CDPs: Why now?
Our new MarTech Intelligence report on choosing a CDP vendor elicits an inquiry into how this category emerged only recently, after years of hype about “a single view of the customer.”
For much of this century, marketing software vendors have promised “a single view of the customer” or “a single source of truth.”
So, why has the category of Customer Data Platforms (CDPs) just recently emerged?
In other words, there clearly has been an articulated desire to house all customer-related data in one place, and there certainly was technology to create a single database.
What’s different now?
First of all, Preethy Vaidyanathan told me, major new business drivers have emerged relatively recently. She’s Chief Product Officer of CDP Tapad.
Data aggregation projects have, of course, existed in previous years, she said, but they have largely been separated into silos. Marketing coalesced all of its customer info, sales did theirs, and the call center theirs — each separately.
But now, she said, “brand priorities have changed” because of changes in some of the key ways that brands boost their revenue, and these changes are driving the need for a single repository across the brand.
The current state of CDPs, and the factors involved in choosing a vendor, are covered in our new MarTech Intelligence Report, “Enterprise Customer Data Platforms: A Marketer’s Guide,” out this week (registration required).
For example, she pointed out, there has been a boom of direct-to-customer (DTC) efforts “in every vertical,” like glasses retailer Warby Parker or meal provider Blue Apron, as well as DTC efforts by longstanding retailers. To market, sell, ship and service directly to the end buyer, you must get all of the sources of customer info together, talking to one other.
Josh Neckes, co-founder of CDP Simon Data, noted that Amazon, Google and Netflix are known for their brand-wide direct interaction with their customers — for which they’ve built their own internal CDP equivalents.
Even brands that still work through retailers understand that the key diffentiator these days, when a customer can reach a competitor in a click, is often a great customer experience that is consistent through every touch point when there is interaction with the brand. Again, that means unified data available immediately throughout the brand, especially the identity resolution that many CDPs offer.
And there is an increasing need to completely and quickly manage the privacy of all customer data, Vaidyanathan noted, which can’t readily be done if it’s spread across departments.
“[Privacy] is no longer just a nice-to-have,” she said.
These days, analyst David Raab told me, “it’s more painful not to have a single source of truth, because there are so many more channels and options.” Raab founded the Customer Data Platform Institute, which has helped to define and promote the category.
Along with the new business drivers, CDPs have also been propelled by technological advances — in particular, new tools for handling unstructured data, such as the huge tsunami of social posts.
‘A new kind of data platform’
Neckes noted that Hadoop, a big data solution that’s been around for years, has developed a mature ecosystem only in recent years. It and other recent technologies, like big data analytics engine Apache Spark and real-time data pipeline Kafka, Neckes said, “provide an opportunity for a new kind of [data] platform.”
In addition to readily storing and accessing unstructured data, this new platform can hold lots more immediately accessible data than ever before, allowing thousands of data points for each customer.
Data architectures involving Hadoop and related tech, Raab said, “can store data you weren’t planning on having.”
In a traditional data warehouse, he said, “the IT department had to map the data into data tables for a relational database.” By contrast, he said, Hadoop is a file system built around such simplified data structures as identifier (such as eye color) and a value (blue).
This means that when a new communications feature is added to a popular social app, for instance, the data repository doesn’t need to plan and map it before acceptance.
“It just shows up in the data,” Raab said. This is essential to the concept of a CDP, which is intended to house and make available every kind of data relating to customers and their interactions with the brand.
Evolve into mammals?
Marketing automation platforms, data management platforms and customer relationship management (CRM) systems have been primarily focused on structured data, he said. CRMs, for example, were initially designed for sales agents and call centers, not as multipurpose centers for all customer data and identity.
Raab noted that when an existing marketing platform like Salesforce bought a social media firm like Radian6, the resulting influx of unstructured data largely stayed in silos or was imported when needed.
Salesforce doesn’t offer a completely unified, immediately accessible view of all structured and unstructured customer data, he said. But Raab thinks Salesforce’s recent acquisition of marketing intelligence and analytics platform Datorama could evolve into that marketing platform’s operational CDP.
Legacy customer systems are also more limited in their data flows than the more modern CDPs, Neckes said, since they were primarily designed to acquire data, rather than both acquire and distribute. And CDPs are “orders of magnitude” better at using their data in real time for orchestration, he said, such as setting up and running campaigns across every channel or modifying actions in real time as needed.
All of these factors — including huge amounts of data, unstructured data, two-way data flows and real-time orchestration — mean that CDPs are built specifically to address more data, and more kinds of data, in a real-time environment. These attributes are exactly what modern marketing applications — like identity matching, predictive marketing, analytics, real-time multichannel campaigns and operational uses of AI — need to operate at their fullest.
Given the drivers and tools for customer data management in this current evolutionary epoch, can the legacy dinosaurs ever become nimble, modern mammals? I asked Neckes.
Yes, he said, by buying CDPs and rebuilding around them.