Separating Myth from Reality: CDP Components and Composable CDPs

Separating Myth from Reality: CDP Components and Composable CDPs

Composable CDPs: Separating Myths from Reality

As Customer Data Platforms (CDPs) have proliferated—there are hundreds of vendors in the market—a new camp has emerged: the composable CDP camp.

What is a composable CDP? The concept is pretty simple: instead of buying an all-in-one CDP solution from a vendor like Segment, mParticle, or Treasure Data, just build a CDP by assembling various CDP components and integrating them with a data warehouse. Some people refer to this concept as an “unbundled” CDP.

Proponents of composable CDPs are becoming more and more vocal, even predicting the death of traditional CDPs. And while they make some nice points, they’ve also overstepped on a few key points about traditional (integrated) CDPs. Here are five claims that the composable CDP proponents make that are actually myths.

Myth #1: Traditional CDPs restrict access to your data

While this may be true of some CDPs, it’s not true of all CDPs. Some CDPs—Segment is a good example—store the data in their data warehouse and restrict access to it: only Segment employees have access to the data while it’s inside of Segment. 

But not all CDPs are built like Segment. Treasure Data stores the data it ingests in a data warehouse. But—unlike Segment—Treasure Data makes that data accessible to Treasure Data customers. In fact, Treasure Data offers multiple query engines for their customers to use when accessing their customer data. 

The best CDPs provide the same levels of data access as the best data warehouses.  

Get Treasure Data blogs, news, use cases, and platform capabilities.

Thank you for subscribing to our blog!

Myth #2: Traditional CDPs don’t have tooling for data teams

Again, the reason this claim is a myth is because it only applies to some—not all—CDPs.  

Indeed, CDPs primarily cater to marketers, emphasizing marketing capabilities like segmentation, journey orchestration, and predictive modeling over data capabilities like querying, transformations, and automation. But not all CDPs are created alike. Treasure Data, for example, was originally conceived as a big data platform, so its users enjoy many of the tools and capabilities that data teams require: SQL for querying, Python for automation, and a Command Line Interface for doing it all programmatically.  

CDPs built on solid data foundations will come with the tools and capabilities that data teams love.  

Myth #3: Traditional CDPs can only ingest some of your customer data

A fundamental capability of any CDP worthy of the name is the ability to extract, transform, and load (ETL) customer data. As such, most CDPs can ingest a wide range of data, including SaaS and POS data. 

With hundreds of CDP vendors in the market, there are bound to be some with weak data ingestion capabilities, but these are the exception and not the rule. The strongest CDPs in the market can undoubtedly ingest SaaS and POS data—in addition to many other types of customer data. For example, Treasure Data offers plug-ins, APIs, SDKs, workflows, custom scripts, and pre-built connectors that can ingest practically any kind of customer data.  

Myth #4: Traditional CDPs lack analytics & insights capabilities

Data warehouses have long been associated with insights & analytics. The whole idea behind a data warehouse was to centralize data in one place so that analytics teams could query it, discover insights, and then share those insights with decision makers.    

On the other hand, CDPs have long been associated with marketing use cases. The idea was to facilitate and distribute the customer data to power marketing campaigns. Analytics wasn’t a priority. Instead, many CDPs focused on integrating third-party analytics tools rather than including analytics in the platform. 

But as the CDP market has matured, the leading CDPs have developed robust analytics capabilities—capabilities that rival the analytics tools that run on top of data warehouses. For example, Treasure Data users can use SQL to run sophisticated queries, Treasure Insights to create beautiful dashboards, and various self-service machine learning capabilities to build predictive models. The analytics capabilities of the best CDPs rival those of the best data warehouses.   

Myth #5: Traditional CDPs cost more than composable CDPs

The logic behind this fallacy goes something like this: the components that make up a composable CDP are either cheap or free to use. Traditional CDPs can be expensive. So, you can save money by building your CDP on top of your existing data warehouse using those inexpensive CDP components.  

It’s true—CDPs are potent tools that require a sizable investment to implement and run. It’s also true that some vendors offering CDP components offer their products for little or nothing. But just because the individual sticker prices of CDP components are small doesn’t mean that the total cost of owning a composable CDP is.  

For one, this line of thought ignores the costs of integrating all those independent CDP components. With a traditional CDP, each component is designed & built to work with the others; the cost of integrating CDP components is borne by the CDP vendor. With a composable CDP, the user must pick up those integration costs.  

Secondly, those free or low-cost CDP components might work for organizations with small aspirations for their customer data, but they won’t work for those with a big vision for their customer data. If you want to do big things with your customer data, you will have to pay for big capabilities. 

Finally, if you build a composable CDP yourself, you’ll also have to maintain it yourself. Maintenance will likely require a team of architects, analysts, and engineers. Their salaries might not show up on the sticker price of a composable CDP, but they’ll certainly show up on the bottom line of whoever’s paying for the CDP. With a traditional CDP, the vendor absorbs those costs.  There’s no free lunch—you get what you pay for when it comes to building your own composable CDP. 

How Treasure Data Can Help

With over 150 vendors calling themselves CDPs, it’s easy to find some with weak offerings. Against these weaker CDPs, a composable CDP might be a better option. But the strongest CDPs in the market can more than match composable CDPs: they are built on solid data foundations, providing all the benefits of a data warehouse in addition to the capabilities of a CDP and the value of a fully-integrated solution. 

Treasure Data is one of the most robust CDPs available in the market, which explains why it was named a Leader in the In IDC MarketScape for Data and Marketing Operations Users. You can learn more here. In other words, traditional CDPs are far from dead.  

To learn some guidelines around deciding whether an integrated or composable CDP makes the most sense for your business, download this white paper by the CDPI.

Composable or Integrated CDP? Making the Right Choice for Your Business. Download the white paper.

Jim Skeffington
Jim Skeffington
Jim Skeffington is a Technical Product Marketing Manager at Treasure Data. He has years of experience working with data, including as a financial analyst, data architect, and statistician. Recently, he was recognized by the Royal Statistical Society for his thought leadership in the fields of statistics, data science, and data research. He is also proud to serve as a Captain in the United States Marine Corps.
Related Posts