We review CDP basics here before diving into the key steps you’ll need to take for a successful customer data platform implementation.
What can a CDP do for your business?
Do you think you understand your customers well? Do you have complex MarTech systems with a varied amount of sources? Do you know if you are able to personalize brand messages and campaigns to high value or lookalike customers easily and using first-party data?
These are some questions you should ask while thinking about a CDP. Customer data platforms bring all of your customer data together, stitches profile, creates segments, and allows brands to easily execute and activate personalization and campaigns. Understanding what a CDP can do for your organization will help you make the CDP implementation journey easy, so this blog starts with the basics. If you’ve already made up your mind about which CDP to use, skip ahead to the “Key Steps” section.
Customer data platform basics / What is a CDP?
What does a CDP do?
How does a CDP work?
What is the primary benefit of using a CDP?
What are the primary features of customer data platforms?
- Data Collection – CDPs ingest inputs from any customer data source and store it in a usable format.
- Profile Unification – CDPs associate and resolve IDs to a person, append external data, and de-duplicate profiles. This resolves ad exposure and conversion to a person to ensure maximum accuracy of frequency, reach, and conversion.
- Segmentation- CDPs create customer segments that can be activated on in multiple destinations/channels. Prediction and Decision (optional) – Some CDPs offer out-of-the-box machine learning (ML) modeling and some import custom models from SAS, R, etc., which can be used to deliver content or offer optimization, next-best actions recommendations, and journey analytics.
- Activation – CDPs connect audiences or segments to (external) activation systems (e.g., email, social, website, commerce, advertising, loT).
Who needs a CDP?
- Multichannel marketers with integrated marketing goals (as opposed to siloed marketing channel units pursuing their own channel goals irrespective of customer experience)
- Large-scale advertisers
- Marketing executives with automated customer lifecycle marketing initiatives
- Marketing executives with limited technical resources in their marketing department, since the CDP does require some technical expertise. Marketing executives with challenges accessing and leveraging customer data sources
- Marketing executives striving to communicate to customers with onsite personalization
What are use cases solved by implementing a CDP?
- Unify sales, analytics, and marketing performance data
- Calculate and predict customer lifetime value (CLV/ LTV)
- Cross-device or cross-channel personalization at scale
- Targeted marketing for cross-device or cross-channel activation
- Capture consent records
How do I select a CDP?
- The first step in selecting the right CDP is to define your goals and desired business impact.
- Next, define your use cases and requirements. (We focus a lot more on these steps in our “How to Select a CDP” video. Check it out.)
- When you’ve done that, prioritize your use cases and requirements.
- Evaluate different platforms across your use cases and requirements.
- Request demonstrations from the platforms on your shortlist.
- Finally, select a platform.
How long does it take to implement a CDP?
A smaller CDP can be rolled out for initial use cases in 90 days. As POC use cases are addressed, then the processes are automated and deployment issues are addressed, which could take 6 months in total.
Larger CDP implementations with a base of two or three use cases can be fully implemented within six months, and it should take around a year to realize its full potential.
Within six months of implementing a CDP, with people and processes in place, you should expect to see personalization and segmentation reduce costs and drive higher revenue. Your CDP will mature from managing simple campaigns and newsletters through multichannel journey orchestration, A/B testing, and some personalization and triggers through full alignment of people, processes, and technology.
Within a year, the CDP should be providing democratized data across your teams, which can lead to further value-generating initiatives with AI and Next Best Action informing communications.
Within two years of implementing a CDP, you can expect to see increased agility, efficiency, and effectiveness as your organization adopts a customer-centric, data-driven culture.
Key steps for CDP implementation success
A successful CDP implementation helps a brand become more customer-focused & data-driven. A good implementation helps you understand your customers better and enhance your customer experience, which results in increased revenue and cost savings.
All of this can’t happen without communication across teams, however, so it’s important to plan, gather information, determine your strategies for segments and identity resolution, and to define your plan for action (set expectations and prioritize how you’ll use the data), and finally, to monitor the results of your CDP implementation so that you can measure and report on the value. Keep this broader arc of goals in mind as you take these nitty-gritty steps to implement your chosen CDP.
At Search Discovery, we break up our services around CDP consulting into three steps: vendor evaluation, CDP implementation, and CDP activation.
Step 1: Discovery
For a successful CDP implementation, It is important to set customer-centric goals. Start with identifying the right stakeholders and implementers within the organization. Learn what they know about the customer and your data, as well as their expectations from CDP. To start this process, do the following:
- Identify the critical data sources and tools/platforms that exist within the organization that support the enterprise data strategy and roadmaps.
- Based on the knowledge, set simple goals and do use case workshops.
- Prioritize and choose a few initial use cases that align with the goals.
- While choosing the use cases for initial implementation, score use cases with business impact and technical feasibility.
- If needed, do a few vendor demos to understand the CDP product offerings.
Step 2: Technical Solution Design
- Document technical requirements for data ingestion of the required data sources.
- Check CDP’s capabilities for batch v/s streaming needs.
- Determine constraints and capabilities to capture data changes.
- Create an Entity Relationship Diagram (ERD) – Map the data source architecture.
- Setup sources to pull data from files and connect API services to CDP to pull data on schedule
- Data mapping – If CDP requires a schema, make sure it is structured and formatted as required. A CDP that favors schema-less ingestion allows various types of data sources to be easily ingested and integrated.
Step 3: Data Ingestion & Configuration
- Data preparation – Create ETL transformations and perform any de-duplication and filtering if required.
- Review consent records – These should include whether a visitor opted into various levels of targeting/tracking and on which platforms.
- Configuration – Model the data and configure the design
- Identity resolution
- If your ID resolution is simple and known, you can use deterministic matching to merge incoming customer data based on known attributes like phone number, emails etc. With first-party data, deterministic matching works well as it’s directly available from the customer.
- If your CDP allows probabilistic matching, and you are doing fuzzy matching to leverage IP addresses, browser, devices, and other behaviors, you can do probabilistic matching that gives some statistical ways of stitching identities but does come with a margin of error. This is used more for anonymous user acquisition.
- Some CDPs leverage AI to improve accuracy and prediction with identity resolutions. A few CDPs are also integrated with third-party identity providers for ID resolution.
- Ingestion – Create data connections and delivery
Step 4: Deployment
- Create the segments for activation.
- Based on the use cases, use audience creation tools to create customer groups based on profile and behavioral characteristics.
- These audiences will be used to activate campaigns and orchestrate journeys for the customers in the audience group.
- If the CDP allows, check overlaps and distribution in comparison to similar audiences.
- Some CDPs might allow you to check out lookalike audiences using their modeling, or allow you to use your custom model. You can also get lookalike audiences in the activation platform where you are sending the audiences.
- Send the audience to the destinations for activation.
- You may create either consent groups and/or exit audiences to exclude people from the activation or from the customer journey. One example is to suppress ads for people who are already customers.