During and after my time as the lead of Debezium, a widely used open-source platform for Change Data Capture (CDC) for a variety of database, I got repeatedly asked whether I’d be interested in creating a company around CDC. VCs, including wellknown household names, did and do reach out to me, pitching this idea.

On the surface, this sounds tempting. CDC, and Debezium in particular, are widely used in the data sphere. So taking a few million seed capital and building CDC-as-a-Service sounds like an attractive idea, doesn’t it? Living the start-up life and creating a unicorn-to-be, oh what a sweet dream. But having worked on CDC for quite a few years now, I am convinced that this wouldn’t the right thing to do.

The reason being that CDC is a feature, not a product.

By that I mean that CDC is an incredibly powerful tool, a huge enabler for working with your data in real-time, enabling a wide range of use cases such replication, cache and search index updates, auditing, microservice data exchange, and many others. Liberty for your data—​rejoice!

But it’s that, an enabler. CDC isn’t really that useful on its own. You ingest data change events into Kafka, and then what? At the very least, you want to have sink connectors which take that data and put it elsewhere. For a successful product, you need to solve a problem people have. And that problem rarely is "Take my data from Postgres to Kafka", and much more often is "Take my data from Postgres to Snowflake/Elasticsearch/S3". Very often, you also want to put some processing to your change event streams, e.g. to filter, transform, denormalize, or aggregate them.

In my opinion, CDC makes sense as part of a cohesive data platform which integrates all these things. These, and more: also data governance, schema management, observability, quality management, etc. Another angle for CDC productization could be to marry it closely with a database. Imagine Postgres provided out of the box a Kafka broker endpoint to which you can subscribe for getting Debezium-formatted data change events. How cool would that be? But again, that’s a feature, not a product.

Now, there have been a few start-ups focused on CDC lately. Two that stuck out to me were Arcion and PeerDB: They got acquired quickly by Databricks and Clickhouse, respectively. As I suppose with the goal of turning them—​you’ll guess it—​into features of their data offerings.