My thoughts on Databricks CDP
Today, Databricks launched a “CDP” called Customer Lake. I have two thoughts
First… welcome to the party, Databricks :) Since the beginning of Hightouch, we've been evangelizing that the data warehouse should be your Customer 360, and we're excited to see you build features like identity resolution to help data teams make that a reality. We've been waiting years for this. Sure, we have overlapping capabilities, but who cares– every company should have a strong Customer 360 in their data platform.
Second… modern marketing is deceptively complex. Being the CDP for marketing teams isn't a "feature." There are already plenty of lightweight "CDPs" bundled into every marketing tool. The reason customers trust Hightouch is because marketing teams recognize how foundational and critical the CDP is to modern marketing and they want a best-in-class solution.
We're excited to keep partnering with your team and your customers. We'll keep advocating for customers to own their data and their Customer 360 in the warehouse, whether that's Snowflake, Databricks, BigQuery, or any other platform they choose. We believe that's what's best for customers and the ecosystem.
That said, we believe in clarity, so here's just some of what it actually takes to serve modern marketing teams. We’ve learned this from serving 1,000+ customers.
And every marketing technology buyer knows you can't judge a feature by its cover. What matters is how it's built, how it handles nuance, how it fits your workflow, and whether it actually works the way it's described. Reaching product maturity takes time and real-world reps with customers to get right.
Plus, marketing never stops changing. We shipped 6 major CDP releases in 2025 alone. And just this week, we launched the Agentic CDP, which turns the CDP from a system of action into a proactive system of intelligence… one that actively suggests new campaigns and opportunities to marketing teams, using data from the warehouse and broader marketing context (past campaigns, creatives, calendar, product briefs, etc.).
In the meantime, we're here and we have no plans of slowing down. It's an ecosystem. We'll keep advocating for what we believe in: data sovereignty and best-in-class solutions built on top. Because that's the whole point of owning your data—freedom of choice.
I can't endorse this statement enough: "AdTech / Data Onboarding that combines 1st- and 3rd-party data, so companies can coordinate CRM and advertising in one platform instead of buying separate platforms." And I see company after company, agency after agency, simply not understand this. Or maybe they say they do, but their actions reinforce silos. I've been on many sides of the AdTech/MarTech split: At the center of the enterprise, in an LOB, as a practitioner in these channels, and now agency-side building products. For those not paying attention to Databricks and Hightouch, you may want to start skating to where the puck is going rather than to where it's been.
Enjoyed this point of view. I see a real differentiator being the edge case handling already baked into Hightouch from their 1000+ customers and feedback over time. I imagine that level of product feature readiness is hard to replicate quickly and is so valuable in a world where core feature sets can appear similar across tools.
When you are trying to be best-in-class, focus matters. Marketing ROI is so crucial that it's hard to settle for a generic product
I’m glad DB gets into the CDP game, it’s an explicit acknowledgement of something we’ve all known for a long time: a CDP is just a loose collection of parts that allow you to capture, federate and control user and event data. But just because you can do these things and call yourself a CDP doesn’t make you a great one. The whole reason I bought HT back in 2020 was because storage and federation are two completely different jobs, and require dramatically different skill sets. For example… Segment tried to become an RETL. Their solution hasn’t been updated in years, and it’s the details that makes it insufferable - recently went through this w a huge company… literally no alert monitoring log level details on syncs. But you need this to create a self fulfilling agent loop. I suspect the same thing will play out here: AI has made it easier to do stuff. More poorly for teams to claim something they are not. All the while the tools that matter are the ones that are paying attention to the details of how humans and AI use tools.
The term CDP is like using the term 'contractor' when one person thinks plumber and the other person thinks electrician. We need to start formally differentiating tactical customer activation capabilities specialised for media, CRM, platform personalisation etc that sit within so many other tech systems Vs something that genuinely orchestrates customer data and experiences in the way that actually drives true incremental growth across the entire plethora of customer touch points. I might see if I can get a fireside chat going between 4 major players all saying they're a CDP but with very different capabilities to really highlight this. Maybe you would join me for that...