On the Requirement for Customer Data Platforms
I first encountered the requirement for Customer Data Platforms long before I had ever heard of them. This was back in 2014, and while David Raab had coined the term CDP as early as 2013, I didn’t know that yet. I was doing a discovery call with a big important client, and I was looking for a data source for the simplest possible use case. I needed an up-to-date list of current customers, so we could exclude them from a campaign.
This turned out to be a surprisingly difficult request, nobody seemed to have a data source that could tell me easily who the client’s current customers actually were, so I was on the phone with the client’s top marketing automation expert trying to figure out where I could get this data. I knew I was on to something because I really got him talking and he was holding forth at length about their data architecture when he came up with this gem of deep wisdom:
“In an ideal world,” he said, in the kind of dreamy sing-song voice people use when talking about a fantastical utopia, “[marketing automation] would feed off this universe that had all the information, right, we could just stick a nice ETL tool between it and start just plugging away, and unfortunately we just don’t have that big data store out there that has all the information we’re looking for.”
Wow, I thought, what an excellent idea!
Most people (especially at that time) just plug their MAP into their CRM and call it good. The CRM is treated as the Source of Truth, and they just segment their audience by whatever data they get through CRM integration. However, CRM data is notoriously messy and incoherent. CRMs were not primarily designed to be data cleansing and repository tools. All the data you need is rarely in the CRM – the CRM wasn’t really designed to be a central data hub and integration point.
Having a central repository of clean, comprehensive data to connect to marketing automation would solve so many problems with campaign segmentation, message personalization, reporting and analytics, and more.
There is a traditional approach to solving this problem. The client in the example I gave above attempted to do so – they made several concurrent attempts, in fact. These proved unsatisfying because they used the approach known as “EDW” (Enterprise Data Warehouse).
Unfortunately, these EDW projects are huge and cumbersome. The data model needs to be explicitly specified. This means making a lot of decisions, and at the enterprise scale, that’s a slow process. The result is a multi-year, multi-million dollar project that rarely gets you everything you want. In our example, after two years of effort the client was able to provide us with a periodic export of an extremely messy customer list, most of which was missing email addresses. Not helping.
So marketing and sales continued to languish in this messy data environment, with little hope in sight.
I ran into David Raab at the MarTech Conference in 2017. He was on fire about this new technology he was calling “CDPs” (Customer Data Platforms). He had set up this organization he called the CDP Institute to promote the category. He gave an amazing presentation in which he played a song by his son called the “Customer Data Platform Blues” and I thought “wow, that really sounds like the technology we’ve all been waiting for!”
I was nervous as a 15-year-old fanboy to talk to him, he’s one of my idols, but I got up the courage to go up and ask him straight-up, “is this the tech we’ve all been waiting for, is this going to revolutionize our field?”
He said, “Well, not quite, because in a few years machine learning is going to take over everything, and that’s going to be really revolutionary. But in the meantime, yes, this is a solution that can solve a lot of the problems we’ve been having with marketing data for a long time, right now.”
Since then I’ve been playing with CDPs every chance I get, and I’m convinced that the technology has enormous potential. CDPs generally use a non-relational database to consolidate marketing and sales data. This is much faster and easier than the EDW approach, because you can combine data from many sources into a single database without necessarily specifying a complete data model or mapping all of the fields to each other. There are drawbacks to this approach, the consolidated data you get may not be completely unified and normalized unless you put in the effort to do so, but you can pick and choose which data you want to map and which data you don’t.
Areas where a CDP can genuinely help a B2B revenue program:
- Account-level data consolidation – get your account data structure cleaned up, get all of your people associated with accounts, and measure all of your metrics at the account level.
- Segmentation – segment your audience based on data from all of your data sources, not just what’s available in the MAP. This could include data from web analytics, data from your accounting and customer tracking systems, even data from within a SAAS product.
- Customer journey tracking – with all of your data about your customer consolidated in one place, you can get a unified view of where they are in their customer journey.
- Personalization – data consolidation enables accurate personalization of content based on all the data you have about the prospect, avoiding the risks of data duplication and inaccurate data leading to incorrect personalization.
- Analytics – with your data consolidated and cleaned up, you can get comprehensive analytics all in one place, based on a single source of truth so that you can track the metrics you really care about across all of your systems and processes.
CDPs are not the ultimate solution to all of our data problems – I think Raab is right that AI will continue to revolutionize the field going forward – but they solve a lot of problems quickly and easily that used to be time-consuming and difficult. With this technology we can do in a few weeks what used to take months or years, and that translates into revenue growth in a world where the fast fish eats the slow fish. If you’re having trouble getting your data together, I would highly recommend you investigate CDPs.
If you need help evaluating and/or implementing a CDP in your organization, give us a call. We can help you pick the right technology for you and help you set it up to integrate with your existing tech stack and revenue operations. We can help you configure your data model and consolidate your data into a single comprehensive customer view.