“We’re Always Racing Our Clients” — but Ad Agencies No Longer Know How to Stay Half a Lap Ahead | What Changed the Advertising Industry ②

Why have ad agencies, whose whole business is making “creative,” entered an age of “product R&D”? What are the “data” and “technology” everyone keeps talking about, and can the things agencies are building keep them the authority on what consumers want?

The client (Party A) writes the brief, the agency (Party B) makes the creative and buys the media, and together they reach as many target consumers as effectively as they can. That is how advertising has always worked, and it is now under enormous strain.

Never before have people in this business been so confused, and so unsure of themselves, about consumer insight — the foundation and the core of the whole industry.

Mobile life changed everything. Consumers now hop between screens and spend their time and attention however they like, and the marketer’s article of faith — put the money where the consumer is — has lost its bull’s-eye.

That the television era is over is close to industry consensus. But in the dizzying tangle of digital marketing, no one can yet say what the most effective marketing actually is.

The search began at least three years ago. Judging by what is happening now, and by our interviews, the anxiety has not gone away. Every role along the chain, however, has changed fundamentally.

Brands that once handed all their marketing content to agencies are now building in-house teams and making it themselves. Agencies may be living through their most R&D-obsessed period ever: data labs, incubators, data-integration platforms. Media companies that used to sell nothing but time and space have been pushed into making creative, because advertisers want ads that fit the outlet’s own tone and style. And Facebook, the standard-bearer of social media, suddenly finds itself squaring off against Google…

That is why we are running this series. In four articles we will look at how each role in the advertising chain is changing, and how each is rewriting the rules. This is the second.

When Jin Jing joined Dentsu Aegis Network as its chief information officer in 2014, the global advertising group had no clear digital target. Now it plans to make its business “100% digital” by 2020. Today the figure is above 50%.

Before Dentsu Aegis, Jin Jing spent 18 years at technology companies, including nine years in Microsoft’s technology division. Walk through any agency today and you will find more and more people like her, hired out of information technology and handed serious responsibility. Titles like “chief digital officer” and “chief information officer” have become common in the past few years.

The industry has moved into digital marketing as a body. The research firm eMarketer says digital ad spending in the United States will pass television for the first time this year, and that by 2020 the gap will be much wider: US$113 billion for digital against US$77.9 billion for television. China’s digital ad market is growing at more than 20% a year.

None of this is surprising. Young people moved from the television to the computer, and now they live inside apps. The phone is gripped in the hand all day, entrances to information are everywhere, and willingly or not, everyone has gone digital.

So “digitalization,” “big data,” “programmatic buying,” “DSP,” “DMP” have become the words the industry argues about. A forum GroupM, part of WPP, held this September was named, bluntly, “Data Gathers Wisdom.” The Dentsu Aegis leadership summit was blunter still: “Using Data, Insight and Technology to Talk to Chinese Consumers.”

The truth is that the louder an agency’s enthusiasm for these subjects, the deeper its confusion about them.

“You have to see the change before your client does. Once the client sees it, they leave us. We’re always racing our clients, and we always have to be half a lap ahead,” Wu Jun, managing director of media buying at GroupM China, told QDaily by phone. The problem is the one we described in the first article of this series: on the question of how to stay ahead of advertisers, agencies are more lost than they have ever been. In a media environment this fragmented, they have lost their grip on both media resources and consumers.

“The hardest challenge facing agencies that provide market strategy is that they keep drifting further from the centre of the business,” said Nitin Nishandar, regional director for brand and communications in Asia-Pacific at Kantar TNS. “Agencies are still one step short. They can advise the advertiser, but they can no longer react to consumer insight and feed it back quickly enough.”

Agencies were not slow to notice. They have always been alert to change, and they are looking for answers. Publicis Groupe announced an incubation programme late last year to invest in digital startups. Dentsu Aegis recently set up a “data lab” to connect the data held by its various brands. GroupM launched three digital marketing products in one go at a forum a few weeks ago.

All of them are chasing the same question: how to use technology to find new ways of reading consumers, and how to use precise data to spread media money so that advertising works harder.

