Agencies that build data infrastructure for companies, such as Dacker, have pushed the data warehouse into the spotlight. More and more companies now treat the warehouse as THE source of truth.
Out of that new source of truth, a new kind of Customer Data Platform (CDP) has emerged: the composable CDP. It keeps taking market share and is redefining how companies activate their customer data.
As a reminder, a Data Warehouse is a technology that brings structured data together from one or more sources so it can be compared and analyzed. Examples: BigQuery, Snowflake and so on.
Here is a short piece to make the key differences between the CDP options clear.
Contents:
- What is a packaged CDP?
- What is a composable CDP?
- So what actually separates packaged and composable CDPs?
- Which way should you go?
1. What is a packaged CDP?
A packaged CDP, also known as a traditional CDP, plays a part in collecting, storing, unifying and activating customer data — enough capability to earn the “packaged” label. It does all of it through its own system, offering a turnkey SaaS solution to a company that wants to track specific events or collect behavioral data on its app or website, and then activate that data in marketing and sales tools.
These solutions offer plenty of features: data unification through identity resolution rules, audience creation and management, and so on. Recently, some CDPs have built features that plug into your existing data warehouse and use that data to send it on to marketing destinations. All of it is integrated into the platform — though enabling some of it sometimes costs extra. 💰
The best-known solution on the market is still Segment, but there are others such as mParticle and more. Depending on your needs and each product's limits, one may suit you better than another. It is not a decision to take lightly, since these solutions are generally quite expensive.
2. What is a composable CDP?
The CDP composable, unlike the packaged CDP, offers only certain building blocks. It slots more easily into your existing data infrastructure, typically integrating with your data warehouse. That is the whole point of the composable CDP: it aims to be far more flexible to integrate. If what you need is to activate the data sitting in your warehouse, it may well be the right answer, because it addresses that specific need.
What these vendors usually put forward is that you get all the benefits of a traditional CDP while staying independent of any single vendor's ecosystem. The most widely used options on the market are Hightouch and DinMo (a French solution 🇫🇷)
What makes a composable CDP a good one?
- A composable CDP should integrate with and run on your own data infrastructure.
- A composable CDP should leave you in full control of how your data is structured and stored.
- It should fit your existing technology, processes and workflows.
Worth reading too: our articles on Customer Data Platforms 👈
3. So what actually separates packaged and composable CDPs?
You have probably worked it out from the definitions above. A composable CDP is more flexible; a packaged CDP gives you a turnkey solution.
The other important difference is how long use cases take to deliver. Some go live quickly, particularly through reverse ETL. Both packaged and composable CDPs offer that. But for other activation cases, the ones tied to on-site tracking, a traditional CDP (the kind with a data collection capability) usually takes far longer to implement, depending on how available your technical resources are. A packaged CDP project can run from 4 to 12 months; on average, count 6 to 8 months before a good implementation starts paying off.
4. Which way should you go?
When you are deciding between a composable CDP and a traditional one, what matters is the specific needs of your business and your teams.
You are choosing between the packaged option — more complete in principle, but potentially demanding a lot of implementation time and effort — and the composable CDP, which will slot neatly into your current data architecture (assuming you have a stable, reliable architecture with a data warehouse, for instance).
First-party data matters more and more for marketing performance and personalization, which makes the case for a CDP stronger than ever. For companies already investing in data infrastructure, or planning to, a composable CDP is the better bet. If you are starting from nothing, a packaged solution may be the wiser route. And your budget will weigh in too.
In closing
When you weigh a traditional CDP against a composable one, the central question is this: do you need a predefined, turnkey solution, or a flexible system shaped around your architecture? 🤔
Bear in mind that traditional CDPs such as Segment are following the market and now offer features that belong to composable CDPs, reverse ETL among them.
Even so, plenty of companies keep migrating to composable CDPs such as Hightouch or DinMo, usually for cost reasons and to simplify the setup.
If you would like to talk through your CDP questions, get in touch with us at Smart Bees.
