We are in an interesting new era of marketing and technology. The last 20 years have been really fascinating to watch and be a participant in. I started my career in the first “walled garden” era, in which big publishers tightly controlled access to their audiences through the gates of salespeople and insertion orders. Big advertisers paid big money, upfront, to access television and print audiences at scale through the mechanism of salespeople.
Slowly, we entered the digital age where the same business model (salespeople and insertion orders) controlled scaled access to newer walled gardens like Yahoo! I remember signing some seven-figure deals for homepage inventory as a young marketing director. Those were the glory days.
Later, some really smart people figured out how to turn the industry upside down. With the browser cookie, marketers could gather access to people data and create their own segments — without paying the Yahoo! tax. Early ad networks started to find “business readers” and “soccer moms” at scale all over the internet, and the power shifted to the buy-side.
Soon thereafter, even smarter people figured out how to trade digital media programmatically, and the power shifted to technology companies and agency “trading desks” which took the lion’s share of advertising out of publishers’ pockets and into their own. Publishers were lucky to get 30 cents on the dollar for their premium digital inventory.
A long period of arbitrage-driven media occurred until a funny thing happened (but slowly, and over time): the principal mechanism for trading people data, the cookie, started to become less valuable and important. DMPs, the machines that publishers and advertisers used to manipulate this currency, faded away. First-party data become ascendant. E-mail addresses, postal mail identifiers, and mobile numbers became – once again – the currency of people data. Thus began the “CDP Era” in marketing, probably around 2016, and we are living in it now.
I recently wrote a book on the topic with my friend Martin Kihn (called “Customer Data Platforms” for lack of a better title), and we speculated that this era in which CDPs are the dominant technology for organizing and managing customer data, would be around for a long time and create many fundamental changes in how companies go to market, engage their customers, change the way we think about marketing and advertising, and even threaten to solve the seemingly impenetrable problem of attribution. Are we finally ready to deliver on the promise of the “right person, right message, right time?”
The answer is, “probably.”
The “probably” part depends on how we think about CDPs, and how we define the problem space. Here are some thoughts on what the new era might look like, based on some questions I have been asking myself.
What is the purpose of CDPs?
Typical answer: CDPs are a data management technology for capturing, transforming, unifying, segmenting, and activating first-party people data. Very true, and not wrong. But why? Generally, the answer is “to improve customer experience,” and that is not a wrong answer either. Better CX drives revenue, reduces churn, and helps people love brands. That’s why modern CMOs are investing, and why there are over 150 “CDPs” listed by the CDP Institute.
The problem is that most of us have too narrowly defined “customer experience” itself. While one of the primary drivers of this technology is, and will continue to be, driving more personalized experiences in marketing (better e-mail) and advertising (more relevant ads), true CX must go beyond those channels, and consider more human interactions – call center conversations, sales interactions, moments between an in-store customer and a clerk at the point of sales terminal. Even what happens at an ATM screen.
This “more than marketing” approach lies more in the domain of enterprise software companies, especially those with a CRM focus, as “customer relationship management” is the basis for provisioning those types of “real life” experiences. Traditional CRM is maturing quickly towards digital, and laying the infrastructure foundations to support more realtime capability, storage and processing of massively scaled data, and adding capabilities for pervasive intelligence (AI and ML) and automation to their stacks. How these advances over the next 3-5 years among the leaders in CRM will have the greatest impact on the category, as most software buyers are eager to consolidate their technology stacks, rather than be buffeted by yet another expensive “era” in data technology.
CDPs are Foundational
If we agree that CDPs are here to improve customer experience, and we also agree that we must define CX more broadly with a “beyond marketing” imperative, then the next question we need to ask is, “what are the things that will drive success in CDP?”
The first thing is obvious: identity. As I wrote 5 years ago, building a modern data management infrastructure begins and ends with mastering customer identity – the notion that knowing your customer is the key to everything and is the basis for the modern technology stack.
In this oversimplified view, identity – both known and unknown data management – is the foundational layer, driving success in intelligence (scaled, unified data drives success in machine learning) and orchestration (one “golden record” that can be activated across many disparate systems).
But the “know” element of the modern technology stack can’t just be about managing the many different keys and identities of consumers. Although it is critical to create a single profile from dozens of different identity keys (e-mail addresses, cookies, device IDs, mobile advertising IDs, et al), the real challenge is mastering the ability to continually enrich that profile with the metadata that drives value in experience. Namely, what are the last e-mails someone opened, the last five SKUs they bought online, the last three calls they had with the call center (and their outcome), and how many loyalty points do they have? Taken individually, any of those data points can drive outsized performance in any number of channels. Taken together, they can reveal true intent and can completely change the way brands engage consumers and go to market.
