How many electronic devices have you used today? Perhaps you checked your phone after waking up, worked from your desktop or laptop, and interacted with a chatbot on a website you visited during a break—all in the space of one morning. To provide a unified brand experience—which customers now expect—marketing technologies must identify and synchronize all of these devices. At the same time, there’s a good chance that as you read the news you learned about another ethical or security concern about organizations’ use of customer data. How can the seemingly contradictory wishes or personalization and privacy be reconciled?
In Chapter 2, you learned about direct marketing as an alternative to mass media, and the progression of tools from database marketing to customer relationship management and customer data management platforms. Here you will learn about the four essential facets of the CDP, and how the CDP emerged to address new developments in technology, privacy, and regulation. Figure 3-1 shows this history and progression of marketing technologies.
As Figure 3-1 suggests, it’s best not to think of each new technology as completely supplanting the other but instead as being additional slices of an overall marketing stack.
A CDP doesn’t necessarily replace existing marketing technologies like a CRM or DMP; instead, it’s used to integrate these and other data sources in creating a unified source of customer data.
While each has its purpose, the CDP is the only marketing technology that is designed around an identifiable customer and is meant to holistically enhance the customer experience. Given changes in technology, privacy, and customer behavior, these attributes make the CDP an essential slice of the marketing stack.
Sending one-time campaigns to large segments of customers via email or other static channels was a breakthrough in the 1990s. But marketers in the 2020s are encountering whole new sets of challenges and opportunities to deliver a customer-centric approach, and older marketing technologies just aren’t up to the task. These forces are described in the following sections.
While the term “Big Data” to describe the explosion in data reached its apogee in the mid-2010s (research firm Gartner famously declared it passed the top of its famous “Hype Cycle” in 2014), there’s no question that something dramatic happened in the generation and collection of data in the last 20 years. Customer data available to marketers was no exception.
The World Economic Forum estimates that by 2025, 463 exabytes of data will be created each day globally. From email to web searches to social media posts, much of this content is created by individual consumers. Compare this to the limited engagement channels that customers had a few decades ago to communicate with brands.
To accompany this surge in volume of data is an increase in variety, not just in media but in devices. At the turn of the 20th century, just over one in two US households had one desktop computer. By 2014, more than 90% of US households had at least three connected devices, and not just desktop computers—from laptops to mobile phones to tablets, consumers are consuming and generating more data from more sources.
Finally, today’s consumer-generated data moves faster than ever: consider the always-on nature of social media versus engaging with a brand via telephone or mail. As data moves faster, it often also loses value or relevance faster. This is especially the case for connected devices such as the Internet of Things (IoT), which generate data in real time that is meant to be acted on just as quickly. For example, the best time for an individual to know about when their order is ready for pickup or when to follow up with a potential lead is immediately—not hours or days later.
Figure 3-2 shows the growing sources of data that marketers must integrate to deliver a customer-centric approach.
This explosion of data is only set to continue: networking and telecommunications giant Cisco predicts there will be 29.3 billion networked devices by 2023, up from 18.4 billion in 2018. What’s more, information services provider IHS Markit predicts there will be 125 connected IoT devices by 2030.
Earlier marketing technologies were not designed to handle the size and scope of today’s data. Remember that in Chapter 2 you learned that relational databases and data warehouses are meant to handle well-structured and predefined data, such as transactional or demographic information. These architectures struggle to accommodate diverse and unstructured data such as tweets or images. Databases and data warehouses also tend to extract and load this data on a predefined basis, such as daily or weekly. That means they often fail to capitalize on the velocity of data.
As we discussed in Chapter 2, earlier direct marketing approaches focused largely on email, with other channels like social media and brick-and-mortar offering separate, unintegrated touchpoints. This is known as a multichannel marketing approach. Customers may have received some personalization from them, but each channel was largely independent in its campaigns. For example, a customer may be given a website address on a shopping bag in attempt at cross-channel promotion, while physical stores and ecommerce remained two disconnected touchpoints for the customer.
