Goodbye Customer Data Platforms

Last night Myles Younger shared the a16z primer The Rise of the Composable CDP. As I was reading I was confused when it was written. Composable CDPs rose about 2 years ago after the SaaS vendorscape of CDPs with their long integrations and high fees were platformed themselves by what they were running on, namely Snowflake. 

I do understand why it was written. Clearly their portfolio company ActionIQ (AIQ) is trying to reposition as a composable CDP. I tweeted about that a month ago.

As you can see this piece co-written by AIQ CEO puts AIQ into 2 buckets. Hardly composable.

It’s also a given this slide is “talking your own book” as there are a number of other a16z investments in those blocks and many other companies left out.

The chart should not be seen as even the current state of the industry let alone the future state as anyone who is actively involved in the ascendance of cloud data knows. 

Nothing against AIQ. Tasso and team are brilliant database experts. I spent some time with them in 2018 and I was blown away. 5 years ago AIQ was new and best in breed. However, that was long before the data warehouse and master data management became table stakes for serious businesses. It was before the wave of reverse ETL (which a16z also invested in with Census). It was before the cloud companies themselves decided to roll up on marketing and advertising

The term “composable CDP” was necessary to differentiate these rETL companies like Census and Hightouch from stand alone CDPs. And it worked. In response the industry has rebranded stand-alone CDP first as “integrated CDP” and then as “packaged CDP.” Integrated and packaged into what you ask? The data warehouse of course.

The whole thing is silly and confusing to the market. For VCs hundreds of millions of dollars in value is at stake with categorization so the attempt to reframe is understandable. The most value for any startup is when you can create a market.

However, what some people are whispering – and I will stand on the side of and give some voice in this post – is that with the data warehouse and the surrounding and ascendant ecosystem of applications and microservices (composability) we have not only moved past CDP vendors, we have actually moved past CDP all together. 

The customer data platform is the data warehouse. 

That statement is nerve racking for many since it’s not the Martech SaaS companies of the last decade that are going to be built on this data platform. Heck they shouldn’t be. Conversion rates have been stuck at 3% for an entire generation of marketing technology.  

The term I most like is “Warehouse Native” as Myles Younger presented it recently in his presentation about the convergence of MarTech and AdTech.

The brand’s cloud is going to handle everything data and the data warehouse or lakehouse is the platform. Owned and paid, merchandising and product, finance and forecasts. Applications and services will operate inside databases filled with rich AI ready data. That is what is actually rising.

The bigger market implication to this outside of obliterating an acronym is that pricing models are going to change. The end of CDP might also be the end of something bigger. SaaS itself.

The future is consumption based pricing models where seats/users are unlimited and credits based on execution and computation align better to customers and the warehouses themselves. AI is already making headwinds in this area with token based pricing meaning every single AI that sits as native application in the data warehouse will certainly be priced this way. 

So with deference to Sir Elton John, Goodbye CDP, where the dogs of society howl. I’ve finally found the future lies beyond the CDP.

Note 1: I wrote about native applications in the cloud here.

Note 2: Like Myles I also wrote about the convergence of MarTech and AdTech in the cloud.

Note 3: I am talking about my own book too as I have started a company doing AI in the data warehouse as a native application. 


2 responses to “Goodbye Customer Data Platforms”

  1. Tasso Argyros Avatar

    Hi Jon, good to see you are staying on top of things! I am not sure why you think ActionIQ is not Composable? You probably have an old picture of our product (as you said, it’s been 5 years.)

    At this point, ActionIQ can leverage the data lake, natively, for both storage and processing and push data to the channels.

    So really not different from what a Hightouch or Census would do, except we have a lot more functionality on top and we are proven in many F1000 Enterprises across every industry.

    As you mention, we are database people (my previous startup was an actual MPP database and I studied databases in the PhD at Stanford) and we understand the modern data stack better than most vendors in the space. We “disrupted” ourselves and our product to make it Composable because we think that’s the future of the industry. That’s our “HybridCompute” technology and you can read more details on our website (or happy to walk you through it). You’ll see it’s the real deal and not some marketing fluff.

    Either way, appreciate you covering the space and contributing to the conversation.


    1. jonathanmendezblog Avatar

      Thanks for the comment Tasso. I’m sure it’s the real deal. You and team are great database technologists. I guess where we differ is on how to define “composable” and part of my issue is that market confusion is making the term meaningless. One is either warehouse native or not and to me if you are composable it means you are an application or service that does one thing very well in the stack. Of course everyone is trying to bundle as soon as things are unbundled. I don’t see how you can be “composable” if you are bundling apps/services. That is my opinion but we can agree to disagree. I do think the market is very confused and better naming conventions are needed.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create a website or blog at

%d bloggers like this: