What CDP buyers told us about cost, composable and who owns the data.
Hightouch bought Strata on October 9. Strata’s AI agent builds and maintains integrations, which is the part of a CDP that usually needs an engineer (Hightouch). That came six months after Hightouch raised $150M at a $2.75B valuation (Pulse).
Databricks wants in too. In June it previewed CustomerLake, a CDP built right on its lakehouse. CEO Ali Ghodsi said they’re “replacing legacy software.” Gartner expects 80% of new enterprise CDP deployments to be built into or composable with a data platform by 2030 (CMSWire).
So we went back through our interviews to see what buyers are actually doing. 57 buyers were in market for a CDP in the last 6 months.
Here’s what they told us.
It starts with the bill, not the architecture
A director at a $10M-$100M baby tech company rated Segment a 5/10 and is three months from turning it off:
“We did make the decision to offboard Segment, which is a big decision that is unfolding as we speak. So in the next three months, we will offboard Segment. We’re basically moving to the warehouse as a source of truth.”
Director Operations, $10M-$100M baby tech company
“…the reason for choosing Segment was the thought that this would be a centralized kind of end-to-end customer lifetime source of truth. This is just not going to work. It’s not a good way of doing that. We decided to take that big important function and internalize it to our warehouse. With that decision, it made no sense to pay Segment for the cost that they have…”
Director Operations, $10M-$100M baby tech company
Even happy customers are leaving. A marketing director at a $10M-$100M streaming media company rated Segment an 8.5/10 after seven years, and moved to Hightouch anyway:
“The only drawback ended up being cost because they are rated based on users, and as the company grows, that user count gets up there, and the cost could kind of balloon.”
Director Marketing, $10M-$100M streaming media company
Composable is landing on shortlists
54% of 35 Hightouch interviews rated it an 8 or higher
38% of 63 Segment interviews rated it an 8 or higher
Last 12 months. Ratings are the 1-10 satisfaction scores buyers gave when we asked them to rate their stack.
A product director at a $100M-$500M beauty subscription company is replacing Segment in the next 12 months:
“certain vendors who have composable Customer Data Platforms (CDPs) are finding their way to the top of our list, more so than more traditional CDPs.”
Director Product, $100M-$500M beauty subscription company
The appeal isn’t the warehouse itself. It’s what marketers can do once the data is there. A marketing director at a $1B-$10B lending company picked a composable CDP to get rid of manual list pulls:
“Composability was critical because we needed to ensure that we could compose the audience segments with a self-service tool. Previously, the challenge was there were armies of people in the organization focused on manually pulling the audience list every week…”
Director Marketing, $1B-$10B lending company
And when buyers switch, they want it to feel like nothing changed. That streaming media director on why Hightouch beat Adobe and RudderStack:
“…in terms of a seamless transition and making sure that the people in the C-suite don’t notice a change or the people who are building audiences don’t notice anything different, I think that we were closer aligned with Hightouch than RudderStack.”
Director Marketing, $10M-$100M streaming media company
The catch: someone has to own the warehouse
Composable doesn’t win every deal. A marketing director at a $1B-$10B apparel retailer liked Hightouch, and the company had it, but the team picked a packaged CDP instead:
“The main reason was concerns about time to onboard that platform. Honestly, it was a political decision around who manages it, since that composable Customer Data Platform approach relies on data in the data warehouse, and that is managed by the analytics team, which the business didn’t want to be reliant on.”
Director Marketing, $1B-$10B apparel retailer
That’s the real fight. Warehouse-first means marketing depends on the data team. Some marketing teams want that. Plenty don’t.
What to do about it
- If you sell a packaged CDP, your pricing is the attack surface. Pricing that grows with users or profiles is what pushed out a buyer who rated Segment an 8.5. Get ahead of the renewal conversation before your customer runs the math on the warehouse they already pay for.
- If you sell composable, sell to marketing, not just data. The deal you lose is the one where marketing doesn’t want to wait on the analytics team. Show the self-service audience building first, the architecture second.
- If you sit on top of the CDP, connect to the warehouse directly. ESPs, personalization and ad tools that read straight from Snowflake or Databricks stay in the stack no matter which CDP wins. Braze and Adobe are already CustomerLake launch partners.
Moving the CDP into the warehouse sounds like a data decision. In these interviews it was a budget decision and a who-owns-what decision. Sell to both of those, and you’ll be fine wherever the data ends up.
Jon
P.S. We’ve got every CDP conversation on the platform, including how each buyer rated Segment, Hightouch and the rest, and who’s up for renewal. If you want the breakdown, LMK. Best case, you spot your next opp. Worst case, you know what you’re up against.