Reverse ETL CDPs vs Traditional CDPs
You have a data warehouse, which is your single source of truth when it comes to what your business does online and offline. However, activating that data can be very difficult, time consuming and expensive in people, time and money.
You would need to build an ‘Activation layer’ to action your data against internal and external API’s, you would need to build audiences/segments against the data and then you need to work out what the trigger points should be, how and when to connect and so on. Oh, you will also need to keep up to date on API’s that change over time (Google are very good at doing this…). You may have also have tried a Customer Data Platform (CDP) before, thinking this would solve all your data activation problems but maybe it didn’t quite bring the return on investment you had hoped for… Not unusual to hear either…
However, fear not! Although CDPs have been around for a while now, Reverse ETL CDPs are still reasonbly new within the marketplace and are becoming more popular, for good reason!
More traditional CDPs are becoming reverse ETL ‘compatible’ - whilst that strikes me as being a car automaker who used to use engines but is now going electric and ‘adapting’ your drivetrain to ‘fit’ the new propulstion system, we predict you are likely to continue seeing more and more ‘traditional’ CDPs becoming ‘Reverse ETL’ compatible.
Instead of needing to take your data warehouse to a CDP, bring the CDP to your data warehouse! Let me explain…
First of all, ETL stands for Extract, Transform, Load - this is a process of extracting data from one place, transforming the data to be able to load it somewhere else. Most cloud based data storage platforms used by most businesses these days tend to fit into one of three categories - Microsoft Azure (not that common, at the moment), Amazon Web Services (AWS) or Google Cloud Platform (GCP). Lately, a LOT has happened in the GCP space and I have personally found AWS to be ‘struggling’ to keep up with the level of integration and innovation - that being owned by Google, currently.
You might also of course have a more traditional Teradata, Oracle or Microsoft data warehouse - and that’s great, but you can still benefit from working with reverse ETL CDPs, you’ll just not have the same level of flexibility and integrationc capabilities as a reverse cloud based data warehouse would. More on that for another day…
So when we talk about Reverse ETL CDPs, now you know what you could do with such a platform - how do they compare with the more ‘traditional’ CDPs?
As always, the quality of data inside a CDP - whether traditional or reverse ETL depends on the data at source. For most, this depends on the data layer, in app or in web. Although a data layer isn’t technically needed to use a CDP, it is most definitely recommended! Tag Management Systems like Google Tag Manager, Tag Commander, Tealium iQ and others typically handle the data available in the data layer in your app or web pages. That data is consumed and passed down to the CDP for audience building and activation purposes.
However, sometimes data layers miss a key ‘ingredient’ needed for audience building and activation. With reverse ETL CDPs, that data might exist in an external data warehouse instead! This will enable the data attribute to become available to configure in audience builds and data points to go to data end points, whether display advertising, personalisation, search bid management platforms, realtime scoring and more.
There are three main differences between traditional CDPs and reverse ETL CDPs. It is worth knowing which is right for your business and its goals before making the jump into a CDP as picking and implementing a CDP can take time and in the end, the return on investment may not be what you were expecting or hoping for. Simply having a CDP alone will not solve your data activation problems…
Due to their design, traditional CDPs have a more limited support for specific use cases, whilst Reverse ETL CDPs are considered to enable and provide more flexibility through more complexity, thanks to their design.
It is probably true that the traditional CDP is generally considered more ‘user friendly’, whereas reverse ETL CDPs can require a little more consideration and planning around use case design and implementation.
Traditional CDPs are widely considered realtime platforms (great for companies closely aligned with marketing goals), whereas reverse ETL CDPs are generally more considered ‘batched’. However, don’t forget about the realtime Reverse ETL CDP’s also…
When it comes to deciding which type of CDP to use (vendor choice aside), it is important to understand what your use cases are. If you don’t know, then you could make a poorly informed decision and trust us, the vendors won’t help you here…
Traditional CDPs are considered the more ‘marketing friendly’ platforms. If your business is one which can change/pivot frequently depending on the platform, such as ecommerce, then Traditional CDPs might be better for you. However, if your use cases and business needs are more complex, or your environment more structured around regulation or process such as insurance or financial services, or where data modelling for real time decisions around fraud, money laundering, subscriber churn or modelling for affinity on products, services or content, then reverse ETL CDPs could most definitely be a better choice.
Some of our customers have asked if they can use both traditional CDPs and reverse ETL CDPs (at the same time!). Whilst technically you can, we would advise against it - you would be double tagging, doubling the audiences, making and sending multiple data sets to the same data end point creating duplication - making life rather complicated for yourself and your teams, but you would also be creating another fragmented data silo - this is the very problem CDPs try to address!
But if you’re still not convinced, consider this typical approach - If one CDP has a view of an individual (maybe someone who abandons a purchase) and the other CDP has a view of that same ‘someone’ being a buyer, which marketing message should be sent to that ‘someone’?
So although technically possible, we would not recommend it! You would be introducing a new level of complexity which would need careful execution and audience management.
You don’t need or want to make life any more complicated than it already is…!!
