If data is the new oil, Customer Data Platforms (CDPs) want to be your refineries. CDP vendors promise to bring together customer data and stitch it together into unified profiles so news marketers can segment users on the fly and dynamically personalise user experience. 

Yet at the same time, analysts of Forrester Research claim Customer Data Platforms over-promise and under-deliver.

So, how is a news media publisher to decide whether or not to move foward with a CDP? That was the subject of an INMA Readers First Webinar on Wednesday, with INMA Researcher-in-Residence Greg Piechota and representatives from several major media companies.

The Economist

First to present was Steve Lok, global head of MarTech for The Economist. When it comes to Customer Data Platforms (CDP), Lok said there has been a lot of recent activity, from new vendors to sideways lateral moves.

“There’s a really important question around whether or not we can actually get value out of the CDP platforms,” Lok said, adding that The Economist didn’t get there overnight.

“Whether you want to look at this from a temporal basis or a maturity basis, it’s not really the thing you can jump into tomorrow. If there isn’t a level of data literacy and ability to understand how you really are generating revenue out of the data that’s available to you — and then being able to separate exactly the kind of third-party data that you have in your environment — it can be very difficult to understand not only how to get value out of the CDP, but [understand] the differences between a CDP and DMP, for example.”

DMP operates on third-party data, with no direct customer information, while CDP is a first-party data platform. CDP requires the ability to master your own data profiles.

Data literacy has to be built up over time, Lok said.

“One can jump into it more easily these days, I think because of the fact that CDPs have been moving so quickly,” he said. “It’s very important when trying to make CDPs work in your scenario ... to look at the origin of the company and strategically where they have come from and where they’re going.”

If a company is highly advertising focused, for example, Lok said it would be important to look at the CDPs that have come from the DMP world. “It’s very important to define exactly what your needs are first, and what your company is like, and then the kinds of interactions and transactions that you’re trying to drive people to.”

CDP: Intelligent profiles with data science-based insights

With The Economist being a subscription-driven business, the company needed to find a CDP that had a content engine that was part of the platform itself.

“Being a content-driven business, it was really important for us to have marketing profiles and behavioural profiles that were built, that had a constant connection with the relevance of the content that we had on our site,” Lok explained.

One of the most important aspects of establishing the value of a CDP for The Economist was that, in looking at an individual profile or a set of profiles, the segments could be defined not only by where they were coming from or time spent but also in relation to the content — the user’s interests and intent.

“If we’re targeting people to try and get them to convert and to push forward a point-of-value exchange — to get them to buy something — it was extremely important that we understood exactly what it was that triggered that.” This directly informed the type of marketing that The Economist would direct at that user.

Beginning in 2015, Lok’s team started leveraging DMPs to get their targeting better. “We had linked the segmentation that we were leveraging from particularly social search and display, and testing out different targeting groups.”

Adding technologies over time is what fueled the strategic transformation. “The business finally understood the power of data and what you could do in terms of understanding exactly who your readers were and moving the entirety of the business into a position where we could actually respond very quickly to signals in the market,” Lok said. With this history, at the point where the CDP came into the equation in 2016, the team was able to actually utilise it successfully.

However, he was quick to point out the journey was not easy. “Many companies fail with CDPs because they bring them in and they expect the CDP to tell them what to do,” Lok said. A company needs strong use cases that they can leverage the CDP against.

Dow Jones

Next up, Chief Analytics Officer Guthrie Collin of Dow Jones in the United States shared his team’s experience.

Any company should really focus on why it should invest in a Customer Data Platform, Collin said: “At the end of the day, you can use data science only if you have the data available.”

When you create data science applications using a CDP, you can:

• Grow subscription revenue with personalised offers.
• Reduce cost per acquisition with channel optimisation.
• Create and nurture content habits and reduce churn through automated and personalised engagements.
• Grow ad revenue through first-party segments.
• Drive news coverage through audience analytics.

