feelings of joy and celebration around family reunions in hometowns are popular motifs around Chinese New Year campaigns, but actual consumption patterns are driven by many more complex emotions and behaviours. 

Silver Sun, general manager of planning at Alimama, debuted new research on the behaviour patterns and feelings of urban middle class consumers during Spring Festival holidays. 

Silver Sun, general manager of planning, Alimama

The study found that CNY celebrations were a time of anxiety for many as young professionals return home to be quizzed by friends and family about their professions, salaries and relationship status. Search terms like “forced marriage”, “rent a girlfriend” “anxiety” and  all spike during this period.

Based on the searches of 70 million core urban middle class consumers as well as more than a thousand questionnaires and qualitative interviews, the survey found some of these anxieties are also channeled into escapist shopping behaviour.  Not only is more spent on home movies, entertainment and online games but online browsing as an activity also rises, Sun said.

See full story.

Data = Revenue

A panel on effective data monetization found that big data is not necessarily better data.

Susan Ren, data and digital development senior director at Unilever noted traditional companies have been spending a lot of money to collect, analyse and store data and “they start to become IT companies.”  

But even with the help of AI data mining, Ren noted that if you don’t see value in the data then machines won’t find meaning in it. “There must be a benefit to the customer or big data will become a big bubble,” she said.

L-R: Max Yin, FT Chinese (moderator), Jack Yu, Reload, Susan Ren, Unilever

So the two V’s are important, Ren said: value and velocity, the latter in reference to the speed of business brought online to harness data.

Reload Communications' chief strategy officer Jack Yu reminded the panel that traditional companies aren’t necessarily at a disadvantage when it comes to data usage.  “Before big data we still worked with data,” he said. “We owned it analysed it and made connections” and this has continued. 

Be he noted just as agencies have had to enlist data scientists alongside creative thinkers, so do traditional companies need to apply creative art to the data to make it shine.

Ren reminded the delegates that a good litmus test is to see how often you’re able to put your data to work.  “The goal is to create a data hub and to create synergies between the hubs,” she said, so that core data is being used to inform other data sets. 

AI the storyteller

AI has been the red hot topic in marketing this year, and DMA was no exception, as panelists on an AI-focussed panel delved into real applications for AI in the industry.

At its most basic level, said Eric Weng, director of client solutions with Carat China, AI elevates the efficiency of the daily work of marketers in programmatic and media buying. For a travel booking platform such as Ctrip, chatbots and machine learning are thought to be able to improve user experience due to the huge volume of data these applications are able to handle.

This prompted Q. Chen, GM of advertising and sales in eastern China for MIUI at Xiaomi, to raise the point that improving user experience is at the heart of the role of marketers, and that should be their motivation when using AI. “IoT and AI speakers used in the households are the most common examples on AI applications that consumers find useful and enjoyable,” he said.

In terms of data, Kevin Guo, general manager of strategic cooperation and marketing innovation with Ctrip, said bigger is always better. “Look at what Tencent and Alibaba have achieved. We can always expect better results when a huge volume of data is used at a large scale. China has a huge potential for big-data applications due to our big population.”

Jeff Ji, partner and SVP at @comm, who moderated the session, said that further applications of AI can be expected in content marketing.

Guo discussed the poem generator (小诗机), a campaign that Ctrip launched in spring this year, where AI was used to create poems to accompany photos uploaded by users on the platform. “Travel was a huge inspiration for the great Chinese poets from the Tang and Sung dynasties," Guo said. "But in modern times, it is very rare to find a poet among us. But AI can now write poems easily following the metre system and data gleaned from our users’ reviews and location details. It became very popular because users thought it was fun and exciting to receive a poem in less than three seconds. That was an example of how AI can be used in storytelling and to create shareable content.”