If we agree that better, unified customer data is the key to driving better intelligence and engagement), then what types of data truly matter, and how should we think about the next phase of CDP? I believe there are three fundamental shifts that are occurring in the CDP space, and each requires a new way of thinking:
Rethinking the Loaded Term “Identity”
The first thing to focus on should include recasting the way we think about “identity.” In a world where digital-only interactions with brands have skyrocketed from a baseline of 25% to over 50% in the past several years, accelerated by the pandemic (see McKinsey), the way customers access digital experiences is more important than ever. It’s very odd that a key foundation of customer identity management – the CIAM category – has not had a real seat at the table with marketers and advertisers.
Customer Information and Access Management, or CIAM – services including customer registration, self-service account management, consent and preference management, single sign-on (SSO), multi-factor authentication (MFA), access management, directory services, and data access governance – has lived almost exclusively in the domain of the IT buyer. For players in the marketing technology and adtech spaces, “identity” meant cross-device graphs, user matching with big platforms like Google, and maybe “onboarding” through services like LiveRamp. It’s interesting to consider that all of those critical services were delivered without the intervention of the customer. As we move into a world in which first-party data is ascendant, and privacy regulations demand more direct consent from customers themselves, CIAM capabilities become the core building block of the “know” layer of the stack.
This fundamental shift envisions much more collaboration between the CMO and CIO, a further alignment between traditional CRM and the marketing stack, and a much faster evolution of CDP from a marketing-specific packaged product offering and a fundamental layer of the overall enterprise software stack.
Rethinking “Commerce” Data
Why is “retargeting” still such a big market? Why do we seem to see ads for products we searched for, talked about, clicked on, or otherwise merely thought about appear in our social media feeds and on display ads? It’s obvious and effective – even a small amount of real purchase intent is a great predictor of what people will buy.
“If you bought this, then you’ll buy that” product recommendations and the marketing and advertising tactics they drive will continue to be a big part of the landscape going forward. However, as backend enterprise systems migrate closer to the marketing technology stack, there are opportunities to go beyond the obvious (retargeting, product-driven customer journeys, data-driven call center recommendations) and start thinking more broadly about what types of data drive revenue.
Consider this: Every commerce operation that fell to its knees during the pandemic had nothing to do with a failed shopping cart or broken check-out process, but rather, was a result of empty product catalogs caused by the collapse of supply chains. In other words, eCommerce failed during the pandemic not due to leveraging data-driven marketing tactics (commerce and marketing systems have been closely linked for some time), but due to the lack of connectivity between backend inventory management systems and front-end engagement tools. Basically, companies were marketing products that could not be fulfilled and delivered.
What if there was a tighter connection between the systems driving manufacturing on the back end, and the marketing of products on the front end? For centuries the way commerce worked was people predicted how many of something could be sold, manufactured it, and used marketing to create demand. Over the next decade, we will see that flipped on its head. People will buy what they want, and near-realtime supply chains will manufacture exactly what was ordered. This requires a new way of thinking. Show people products that don’t yet exist, create demand, manufacture, and deliver on-demand. This is starting to happen today, and it threatens to overturn everything the traditional marketer understands about demand generation. Plugging supply chain data into “CDP” is maybe the most interesting opportunity in the category.
Rethinking CDP as a Category
One of the arguments we made in the Customer Data Platforms book was that there were three types of CDPs:
The first category we called Systems of Insights, which are described systems that were principally concerned with managing customer data, creating a common information model, segmenting customers, and analyzing and activating data. These types of CDPs are most akin to CRM and MDM, and are considered “systems of record” for customer data.
The other type of CDP, a “System of Engagement,” maintains a realtime customer profile, and is mostly concerned with making sure the right customer gets the right message, offer, or action in realtime. These are systems more akin to journey management or Realtime Interaction Management (RTIM) systems. They are great for creating personalized engagement at scale, but not really where an enterprise would actively manage their first-party data.
The third category imagines an “Enterprise” strength offering that is the next of both worlds. Big enterprise software companies are working on building these today–systems where pure data management capabilities are structurally intertwined with the systems that deliver engagement (marketing, commerce, service, et al). In a nutshell, what if you had an infrastructure where a persistent customer profile and the ability to activate it was the foundation for all of a company’s systems? That would solve the problem of siloed data, provision an incredibly rich customer profile, make ML and AI smarter, and help drive experiences at scale across every touchpoint.
As the “enterprise” vision of the CDP becomes more of a reality it means that the way we think of “CDP” today is more of a framework for thinking about “data management” writ large, it will render many of the “CDPs” in the market today as point solutions, and it will transform the way we connect the backend systems (broadly, “ERP”) to the front end (even more broadly, “CRM”).
The New Era
In this new era, we will see identity redefined as consumer-driven through preference management, and more intelligent ways of connecting user data with the systems that need it. We will finally start to see the supply chain management and customer experience coming together – threaded through data and identity infrastructure, and we will start to see the emergence of what we can think of as true “enterprise industrial-strength CDP” as businesses seek to connect the way they operate their enterprise with how customers interact with it. It’s the beginning of an exciting new era.
[A version of this post appeared here on the Future of Customer Engagement]
The Devil's Work 