A company needs to first define what the core CDP data will be, and will not be, Collin said. At The Wall Street Journal, there was a very diverse set of data. So the company organised it into information that was about the product, information about the customers themselves, and lastly the crossover being information about how customers use those products. The team also used operations data about how the company works.

“When we came to the CDP we thought, we can’t do this all,” Collin said. “It would be nice to have everything linked together. But to realise those benefits we had to focus on just two: the customer data and then the product usage data — how they’re using our products.”

His first suggestion for other organisations looking to employ a CDP is to have a really tight grasp on exactly how they want to be using it and what data they want to use with it.

Second, know that the heaviest work you will do with the CDP has the least tangible impact to the business. “The data science applications you build on top of that data platform has the most impact and takes the least effort — but only if you’ve done all the grind of collecting the data, transforming and curating it in your CDP.”

When Dow Jones launched its CDP, it bundled it with a set of data science applications from basic dashboard analytics to new segments and recommendation algorithms. A tangible benefit of that in the subscription world is that an organisation will have much more robust behavioural segments that aren’t stale, pre-defined lists but, rather, always current.

Creating content habits

Collin shared some examples of how the CDP helped his organisation create content habits. The Wall Street Journal Welcome Series, for example, sends messages via email or social outreach to members. These messages are personalised through differentiating if the user has already downloaded the WSJ app or not. WSJ does the same thing with a welcome series messaging for its newsletters. 

“This is only possible with a CDP that lets us identify patterns in people’s behaviours and then create engagement messaging around those behaviours,” Collin said.

Another way the WSJ uses the CDP is to reduce churn — through the use of anniversary e-mails, for example. This increases subscriber engagement and loyalty by sending a one-year anniversary e-mail to congratulate members, reinforce membership benefits, and prompt engagement actions.

“That’s how we take the CDP and actually deliver benefits of business,” Collin concluded.

MittMedia

Michelle Ludovici, product manager for data at MittMedia, joined the Webinar next from Sweden.

MittMedia built its own Customer Data Platform. “We needed to see what content works for the users, what content is the most interesting, but also who are our users anyway and how are they using our product?” Ludovici explained.

The company developed its own platform for several reasons. First and foremost, to create exactly the platform it needed, but also to be fast and flexible in the development, which isn’t always possible with an outside vendor. Being able to experiment and iterate is another important consideration.

Ludovici shared a slide that showed all the data MittMedia collects on its users: “What is important to see from this slide is that we have an overview about the whole ecosystem,” she said. “Everything that concerns the user.”

Although MittMedia has only one data platform, called Soldr, all various units within the company — whether they are producing content or advertising — can choose their own tools to use. Even third parties outside MittMedia can connect, including robot content providers and external publicists.

“Everything is on Soldr, unified in the same way, and then you can choose if you want to get it out manually or automatically or personalised,” Ludovici said. “It doesn’t matter if it comes out to our sites or to external sites.”

Next, she showed a slide that illustrated all the possibilities in different layers that the Soldr CDP provides.

“The point of this is to show that we can do so many innovations,” Ludovici said. “We have personalisation on this, we have a content API so we are able to make services talk to our server, and then in turn talk to the end customer, which will come more and more we believe.”

A perfect use case, she added, is personalisation.

“Personalisation actually uses the whole ecosystem,” Ludovici said. “We make clusters for our users that puts them into segments.” Geolocation enables MittMedia to serve articles close to where the customer lives. “We can match the article data to the users interests and needs and to where the user lives.”

From there, the team evaluates the CDP and data science to determine where to improve.

For MittMedia, the key was to focus on what needs the company could solve directly with the data platform: “If you start very small, then the scope doesn’t get too big. Because if the scope is too big, then it costs a lot,” Ludovici ended. “We started with safe data we know that we can put to use, data that is critical for the business development. You need to own it, and start with that and then you build on top of it.”

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .