A Case for a bespoke Composable CDP over one off-the-shelf

In my previous articles, I’ve expressed my preference for off-the-shelf solutions for Analytics, and Data Science needs due to their reduced implementation time and effort. However, I believe a bespoke solution for a Customer Data Platform (CDP) is worth considering. While it requires a more significant initial investment, after exploring the market leader in CDPs and going through a month of demos and discussions, I’ve learned that it is worthwhile to build a CDP in-house if you have a modern data stack.

Customer Data Platform

Martech defines a Customer Data Platform (CDP) as a system designed to collect customer data normalized to build a unique profile for each individual. Primarily being a marketing tool, CDPs are a darling of marketing leadership, and I can see why. Marketing teams use CDPs to build out event-based communication strategies, which is vastly better than the alternative, blanket time-based communication.

The main objectives of a good CDP are the following —

• Collect data from customer’s every touch point/interaction
• Map out the customer journey
• Aggregate and slice the data as needed
• Synchronize and maintain data integrity

Good CDPs do these well and become a powerful weapon in a marketing team’s arsenal.

Exploring the Market Leader

Twilio’s Segment is a market leader in this space and is one of the top candidates for anyone needing a CDP in the United States. When our marketing team came to us with a proposal for a CDP, We strongly considered going with them as they had an integrated solution to weave customer journeys and build a 360-degree view of our customers.

We met with Segment’s sales and technology teams to understand what value they bring, what they do, how they do it, and the effort needed to implement their platform at our company. We quickly understood that Segment shines at -

• Connecting to various platforms and collecting customer events
• Transforming events and synchronizing data
• Building customer personas, traits, and audiences
• Data Quality management

Limitations of Integrated CDPs

Although we were eager to partner with our marketing team to bring Segment on board, within a few weeks of product demos and back and forth, it was clear that Segment and other integrated integrated CDPs have their limitations.

• Pricing — Integrated CDPs base their price on the number of visitors tracked per month (MTUs). Based on the traffic on your websites and your customer base, the cost can shoot up quickly to hundreds of thousands of dollars a year. For our member base, their estimate exceeded $200k a year.
• Implementation Challenges and Resource Requirements — Although Segment and its competitors provide connectors to a plethora of platforms, it requires you to embed their tags and javascript snippets into every web page and app screen. This takes hundreds of hours of engineering work and coordination. In addition to being resource intensive, you are tied down with Segment because of this unless you are willing to redo tag management for another vendor. Our front-end team had already implemented a template solution in GTM to track data customer data with Google Analytics 4, which helped us a lot.
• Control over Data — Segment stores data logs on their servers and charges extra (on top of the $200k) to push logs and events to our data warehouse. Bringing event data into your data warehouse is critical to analyze and identify patterns in your data.
• Limited Flexibility — Integrated CDPs

The Alternative

An alternative is to build your own Customer Data Platform using your data warehouse, ETL and reverse ETL infrastructure, and Data Science engine. Although this sounds like a lot of work, the cloud and serverless solutions are making it easier to build these without being too labor or cost intensive.

A composable Customer Data Platform (CDP) is a CDP that is designed to be modular and flexible, allowing organizations to build and customize their own CDP solution using a set of reusable building blocks.

In a composable CDP, the platform is broken down into smaller, more manageable components, such as data ingestion, identity resolution, segmentation, and analytics. These components can be mixed and matched, combined, and extended as needed, allowing organizations to create a CDP solution tailored to their specific needs.

A Warehouse-centric Customer Data Platform (Generic schema)

I will follow up with an article detailing our approach to the Composable CDP.

Advantages of a warehouse-centric Composable CDP

In addition to solving a lot of drawbacks of an integrated CDP, a Composable CDP has key advantages-

• Price — Even if this increased your warehouse cost a bit, with storage being so cheap nowadays, you would save hundreds of thousands of dollars!
• Data privacy and control — With privacy regulations like GDPR and CCPA becoming more stringent by the day and data breaches becoming more common, not storing your data in another 3rd party platform is a pretty good argument. Moreover, CDPs become the source of truth of your customer data, and you are tied to your vendor for years on end.
• Machine Learning and Data Science — Although Segment is excellent at building customer personas, you must use the data to build your communication strategies. As I mentioned, bringing customer event data into the warehouse and data science engines will allow you to build insights around your customer data. This can be leveraged in personalizing actionable intelligence.
• Break the silos — CDPs being marketers' tools, often siloed into marketing. However, building a CDP with the warehouse can be used by your product, customer care, and other teams to ensure you deliver a consistent, personalized message to the customer no matter how they interact with you.

Conclusion — build or buy?

Segment and other similar platforms are market leaders for a reason. They cut down on a lot of effort in weaving together a 360-degree view of your customers. These powerful tools help you move from a time-based marketing communication plan to an event-based one.

I’d suggest going with Segment or other competitors if

• You have lean data and engineering headcount.
• You are in the early stages of the data journey (if so, you may find this helpful).
• You are a marketing leader short on data science and engineering resources.
• You are in marketing and would rather control your strategies with a marketing tool vs. rely on the data teams.

I’d suggest building your own Composable CDP if

• You have a robust data architecture collecting data from multiple platforms and pushing data into business applications.
• You want control of your data and are ready to manage pipelines.
• Have a team of Data Scientists or Machine Learning Engineers to build out custom models to help marketers provide a tailored and intelligent customer experience.
• You have other business teams (customer care, product, etc.) that could use customer insights instead of data being siloed in marketing.