Demystifying CDPs, Composable CDPs, ReverseETLs and Hybrids ????
There has been a lot of online chatter recently on what is a CDP versus some of the other new category terms, leading to heated (but entertaining!) arguments on social media and, worse, market confusion
First, if you haven't checked it out already, please read the a16z's blog post on the Rise of the Composable CDPs. It's a good starting point and among other things it offers a clear definition
But here's where the confusion comes in. Most of the "Composable CDP" vendors are really ReverseETL vendors that repositioned as CDPs late in the game. Early on, ReverseETL vendors explicitly avoided positioning as anything like a CDP, and in fact they were making many grandiose claims ("Friends don't let friends buy a CDP!", "You don't need a CDP! Use us and your Data Warehouse will become your CDP!"), which at some level were entertaining and rooted in those vendor's lack of understanding of what a CDP does versus doesn't do. Thankfully people do learn, and we rarely hear similar arguments these days.
On the other hand, from the core CDP vendors only ActionIQ has evolved so far to meet the Composable CDP definition. Moreover, we offer the only Hybrid solution in the market, meaning an Enterprise leveraging ActionIQ can have their CDP deployed in a Composable way (meaning, using their data warehouse/lake/lakehouse for storing and processing their data) but also they can have some data stored and processed in ActionIQ, as needed. That's totally a choice we leave up to our customers and it's transparent to the ActionIQ business user, making it a very powerful deployment tool.
In addition, ReverseETL vendors, showing up late in the CDP game, lack a lot of standard CDP functionality. For starters, those vendors have admitted that they don't meet the CDP Institute's definition of a CDP. And despite the ChatGPT hype, software still doesn't write itself! So having had a late start, they lack several critical CDP features
Moreover, the orthodoxy of ReverseETL products being only a push-down tool, versus having a Hybrid option like ActionIQ, looks benign on paper but removes much needed optionality in many real life situations
领英推荐
By the way, why do we know these things? Just go to actioniq.com and check out the long list of Global 2000 logos we have there. These are brands that we are proud and humble to call customers, most have partnered with us for many years now and together we have delivered outstanding business results
So given the above, we can now summarize the confusion. Most Composable CDPs, specifically the ReverseETL vendors, are not CDPs, per the CDP Institute's objective definition and what, in my experience, has historically been considered core/essential CDP functionality. But also, most CDPs except for ActionIQ (and whomever else is willing & able to truly evolve their core architecture in the future, which is not trivial) are not Composable CDPs despite the term similarity. And to make things worse, the same ReverseETL vendors who essentially renamed themselves to Composable CDPs after claiming "CDPs are dead" will sometimes quietly drop the "Composable" part... and even let some of their friends buy a CDP, as long as it's them ??
And to make things worse, the same ReverseETL vendors who essentially renamed themselves to Composable CDPs after claiming "CDPs are dead" will sometimes quietly drop the "Composable" part... and even let some of their friends buy a CDP, as long as it's them ??
At the end of the day, there may well be use cases where a ReverseETL tool can succeed, perhaps smaller/simpler environments with simpler data environments and fewer business user/CX/orchestration requirements. However, as the market tries to understand what tools is best for which job, it's good to keep in mind that calling yourself a CDP is cheap, but evolving a product's architecture
All that being said, we absolutely love the energy and debates in the space! And I look forward to continue seeing the broader category grow, proving all the early CDP skeptics wrong ??????!!
Martech / Digital marketing / CX
1 年It's been very entertaining entertainment in this debate.
Board Director | Forbes Hall of Fame Marketer | Detroit Advocate
1 年Great explanation and clarification. I highly recommend this for a better understanding of what every team should be learning and investing in.
Co-Founder, Stitch & Stitch Predict | media, martech & modelling
1 年Great write up Tasso ??
EIR at Balderton Capital, Independent Consultant, Author of Kellblog, and Co-host of SaaS Talk.
1 年Great piece, Tasso