CDPs and Identity Resolution

Originally published on October 4th, 2018

As you’ve probably heard, CDPs (Customer Data Platforms) are a hot category in marketing tech at the moment. What do CDPs promise to do? At a high level, they claim to deliver the long-sought “single view of the customer.” CDPs will collect all your customer data in one place, organize it, and activate it in any channel. This is not just a marketers dream, but would also unlock incredibly powerful use cases across the organization that I won’t get too much into here.

However, as Jonathan Mendez states in his excellent series of blog posts here, this idealistic “360 degree” view of customer data has been promised since the first CRM systems in 2000 — and has always fallen short. What makes CDPs so different than existing solutions?

The key lies in how they plan to achieve this goal — centering all your customer data around a first-party master ID. To accomplish this, CDPs need to offer what’s called “identity resolution.” According to Martech Today’s CDP Report, every CDP they reviewed offers this feature, apparently at parity across the board. Why is it so mandatory, and is it all the same?

What is Identity Resolution?

First, let’s start with a primer on Identity and Identity Resolution. I define “Identity Resolution” as connecting the unique IDs from every channel where a customer interacts with an organization.

To unpack that a bit, think about all the channels and devices customers and organizations use to communicate. Everything is either digital and physical, those categories can be broken down further. Here are a few of the most common unique IDs used to identify customers across channels and devices:

Physical:

• Postal Address
• Email Address
• Phone Number

Digital:

• Cookie
• Mobile ID (tablet and phone)
• IP address
• TV / set top box ID
• IoT Device ID
• And many others I’m sure I’ve missed…

A point of “identity data” could be a single linkage, such as a SHA2 email hash deterministically matched to a cookie ID. An “Identity Graph” consists of the sum total of linkages between any of these unique IDs.

Identity Resolution — in all its gritty detail — does matter

CDPs, as promised, provide the tools to collect an organization’s identity data and build its own Identity Graph. The reason why Identity Resolution is so critical to a CDP is because it’s the first and most important feature that a CDP will enable — all of the segmentation and activation that happens downstream depends on the scale and accuracy of the identity resolution. Detail should be paid to every CDP on how they accomplish the joining of identity data — its not all the same. Understanding the nuances between deterministic and probabilistic matching, the limitations of mapping lat-long mobile data to postal addresses, and when IP matching should and should not be used will emerge as critical to evaluating CDPs.

The Bad News — Most Identity Graphs are small

With the understanding of how identity graphs are compiled, you can quickly see who holds the largest identity graphs — Google, Facebook and Amazon (surprise!). Typically a single organizations’ Identity Graph is going to be smaller than these players, depending on 1) the scale of their digital assets and 2) their methods of distribution. For example, CPG companies selling through distributors often don’t have the digital customer engagement to trigger authentication events at the scale a popular ecommerce retailer like Warby Parker might. Mileage may vary depending on the type of business a marketer is in.

But let’s look at another large Identity Graph — LiveRamp. Rather than using their identity graph to deliver targeted ad inventory themselves (ie become a walled garden), LiveRamp pioneered data onboarding as a category, and offered it to advertisers directly. This portability of an advertisers 1st party data into any marketing platform became an essential marketing tool, and opened our eyes to the power of 1st party data and identity graphs.

Organizations can enhance their Identity Graph with 3rd party identity data

Over time, identity providers have proven more willing to license the underlying identity data directly. This shift in business models will give organizations leveraging a CDP the ability to augment their 1st party identity graph with 3rd party identity data. The ability to supplement a 1st party identity graph has the ability to dramatically improve the scale and accuracy of use cases that marketers are already implementing today — trigger-based messages acting on an abandoned cart is one of the easiest use cases to quickly scale up.

The next generation of data pipes are still being built

In an ideal world, this 3rd party data could be purchased in real-time on a customer-by-customer basis. Data budgets could dynamically incorporate customer acquisition costs and lifetime value estimates calculated by an organization’s CDP. Buying could be data vendor agnostic, with the buying criteria solely based on customer identity, recency and accuracy.

Unfortunately, most 3rd party data today is bought in bulk, shipped around slowly via FTP and tied to proprietary IDs. The increasing limitations of 3rd party cookies also hamper recognition and actionability of these data points. In better news, companies like Narrative.io are rebuilding the marketplace infrastructure required to make this level of 3rd party data transactions a reality. I believe there is a big opportunity to build pipes between CDPs for data sharing across organizations.

Takeaways

If LiveRamp showed us the power of identity graphs, CDPs will democratize them, allowing every organization the ability to harvest their own identity data and wield it effectively. As scrutiny increases on the features of CDPs, more attention will be paid to identity resolution. But capturing the existing identity data within an organization will only go so far — 3rd party identity resolution providers and next gen data marketplace infrastructure will emerge to fill the gaps.

Originally published at https://www.linkedin.com.