Building Coca-Cola’s Customer Data Platform

The beverages giant brings marketing technology in-house by focusing on first party data and a hands-on, real-time approach to marketing campaigns

American businessman John Wanamaker, whose eponymous department store helped shaped the country’s retail industry in the late 19th and early 20th century, is often best known for his quote: “Half the money I spend on advertising is wasted; the trouble is I don't know which half”.

When Harish Kundargi joined The Coca-Cola Company as its ASEAN Marketing and Digital Technology Head at the end of 2018, the company was wrestling with similar issues that vexed Wanamaker. Decision-makers were considering buying chunks of information from Digital Marketing Platforms (DMPs) to better target its ads; Kundargi disagreed.

“What is essential for all brands, and it certainly was for us, was first party data,” he explains, referring to personally identifiable information (PII) such as names, phone numbers and email addresses that serve as the fuel behind Customer Data Platforms (CDPs). DMPs, the biggest of which include names such as Lotame and Krux (bought over by Salesforce in 2016), mostly work with third party data that are anonymous.

In other words, CDPs can identify on whom the marketing money had been spent and whether it has made a difference. DMPs, on the other hand, do not address the Wanamaker issue of identifying wasted ad spend.

CDPS AND THE POWER OF PII

Because legal restrictions stop DMPs from de-anonymising its data – often bought from telcos – “DMPs will only know my identity as ID12345…but my identity in a CDP includes my name and email etc.,” Kundargi notes. He believed that Coca-Cola should build its own CDP to maximise its marketing efforts.

“What we are building is essentially a Customer Data Platform for us to host all the first party data that’s coming through on our websites, our applications, our tactical promotions that we do, etc.,” Kundargi explains. Because Coca-Cola had outsourced its marketing operations to external agencies, all that the data sat externally. Even when it got some of that data in weekly reports, it was not real-time and lacked the required granularity.

“We built a platform around this internally to say, ‘We need to know at all points in time if a campaign is performing well.’ We have access to all the data now [following the completion of the CDP]. If a campaign is not working in week one, I will take money out of that campaign and put it into something else. We had never been able to do that because it was never in our control.”

Perhaps more importantly, having all that data allows for better identity resolution.

“One of the things you want to do is stitching profiles,” says Kundargi. “If people are coming to your platforms from promotions, from Facebook, or a mobile app, or programmatic channel, you want to be able to stitch them together to make sure that it’s the same user who’s interacting with the promotions.

“Is this the same user who clicked on the promotions 100 times? Or is it 100 different users at different times? That’s the differentiation you need to make. You can only do that when you stitch the profiles together.”

Additionally, fleshed out profiles can also be used to derive correlation, such as family relationships. Kundargi explains:

“For example, if I see the same IP address for five different consumers, I can probably say with 50 percent accuracy that they probably belong to the same household or family. I can’t say with 100 percent certainty because IP addresses are not a reliable metric.

“If a company realises you’re all from the same household, and the decision-maker in the household is X, and you have device IDs being transmitted as well, you can look at who has the latest iPhone and who has a Samsung Galaxy 4, for example. Based on that you could guess fairly accurately who is the decision-maker. And then, you can target your messaging to that specific device or profile.”

MARTECH

All that can be classified as features of MarTech, otherwise known as Marketing Technology. When asked how companies can embrace digital marketing, Kundargi points out that there are lots of vendors in each of the many spheres within MarTech, such as social listening and content marketing. Instead of hiring multiple vendors using incompatible platforms and technology, Kundargi says it might be better to bring operations in-house.

“When it comes to MarTech, you need a person who’s able to draw out the entire architecture,” he muses. “If you have a social listening platform, and you have a consumer data platform, and you connect to ad platforms such Facebook and Google etc. – how do all these connect and come together? When the data flows between these systems, is it connecting through a central pipeline where you have control of the data?

“It’s crucial to the delivery of the best consumer experience. If the systems don’t talk to each other, you won’t be able to personalise content, you won’t be able to target consumers the right way.”

He concludes: “The basics though, remain the same: Acquiring and retaining customers, delivering the right experience, at the right time, through the right channels. In the digital era, it all boils down to applying consumer insights with the right marketing mix.”

 

Harish Kundargi was the speaker at the SMU Centre for Marketing Excellence (CME) seminar, “Building the right MarTech levers to drive meaningful Consumer Experiences and Insights” that was held at SMU on 17 January 2020.

Follow us on Twitter (@sgsmuperspectiv) or like us on Facebook (https://www.facebook.com/PerspectivesAtSMU)


This article  is a repost from Perspectives@SMU
https://cmp.smu.edu.sg/article/building-coca-colas-customer-data-platform

 

 

 


PROUDLY BROUGHT TO YOU BY