Customer Data Platform -CDP Architecture – Deep Dive

Okay!! So now you have a decent understanding of the CDPs. You understood the basic functionalities of CDP, its features, its advantages, and limitations. Now you are eager to get more information on CDP and understand its advanced capabilities; at this point you are looking to figure out ways to get that CDP implemented in your organization. But at this point you need a bit more clarity on how they work. If you don’t know that yet, here is the detailed article for you – Customer Data Platform – The detailed Guide. 

There are different types of CDP’s which are present in the market. It can become so overwhelming to select the right one for your business. That is why getting a little bit deep in the customer data platform architecture, can actually provide you some in-depth understanding of the technology and can help you choose the right one.

How data flows through Customer Data Platform(CDP)

During the complete data lifecycle, data goes through various processes inside the CDP. It’s ingested, prepared, stored, explored, processed, transformed, analyzed, visualized, and activated. Various CDP vendors can use different terminologies, but the purpose for each remains the same. At the macro level, below are the stages in data any lifecycle –

• Connect/Collect: Collecting data from disparate sources such as website logs, Event streams(clicks, scrolling, etc.), Customer Relationship Management Data, POS data, Ecommerce, Social data, or any other source systems, etc. It can be First-Party data, Second or Third-Party Data.  These data are called Data Sources in CDP, collected and stored in data warehouses/cloud data warehouse. This data can be structured or unstructured.
• Transform(Unify & Enrich): Once the data is collected on any data warehouse or any data management platforms. That data needs to be transformed into structured data; it needs to be unified from all data sources. Different variations of data is then merged into one single source. That is why it’s also referred to as a single source of truth. Data is then linked to a single identity of the customer(it can be an Email address or phone number). It’s combined with other second and third-party sources as well.
• Analyze: Data is analyzed to answer some predetermined business questions using predictive analytics, deep analytics, machine learning, and AI.
• Visualize: Once the data is ready, it’s then made available for understanding and gaining better insights or the business intelligence tools. Here, data helps identify actionable insights that can help the company to understand which customer segment are creating value and which aren’t.
• Activate: Data is converted as per the compatibility of external platforms and then shared across all of them. This data is ready to be utilized for the Personalization of Messages, websites, emails, etc. Based on this data, effective marketing strategies and business rules are created. And this data are sent to applications which use Machine Learning and AI to identify and predict customer outcomes, such as the Churn Rate of Customers, Recommendations, etc. Data can further is utilised for acquiring insights, which further leads to advanced product development as well.

www.treasuredata.com

Why it’s crucial to understand the Customer Data Platforms Architecture.

CDP architecture will help determine the right platform that can match your brand’s scope and scale. The platform’s database structure should be agile enough that it scales without breaking as your company data volume increases. CDPs can collect data from Structured and Unstructured data sources, but not all CDPs can do that.

So deciding which type of CDP is appropriate for the present and any additional features or functionalities which can be extended as and when required is paramount. Remember, crafting ROI from the CDP implementation is very critical; at the end of the day, your Business should generate higher returns. Using CDP, you will be able to identify ROI at the single customer level, and also it will help determine the CLV(Customer Lifetime Value).

The main challenge with Customer data platform architecture is building the base of the product which actually gathers data from different types of sources, where data is available in different format and have their own data platform architecture.

This image can help in understanding different components of CDP, which are ingested together to create this data platform architecture.

Five things to consider before opting for a CDP

• Integration:The platform should be capable enough to connect with all data sources(whether its first-party data, Second- and Third-party data sources) that your company might use. It can be Website Logs, Mobile Apps data, ERP data, Customer feedback, Analytics Platforms, CRM data, Offline Sources etc. All data variations should be considered. There should not be any data silos left after onboarding any CDP; otherwise, it will not fulfill the significant purpose.
• Scalable:You will not change your CDP platform now and then, so it’s essential to select a future-proof CDP, i.e., it should handle a large volume of data when there is a requirement. There will be thousands of single customer profiles, billions of events, n numbers of tables & segments will be created, and everything will be interlinked. Data volume becomes a significant challenge moving forward if it’s not accounted for in starting itself.
• Secure & Compliant:Your customer data is the holy grail for your BusinessBusiness. And privacy is what matters most for your customers. Users Private information needs to be kept at credible sources which are highly secured and almost unbreachable. A platform should have workflows that abide by the country’s compliance where your user resides, such as GDPR, CCPA compliance, etc.
• Flexibility:CDP should be flexible enough to be integrated with various types of internal and external platforms. Based on your convenience, you might incorporate platforms from different vendors across your organization. CDP should be created so that not just from CDP’s vendor applications, but all other applications and platforms from other vendors should also be compatible. And customized applications can be created on top of CDP if there is a requirement for the same.
• Clear & Transparent:What’s the use of any data if it’s not understandable by its users. All data should be precise and presented so that all users across the company should easily understand the data. Not just for the specific group of people such as data analysts but all the users. Reporting should be in such a manner that it can provide actionable insights. It should provide a complete view of the customer journey, right from the start till the last touchpoint.

Types of CDP Architecture(Database Structure)

Different data structures provide different capabilities in each CDP. There are broadly three different types of database structures on top of which CDPs are built.

• Relational Database Structure.
• Event-Based Database Structure.
• Profile Based Database Structure.

Relational Database Structure:- This is the most basic form of structure on which a CDP is built. In this DB(database), the size and relationship within the entities need to be defined at the start. The framework is highly structured. It is one of the major drawbacks of this type of architecture because whenever there is a requirement to increase the size of the database or any new entities requirement, one needs to create everything from scratch.

