Would you like to maximize the value of your data ecosystem?

How should the starter (Unbundling CDP) modern data stack be implemented?

In our previous post, we described four stages of the Data Maturity Journey in companies. 

Companies that are just beginning their data journeys frequently utilize a straightforward architecture geared toward data collection and activation. On the opposite end of the spectrum, businesses are using machine learning to provide customers with real-time interactions and recommendations.

A key to extracting value from data is to identify which stage of data maturity your company is in, and then to activate the different components of a modern data stack based on this stage.

Using this starter (Unbundling CDP) modern data stack, we will describe the different steps involved.

In this stage, there are three main challenges:

1. Implement a Unified Data Layer to create Data Consistency.

Using multiple systems, technologies, to create and manage customer data caused inconsistency across tools in the stack.

Implementing this Unified Data Layer, we will be able to gain an understanding of our customers’ behavior that will include:

-Data about their behavior in the product (customer behavioral events).

-Customer profile data.

2. Create an infrastructure that eliminates the need to implement point-to-point integrations between different SaaS tools.

3. Send the customer data you track from your Website or App to applications within your stack. 

In marketing, you may need to send personalized emails to your customers based on behavioral events.

In order to improve a certain product feature, you may need a better understanding of how customers use the product.


A company should implement this starter (Unbundling CDP) stack if it exhibits the following symptoms:

 -You have different tools, and in each of those you have an isolated data sample of your customers’ journeys. You do not have a system that allows you to see your customers from all sides.

-You have implemented different tools that make the same data different depending on the tool you use to query it. Across your stack, there is no data consistency.

-It is impossible to develop more advanced use cases because the data is fragmented across different SaaS technologies (CRM, Email automation, Customer Success, Product Analytics, …).

-You have dedicated development resources for updating specific integrations between SaaS tools.


A 3-step process for implementing a starter (Unbundling CDP) stack :

When it comes to investing in a data stack, the first step is crucial. Data stack decisions you make now will have a significant impact on the future. Getting this first step right is crucial because it will make it easier to progress to later phases, like the Growth Stack and Machine Learning Stack, where you’ll build advanced analytics, enrich data, build predictive models and personalize customer experiences.

1. Set up an SDK that tracks how customers use your solution (Web and App) and collects data about those key events.

Prior to implementing the SDK, you will need:

A taxonomy that explains what types of events and actions need to be registered.

Taxonomy will include :


-Event properties.

-Account or Organization properties.

-User properties.


2. Implementing a Unified Data Layer.

You will have accurate visibility into who your customers are and how they behave with a unified data layer that sends consistent behavioral events and customer profile data throughout your entire stack.

3. Data is being sent to downstream applications in your ecosystem.

Unified data layers ensure that all downstream tools have the same copy of events and user records.

The Starter Stack simplified event and user profile updates, so the architecture is so clean: SDKs in your Website and Mobile App send events with standardized schemas through an integrations layer and update downstream tools:

-Web analytics.

-Product analytics.

-Email marketing and automation.


-Ad platforms.


Components required for implementing the Unbundling CDP Modern Data Stack :

One-way direction data flows.

1. Data source.

-SDK for tracking behavioral events  (APP and Website).

Technologies : Segment, Rudderstack, Amplitude, Freshpaint.


Technologies : Fivetran.

2. Integration Layer.

3. Storage Layer.

4. Transformation Layer.

5. Data Activation.

Use cases to personalize your service based on data :

-Product analytics.

-Email marketing and automation.


-Ad platforms.

-Customer Success.

-Account Health.




Eric Dodds’ idea post inspired this piece of content.

Posted by:Fran Castillo