Notice what has happened. Agencies that used to produce only “creative” are now creating “products.” Agencies never owned a product before. The product belonged to the client; the resources belonged to the media.

Put another way, agencies care about “R&D” more than at any point in their history, and that forces them to deal constantly with technology companies and to learn the rules that technology and data run on.

So what is this “data” they all talk about?

“We asked ourselves what data actually is,” Wei Wei, managing director of mobile marketing at GroupM, said at the digital marketing forum in early September. The conclusion: data is the information you turn into consumer insight, and its purpose is to understand consumers better. Agencies did not grasp that at the start.

The “data–information–insight” method is not new to them. What changed is the definition of data and the range of information it can yield — a range that lies outside anything in an agency’s experience. That is the main reason they are both infatuated with data and lost in it.

Take Getui, a push-notification company that sends the weather forecasts and news alerts to your phone. By sitting inside the software development kits that app makers install, it reaches 600 million to 700 million active mobile devices a month. Those devices tell it which apps people have, and when and where they open them. By tracking that behaviour Getui can build hundreds, even thousands, of consumer models, and can pick out a group with a particular trait — pregnant women, for example — without ever contacting a single one of them.

“Phone use peaks at 10:10 p.m. and bottoms out between three and four in the morning… We found that almost every pregnant woman is asleep before 11, with the phone switched off or offline. So when a girl who uses her phone normally suddenly starts going to sleep before 11 for a stretch, we can guess she’s pregnant,” said Fang Yi, Getui’s chief executive. By watching which devices connect to the public WiFi at maternity hospitals, the company can count the days a phone stays on that network — that is, the length of the hospital stay — and from it work out whether the birth was vaginal or by caesarean, and whether the woman is early or late in her pregnancy.

That is how Getui became a new player in advertising.

The richest holdings of information about Chinese users online belong, as you would guess, to Baidu, Alibaba and Tencent. Inside Alimama, Alibaba’s platform for pooling brand marketing resources, every consumer is given one ID and only one. That ID gathers their behaviour across Taobao, Tmall, Alipay, Youku Tudou and other services, and hangs a long list of labels on it.

When Alimama can put 4,000 to 5,000 labels behind a single ID and sell that as consumer insight, you can imagine the shock to the advertising business. Until then, consumer data meant age, sex, occupation — demographics, in which every individual was a blur. The 76 options Facebook offers under “gender” are, in their way, the same sharpening of the picture, done to serve advertisers who want to know who they are talking to.

How has the new definition of data changed the way agencies work?

Wu Jun still remembers her first job, in 1996, buying media at J. Walter Thompson. Her day’s work was drawing up television schedules for clients.

“Back then in Shanghai there was Oriental TV, Shanghai Television, Channel 8 and Channel 14, and that was it. You made the schedule and you were done. The media environment was simple and the clients wanted simple things.”

Digital media is far more varied, and that slow way of buying stopped working long ago.

Today, through what is called programmatic buying, agencies let algorithms decide which platform an ad goes to, what price is fair and which time slots carry the best inventory. Nothing like the television era, when every new client meant hiring another batch of people to negotiate deals, draw up schedules and divide the budget.

In the United States, programmatic accounted for 67% of the digital ad market in 2016. In China it has been the most talked-about way of buying media for two or three years, and its share of digital advertising is expected to pass 50% this year.

For agencies, programmatic cuts labour costs and raises efficiency. It also means handing part of their craft and their media-buying experience to newcomers, because agencies do not own the technology and must rent it from third-party ad-tech firms. More hands are reaching into the same pie.

Technology’s blunter effect on agencies is on how decisions get made.

“The days of marketers deciding things in a meeting room are over,” said Winder Chen, chief executive of the ad-tech company Kuaizi Tech, at a marketing leadership summit held in early September by Carat, part of Dentsu Aegis. “The campaign has only just started and so has the creative. Right up to the moment the campaign goes live we have no idea which kind of thing consumers will like. The job now is to build a frame for the creative cheaply, and probably several versions of it.”

