The deck said, in its first two lines, that our world had changed and our work had not. We told impactful stories. We knew earned content mattered to China's millennial mothers. And we had two problems that no amount of good content solved: our influence was limited to the reach of the publishers and influencers we worked with, and we had never been able to connect that influence to results in the real world.
That admission is the whole case. Content to Commerce was one of two products I co-created and helped take to market across the network, and I led Influence to Commerce and the social and digital practice for the region. This was its application to Pampers in China. The brand approved it and the test ran.
Why China's mothers were the right place to test it
The evidence for earned content was strong. Smartphone internet use took the largest share of a Chinese mother's media time, ahead of computer and television. Half of mothers researched diapers during pregnancy and another third after delivery. A majority always checked product reviews before buying. After professionals and friends, the most influential source on a baby-product purchase was a recommendation on a parenting forum, well ahead of recommendations on social media.
So the content that moved these buyers already existed, written by other mothers on forums and social platforms, and a brand could neither control it nor reach beyond its natural audience. The question was whether it could be found fast, amplified precisely, and measured honestly.
The mechanism
The approach had four parts. Identify earned content from a dashboard within fifteen minutes of its being posted, across articles, videos and blogs. Convert the strongest pieces into a paid earned-content unit that carries the original headline and the source's name, marked as promoted, with an image cleared by the brand. Target that unit programmatically on behavior, demographics, location and search, across desktop, mobile and social, so it reaches the right person rather than whoever happened to follow the author. Then drive the reader to the original publication, and re-message them with the brand's ad, its owned content or further earned content to shorten the path to purchase.
The measurement side is what made it a system rather than a media buy. For retail, sales lift was to be proven by a third-party panel matching exposure to purchase through store loyalty data, and brand lift by an independent research firm, so the number was not marked by the agency that ran the campaign. For e-commerce the conversion was measured directly.
The content rules were stricter than most brands expect. The promoted unit could carry compelling brand visuals and a compelling headline but no superiority claims. The article behind the click had to be branded, had to convey superiority, and could make no unsupportable claim. Every piece had to complement the language of the paid advertising rather than contradict it.
Three pieces of earned content, chosen to show the range
The deck worked three real examples. A celebrity mother, photographed shopping for baby products in Japan, had led a group of mothers on a visit to a Pampers plant, and the article framed the brand as the first choice of Japanese mothers, pediatricians and overseas buyers. A parenting account with over a hundred thousand page views carried a mother's account of trying four competing brands before settling on Pampers for absorbency and her baby's sleep. And a video from a mother-focused account with more than half a million followers showed a Chinese actor and father pouring water onto four diapers side by side and comparing the result.
Each was a different kind of influence, from a celebrity's endorsement to a parent's comparison to a demonstration a viewer could repeat at home, and each could be converted, targeted and measured the same way.
What ran
The test was defined the way the deck said it would be: a focus product, a market, a budget, and the measurement standards, brand lift among them, agreed before the first unit was bought. A media plan followed, and the mechanism was put in market.
What I would do differently
I would have run the e-commerce model first and the retail model second. The retail measurement chain, from a loyalty card to an exposure match to a lift figure, is the more convincing proof, but it needs a third party, a panel and a quarter of patience. An e-commerce test on a single product would have produced a number in weeks. Sequencing it the other way meant the mechanism spent its first months being explained rather than being cited, and a mechanism is easiest to defend once it has produced a figure someone else measured.
The design is on this page because the design is mine to describe. What the test returned belongs to the brand. The method is describable. The specifics are not mine to publish.