Assembly, activation, and the many flavors of Customer Data Platforms
In this CMSWire story aiming to better spell out the differences among CDP players, Dom Nicastro speaks with mParticle's CMO, David Spitz, to better understand the difference between application- and foundational-oriented CDPs.
Two things are certain about the customer data platform (CDP) industry: There is hype and there is confusion. The CDP Institute came out with a RealCDP program to try to bring some clarity to the market. According to multiple sources, the market has seen a decisive split in what CDPs actually do. The basic promise? Bring together all your relevant data. But are CDPs just bringing together data, or fostering activation and orchestration as well?
CDP Slices: Assembly and Activation
If you ask David Raab, founder of the CDP Institute and creator of RealCDP, the CDP market is divisive in terms of what the vendors offer. “I’m seeing a very distinct split in the CDP market between companies that focus on data assembly and those that focus on activation,” said Raab.
He cited Segment’s position as a “customer data infrastructure platform.” The assembly vendors, Raab added, build a lot of connectors and data transformation capabilities. The activation vendors? They focus on analytics, personalization, orchestration and even message delivery.
Are Some CDPs Marketing Suites?
Some CDP vendors include capabilities like analytics and personalization, does that make them similar to marketing clouds? Yes, according to Raab, who added, “This increasingly makes them look like marketing suites, although in theory they are more open to external inputs and targets than a traditional integrated suite.”
Raab pointed out that this is a natural division since data management and marketing applications are quite distinct capabilities. “Many companies will want best-of-breed for both, which may mean purchasing two different products,” Raab said. “We see that more and more — products like Segment, mParticle and Tealium feeding into products like BlueConic or Evergage.”
Application vs. Foundational CDPs
David Spitz, CMO of CDP provider mParticle, explained in a recent interview with CMSWire that he also sees the CDP market split into two axes: application-oriented CDPs and foundational-oriented data infrastructure CDPs.
The application CDPs usually come from a CRM or ad-tech heritage, Spitz said. “They're not just focused on collecting and normalizing and syndicating the data, they also have some tools that they want you to use, and there might be a different suite in their portfolio,” Spitz said.
To Raab’s point, Spitz called these CDP players “essentially mini marketing clouds.” They’ll compete on some level with Salesforce or Adobe, two of the major marketing cloud players in the marketing technology (martech) space. Foundational CDPs, meanwhile, are tightly-aligned with IT and provide a foundational data layer, Spitz added.
More Than a Split Down the Middle
Some see the division between CDP offerings in multiple categories. Hilary Edelstein, senior analyst at Winterberry Group, said she’s seen the set of vendors positioning themselves as CDPs largely emerge from four legacy categories:
- Data-as-a-Service providers that live in the data layer.
- Identity resolution and tag management providers.
- Legacy analytics and intelligence providers.
- Engagement applications and solutions.
“Each provider has a unique set of capabilities, strengths and weaknesses based on their original expertise,” Edelstein said.
CDPs Must Go Beyond Data Management Basics
According to Edelstein, true CDPs must be able to properly execute on four core capabilities:
- Ingestion and integration of customer data from multiple sources.
- Customer profile management, including identity resolution to create universal and persistent customer profiles of visitors across different states.
- “Real-time” customer segmentation.
- Exposure of customer data to other systems via “out-of-the-box” connectors.
“All of these reside in the data layer and are crucial for proper CDP functionality,” Edelstein said. “We feel only 15 to 20 of the CDPs in market currently fit this definition.”
Raab's RealCDP program measures CDP systems against five capabilities:
- Ingest data from any source.
- Capture full detail of ingested data.
- Store ingested data indefinitely (subject to privacy constraints).
- Create unified profiles of identified individuals.
- Share data with any system that needs it.
"Taken together, these capabilities describe a system that meets the true promise of the CDP," according to material from RealCDP. "Systems that lack one or more capability will have shortfalls that limit what users can do, often preventing them from supporting important CDP applications."
CDP Vendors With Analytics/Engagement Roots
Analytics and marketing activation/orchestration are often provided by CDPs but are non-core capabilities, Edelstein said. “Vendors who specialize in and excel in these areas typically originated from the analytics/engagement legacy categories and developed CDP capabilities over time to promote and advance the adoption of those original functions,” she said. Edelstein cited an example where an analytics provider may have implemented data integration and identifiers to aid its core function: analytical modeling. However, these providers are not truly CDPs, but rather application-centric platforms with “CDP-like” functionality.
Buyers looking to purchase a CDP should take the underlying expertise of platforms into account so they can see where that vendor’s strengths may lie, “depending on the use cases the buyer wants to achieve via the platform, it may also make sense to layer different CDPs together,” she said.
CDP Is a Function By Itself
A CDP is ultimately a function that can be purchased by itself (in a data assembly CDP) or as one with multiple features (in a data activation CDP), according to Raab. “The reason a data activation vendor would include CDP capabilities is that some companies need both sets of features and would rather buy them from one vendor,” Raab said. “Those vendors could still sell to clients who only want one of the two sets of capabilities, so long as the pricing remains competitive,” he said. “As more companies put a good database in place, it will become less valuable to offer that capability as an add-on, which will reinforce the split between the two sets of vendors and likely lead the activation vendors to focus on the marketing application features and less on the CDP features.”
Raab still sees CDPs in a “transitional period” at the moment because so few companies do have a good database. “So vendors,” he said, “will be most successful if they offer both.”
Latest from mParticle
Between two firms: Featuring Hull and mParticle, moderated by the CDP Institute
If you've ever wondered about CDPs and the role they play in both B2B and B2C industries, here is your chance to hear it straight from the source! Listen to Michael Katz (CEO and Co-founder of mParticle), Romain Dardour (CEO and Co-founder of Hull), and David Raab (Founder of the CDP Institute) discuss what CDPs are and how B2B and B2C CDPs differ.
How a CDP helps you secure your customer data
The trust between a customer and brand is the foundation of a strong customer relationship. Part of maintaining that trust is sound customer data management and security. Learn how a Customer Data Platform helps you secure your customer data pipeline so that you can build trust throughout the customer journey.
iOS 14, third-party cookies, and the future of data-driven marketing
This year, at its annual developer conference, a significant announcement that will have a wide-ranging impact on the future of mobile customer experience was sandwiched between the usual product announcements. The IDFA (the ID present on all Apple mobile devices and used across the martech and adtech industries) would now require explicit user permission to use. This change dramatically limits the usability of the IDFA as a form of universal identity that brands and vendors can use to power marketing, measurement, and everything in between.
Testing in production successfully with mParticle and Split
Testing in production allows you to release features with more efficiency and greater confidence, but doing it successfully requires good testing control and data management processes. Learn more about how you can get started and begin optimizing your test cycles.