CDPs: Extending the Table for Better Plates

Looking back, the Customer Data Platform has been the core technology concept that ties my whole career together across several ventures.

😍 I was pleasantly surprised when I asked your friend and mine, ChatGPT, for its summary on the origin of the term CDP and the best modern examples. It correctly explained the core concepts of Data Collection, Unification, and Activation.

It cited Adobe Experience Platform in its shortlist of 6, a key part of which was Audience Manager – a Data Management Platform I built circa 2009-2011 as part of Demdex prior to its acquisition by Adobe!

🤔 Maybe I should be taking more credit for building one of the first scalable CDPs, a few years before the founding of Segment, TreasureData, or mParticle – three other examples of unicorns on ChatGPT’s shortlist?

On the other hand, it’s clear to me that CDPs have not gotten the market penetration they should, and are seldom being used to their fullest potential in the companies where they are deployed. Why is this?

💡 My hunch: Most companies end up “Table-shaped” when it comes to their Customer Data.

By which I mean they have 1 or more ‘Legs’ which represent an application back-end (APIs, database) or a data warehouse, which prop up ‘Plates’ which represent the delicious customer outcomes of good marketing, front-end app journeys, landing pages, chat widgets, analytics, A/B testing tools, etc.

These plates are usually from a variety of vendors for most companies, and thus don’t share data (our ‘Ingredients’ in this tortured metaphor) effectively with each other – or with the legs propping the whole thing up. So you end up with a table that lacks a smooth surface to allow easy building and sharing of more plates with more food, and a lot of frustration when it comes to “extending the table”.

🍽 Want more data faster on your plates?

You end up either writing a lot more expensive bespoke code on each of your table legs (in the case of smaller companies with empowered engineering teams).

Or you get frustrated enough to kick off a massively disruptive and expensive transformation project that never fully solves the problem (in the case of larger more fragmented enterprises).

The process either way likely results in scenes like the one in this picture (if you know, you know). (╯°□°)╯︵ ┻━┻

Meanwhile, most CDPs use cases end up reverting to “CRM with better marketing data” or “BI dashboards with better analytics” because those are the easiest outcomes.

🧠 Can emerging GenAI capabilities push the CDP to the forefront? Can this replace or drastically reduce the overhead of your application back-end to create a smooth table top onto which we can make all the plates fuller with less incremental effort?

I think it can, which is why I’m helping build the future.

If this post made you hungry for more, comment or get in touch so I can better learn about your challenges extending the table! 😋