In the previous article of the series “How to procure the best customer data platform for your firm”, we walked through the steps to take when acquiring a Customer Data Platform.  

So if you’ve read the article, congratulations! You’ve now successfully procured the best CDP! 

Now before you roll up your sleeves and explore all the great features of a CDP, let’s take a step back and set up a roadmap for the actual implementation. Don’t worry, we’ll guide you step by step during this process. 

Continue reading to learn valuable knowledge based on our experience and expertise at Remotion, that will make your implementation smoother.  

 

1. Understand the technology 

CDPs are built on core technology and IT-design methodologies, so you need to answer several technological questions to make the best use of your CDP: 

• Is it a Data Warehouse-like or Data Lake-like environment?   
• Is the data structure based on relations? 
• Does it support real-time data ingestion, or is it batch oriented? 
• Is it event driven?  

Think about the structure of your system: it might even have different pipelines and technology for different parts of the tool. And while channel data could be both supported in real-time and be event driven, other data could come from batch-oriented flows. 

Teaching every member of the project how the system works and how the technology is designed will enhance coordination and save time for the implementation.  

Therefore, understanding how the technology works should always be the first step when implementing for the first time or introducing new people into the project. 

 

2. Harmonize customer definitions 

The power of a CDP often lies in its ability to link data that is logged on different identities towards the same customer ID. Knowing your customer data-structure and tracked information, based on IDs, makes the implementation much easier.  

There are usually three types of customer IDs:  

• Customer ID number: internal arbitrary ID, often linked to a log-in. This is the key ID that should never be changed, as it often functions as a reference to all core customer data such as transactions, profile data… 

• Identifiers: email, phone number, name, personal ID number, or any other internal identifier that may be used in other systems.  
• Non-stable identifiers: cookie IDs, device IDs, IP-addresses… 

The main differences reside in how stable they are, how often they change and how many of them a user typically has. For example, a user could have one customer ID, two email accounts and could even use 15 browsers, meaning 15 cookies on which a company could theoretically log information.  

The power of a CDP is to keep linking the non-stable and stable identifiers towards the same customer ID. A CDP lets the non-stable IDs change linkage, for example, when multiple users share the same browser and log in and out from different accounts.  

Funny note: in the early days of digital business, customer IDs were often the same as account usernames.
Once one of our colleagues at Remotion wanted to change their old and cringe username on a music streaming platform before sharing their playlist (and to avoid being called MrCool1223 for the rest of their work life). They couldn’t do so because the username was also the customer ID. Instead, they had to create a new account and copy all the playlists with the help of customer service. 

We hope you’ll remember their story to avoid this mistake: if you want your customers to be able to change their account names, don’t use them as customer IDs! 

 

3. Set-up of instances 

An instance is basically a part of a system where your data is set. Even if the number of instances you should run can seem straight-forward, it is still important to carefully consider it, as having to change it later would be cumbersome.

When running two instances you have two set-ups of the same functions, but data is not shared among them. Running multiple instances can then become quite complex.  

These are the questions you should refer to when deciding on the number of instances: 

• Do you run several markets on varied language?
• Do you have a lot of separate brands within the same market?  
• Is the campaign management mainly local? 
• Is the offer significantly different between the markets and brands? 
• Is the customer behaviour different between the market and brands?  
• Do you have requirements linked to a specific brand or market? 
• Is sharing data from different markets or brands illegal?  

If you’ve answered yes to one of these questions, you could consider running multiple instances. But if you’re working with a large amount of markets, remember that each instance added increases the cost. 

 

4. Do not forget the product data 

It’s easy to focus solely on customer data when dealing with CDPs. After all, the C in CDP does stand for Customer.  However, in order to understand your clients you need to know the products they purchase and are interested in. And knowing only the name of your customer’s purchased product isn’t enough.

Additional information such as its price range, the quality and material, whether it’s a new product or was bought on sale, could tell you more about the customer’s interests and taste.  

Now you understand why product data is essential to understand your consumers. And isn’t that the main point of the CDP?

 

5. Anchor and align implementation roadmap with stakeholders 

Lastly, all relevant stakeholders should be aligned with the roadmap and all its constituents. 

We’ve discussed procurement and how having the right resources is an integral part of successful CDP work in the last article.
Having the right stakeholders is essential to facilitate the process of understanding technology, harmonizing customer definitions, and setting up instances. Compiling the right resources often entails setting up cross-functional teams with stakeholders from departments such as UX, paid media, digital marketing, BI, IT etc.

The brands who deploy cross-functional teams of stakeholders involved with customer data in varied ways most successfully utilize CDPs. A successful CDP-implementation will also ensure that marketing departments won’t always need BI or IT to gain access to customer data.  Instead, the CDP will help localize customer data to make it effortlessly available for different stakeholders.  

CDPs will also help dissolve information silos – which are created when different departments do not work cross-functionally and share information and knowledge. When the whole team is coordinated, the roadmap must be anchored to let everyone know their role and their duty in the implementation project.  

Some type of “buy-in” from all the different stakeholders is essential for the success of the project. If all stakeholders are on the same wavelength, the implementation will be smoother and facilitated, with the best possible outcome.  

 

(Country road) Take (me) home message 

After reviewing the procurement process of a CDP, we have now also explained a few important aspects that will arise during the implementation process.  

Understanding the technical and functional aspects of a CDP is a vital part of succeeding with the implementation and avoiding setbacks. Here at Remotion, we know from our knowledge and experience that an anchored roadmap and aligned stakeholders, with the right technical design and customer data, will facilitate the best possible CDP implementation. However, working with the CDP implementation is by no means linear. After setting the foundation and establishing the right routines, you have to scale up your CDP. 

 

Do you want to learn more about how to scale up and optimize your  CDP work? Then keep your eyes peeled on our LinkedIn page! We will cover this topic in the third and final part of our CDP-series.