You may not know this, but when you and a friend open the same app at the same moment in the same place, your splash screens can differ. The advertiser and the agency build a library of components — words, pictures — and when you open the app the library is plugged into a far more complicated ad-serving system, which reads a phone user’s past habits and preferences, pulls different components and assembles a different splash screen for you. That is what Winder Chen means by “several versions,” what the industry calls dynamic creative or programmatic creative, a branch of programmatic buying.

But Huang Xiaonan, founder and chief executive of iPinYou, one of the companies that sells programmatic buying, says that in China not many advertisers can do this yet. “They may not have that many creative components, so how creative they can be is very limited,” she told QDaily by phone. iPinYou was the first company in China to offer programmatic buying; Huang Xiaonan previously worked at Procter & Gamble, the world’s largest advertiser.

Her point is that Chinese digital marketing has reached a state where the technology is ready and the creative is not, and a gap has opened between media and content. In the television era there was always too much creative chasing too little media: the creative went looking for the technology and the platform. Now the technology and the platform go looking for the content.

The root of it is that data has produced a new way of working, and it does not fit the sequence agencies are good at — client writes the brief, agency does the consumer insight, agency makes the creative, agency places the ads. That straight line has been broken, and the new rules do not put the traditional agency at the centre.

What is all this agency “R&D” meant to fix?

Agencies have never cared this much about developing products.

At the industry forum in early September, GroupM, WPP’s media arm, released three products and services at once: a consumer-insight tool called Hummingbird, a big-data platform from WPP worldwide called Compass, and a “premium media platform” laying the groundwork for programmatic buying.

Wu Jun says the premium media platform exists because advertisers now want brand building and sales to come as one. She describes it as a set of drawers, with live data deciding which media platform a piece of ad content belongs on.

“In the years when traditional media led, all we talked about was big exposure. You got brand awareness, but how much of it connected to sales was outside what a traditional media agency did,” Wu Jun said. Programmatic buying, she believes, can close that gap — though it also means leaning on outsiders: “Programmatic buying has to pull a lot of technology companies and data companies into the chain.”

Which raises the agencies’ own problem: how do they get enough control over technology and data to hold on to their standing in the chain? This is not what they are good at.

One answer is to hire from the technology companies — Baidu, Alibaba, Tencent, Microsoft, IBM — and you can already see more and more of those people inside agencies. GroupM has created a new post, head of data for media buying, and filled it with someone from HP.

After joining Dentsu Aegis a year and a half ago, Jin Jing took on the job of setting up the data lab. Its goal is to join up the data held across the group’s brands, analyse it, turn it into consumer insight, and give advertisers a sharper picture of who they are buying against.

The lab is not a separate department, Jin Jing says. Its staff come from the group’s various brands, because the more kinds of consumer data you bring in, the fuller the insight that comes out. Working data-first has changed how a large advertising group runs internally: the agencies inside it used to serve their own clients and sometimes competed with each other, and now everything is shared on one platform.

Wei Wei, who runs Hummingbird at GroupM, may describe the mood best. Building these products, she says, is “breaking your own legs and training your inner strength.” Jin Jing at Dentsu Aegis sees the data lab differently — as “a step out” toward a “revolutionary” change.

Everything agencies are doing now comes down to getting more data and analysing it, and they would like the technology companies to open up more of theirs. But there is a second question worth asking: beyond the data, how do consumers actually react to the advertising? Or, as Winder Chen of Kuaizi Tech puts it, “When I change the content, how sensitive are consumers to me?”

Their ambivalence shows. James Harris, global chief digital officer at Carat, thinks “sometimes we use too much technology, and all the technology brings is noise.”

“If a new touchpoint does nothing real for a category, we shouldn’t let it bewitch us,” Nitin Nishandar of Kantar TNS told QDaily.

They are not sure this is how it ought to be. You can read it as agencies sobering up and moving with the current, or as change they have no choice but to reach for in their confusion.

What everyone agrees on is that they need to connect to premium media. But who decides what counts as premium? That is the subject of the next article.

Cover image: Visual China Group