For example, the database size to handle the ‘X’ volume of data has to be predetermined/calculated and based on that, DB will be created. The relationship between two entities, which can click any particular section of the website by an anonymous visitor, needs to be linked to the Marketing Campaign. In this database model, the Campaign becomes the center point of everything. Everything needs to be connected or linked with the Campaign, as this type of CDP is majorly created for Campaign Management, so everything revolves around that.

This structure is not suitable for marketers. Flexibility is the major challenge in this type of structure. One has to be confined with the limited design. And because of this inflexibility, one cannot define their taxonomy within this structure.

Event-Based Database Structure: This type of CDP can collect a huge volume of raw data be it, unstructured or structured, based on events done by users on different touchpoints, such as App, Website events, customer interactions with Ads, etc. And the type of events can click on ads, scrolling on the website, clicking a particular section of the website, filling the form, etc. The database structure is not restricted at the initial stage, just like how it’s done in Relational CDP. Whenever there is a new requirement of adding or removing any required entities, that can be done.

Though this type of architecture can collect a large volume of event-based data, it still lacks one major thing to unify that data with the customer profile. Not all types of data can be easily linked to a user profile. You can collect data and segment customers based on specific event attributes or user behavior as a group. You can export that data to any external platforms for targeting, remarketing, and Personalized Marketing Campaigns as well, but not at each customer level.

It’s more like how a DMP functions. But the main difference between CDP and DMP is that CDP works with both known and anonymous data, but DMP only works with anonymous data. To know in detail about the difference is both platforms, you can check our article –

Understanding the difference between CDP and Data Management Platforms.

This type of platform is suitable for Campaign Management, Behavioural Segmentation, Ads Remarketing, Personalized Customer Experience, etc., but the main thing this CDP cannot do is create a single unified customer profile and build the complete journey map across all customer touchpoints, which is not possible in this type of architecture.

Profile-Based Database Structure: If you want to have a CDP that can genuinely take you the depth of customer insights and help you leverage the ultimate power of CDP technology, then this is the CDP you are looking for. With this CDP, you can get the most comprehensive and updated information about your users. This CDP overcomes all the limitations which are there in both Relational and Event-Based CDPs. You will get data at the individual level, taking your targeting capabilities to the next level.

Everything in this type of unified customer database is linked with the Individual customer’s PII(Personal Identifiable Information) – it can be an email address or phone number. A persistent and seamless journey map is created and made available for all external platforms throughout the customer journey, almost in real-time.

Apart from the functionalities available on both Relational and Event-based CDPs, Single Customer profile-based CDPs can help real-time personalization of ads, messages, and information delivery at the individual level. You can provide assistance to your customers customized for their particular needs and requirements. These CDPs can help you create buyer personas with accurate Identity-resolution, Synthesis, and Activation.

There are many benefits of this data platform architecture, such as a lot of flexibility and scalability. The CDP can automatically update audience details and synchronize them across all external tools(whether they are marketing tools or anything else). You can get details of customers, who are the high-value ones. Personalized and Consistent messaging across all channels. Overall, it will help improve customer engagement and loyalty, which will help with increased revenue and optimized costs. 

Basically it act like a booster to your First-party data and provides an enhanced and powerful mechanism, which your Marketing Team can utilise to understand Customer Behavior and improvise Business Processes. 

How to decide which type of CDP Architecture to choose for your Business?

It seems like a no-brainer, right. It’s obvious; one would want to go ahead and go with the Profile-Based CDP. And why not? It’s the ultimate MarTech, stacked with Advanced Analytics, Customer Intelligence and Customer Journey Mapping capabilities, which you were looking for. With all these functionalities and options, you can create a fantastic customer experience model, which provides a 360-Degree view on your valuable customers and provide valuable information about entire customer journey. 

But as they say, everything has its price. Having a profile-based structure can come with a high added cost. Creating a single unified customer view is an arduous task to accomplish. A lot of parameters are taken into consideration to make that kind of seamless data platform architecture.

If you are starting this customer-centricity journey, you can take one step at a time to move ahead. There is also a requirement to have a good volume of users coming to your site or app(any point of interactions), so this audience profile segmentation can help you at scale. With significantly fewer users, the costing model will not make sense. Plus, you may need to hire some experts at your end, to have a seamless and smooth experience with your CDP implementation process.

You also need to check the compatibility of your current marketing tech stack and Marketing Automation tools with the CDP. It’s advisable first to have a thorough understanding of the products and the functionalities provided by the same. Be clear on your requirements; jot down all the conditions you are having, and compare it with the functionalities of CDP, which can have a significant impact on your Business Goals.

Evaluate following things properly to get the best from available options:

1. What type of Customer Info with what customer attributes are required to be added.

2. Are the Customer Records safe in the customer data warehouse.

3. Does platform has advance capabilities such as deep analytics or artificial intelligence support. 

4. What type of Customer Support Programs or Customer Service does vendor has?

5. Privacy Regulation such as GDPR and CCPA is followed or not?

6. Data is Collected at what Period of time?

Talking to various CDP vendors and getting complete picture, comparing them in all aspects, especially your tech stack compatibility and pricing structure, is very important. 

Conclusion

In the end, it’s essential to understand that you are doing this to increase the overall Marketing ROI and create value for your Business Users. You want to provide the best Personalized experience possible but also considering cost-effectiveness. Getting hang of the technology and working behind the customer data platform architecture is itself a massive task. It’s great that you have taken steps to understand the technology behind the CDP. Enhanced customer experiences means better engagement & loyal customers, and CDP can help achieve your goals.