By contrast, an omnichannel strategy is designed to offer customers a unified experience across all channels. For example, in an omnichannel approach a customer is able to shop for products online, then pick them up at the store. Figure 3-3 shows the difference between multichannel and omnichannel marketing.
Earlier marketing technologies aren’t designed to accommodate for this growth in number and variety of devices and channels. Database marketing primarily focused on offering customers a fixed offer via email (for example, informing of a promotion valid either online or in stores, often not both).
A DMP may engage in multichannel advertising. But that is only one part of the customer experience, and not truly omnichannel to begin with; there’s a difference between engaging with a social media account about a recent order that associates at a local store then follow up on. Moreover, these advertisements are not truly personalized but delivered based on anonymous and pseudonymous data.
A CRM is primarily meant to facilitate a sales team’s engagement with customers and potential customers. It is not designed to unify how a customer interacts with a brand across its physical and digital imprints.
There are many risks to not adopting a multichannel approach in today’s marketing landscape. If marketing teams and advertising budgets, for example, are siloed in separate customer engagements, there will be duplicated efforts, campaign inefficiency, and wasted resources. Customers will get frustrated, as they feel they are given “split personalities” across different channels of the same organization.
An omnichannel approach also facilitates marketing attribution. Consumers today go through a near-infinite variety of channels and touchpoints before making a purchase. Without an omnichannel approach, it’s nearly impossible to integrate and track how these various touchpoints lead to a conversion.
More broadly, the explosion in devices, data, and channels signals a new way in which consumers interact with brands. With new technology and sources of information, consumers choose when to engage with brands and why. A fixed customer journey doesn’t conform to the new role of brands as a source of on-demand data and information to its customers.
Rather than focusing on converting and communicating with top-of-funnel leads, as CRMs and DMPs tend to do, the CMP delivers the right services to customers when and where they ask for it. For example, a CMP can augment traditional CRMs and DMPs with chat-based customer service and omnichannel support, such as the ability to return to a store an item that was purchased online. Through these technologies, organizations can create personalized touchpoints based on a costumer’s individual circumstances and preferences. This has the benefit of keeping the customer informed and engaged throughout their journey rather than focused on converting leads at the top of funnel through advertising.
Of course, to provide individualized experiences, organizations must draw from individualized data. In Chapter 2, you learned about the origins of direct response marketing and the ways in which organizations have historically gathered customer-level data. A dominant tool for the last two decades has been the browser cookie. This technology has met its limitations in offering a truly customer-centric experience for a few reasons. First, cookies aren’t purely personalized but rather pseudonymous data sources. They also have limited value past the acquisition phase of the customer journey, as they are primarily used to serve targeted advertisements to segments of users.
Moreover, consumers have reached a tipping point with this marketing strategy: marketing consultancy Cheetah Digital found that 41% of US consumers delete cookies regularly and 30% have installed an ad blocker. According to management consultancy Deloitte, with the average website placing 12.4 cookies on a browser and reporting data to 3.9 third-party domains as of January 2020, it’s more than understandable that consumers are wary of this tactic. Giving up privacy and personal information simply to be met with the same series of advertisements as they browse the web is no longer an adequate or permissible exchange.
Moreover, changes to browsers and the web itself signal a move away from the cookie’s dominance. For example, browsers Firefox and Safari now block third-party cookies by default. Google’s Chrome browser is set to do the same by 2023. In their post explaining the decision, Google wrote: “Users are demanding greater privacy—including transparency, choice and control over how their data is used—and it’s clear the web ecosystem needs to evolve to meet these increasing demands.”
The web seems to be bracing for this shift: global cookie tracker webcookies.org now finds that, as of April 2021, the average website now sets 5 cookies on average and reports to 1.1 third-party domains compared to 12.4 and 3.9 in January 2020, respectively.1
With the eventual demise of cookies and their ultimate failure to provide a holistic customer experience, marketers are examining how best to gather, analyze, and act on personalized data. Because it’s gathered at the individual level, this type of data inherently includes personally identifiable information (PII): information that can be linked to a specific individual. Figure 3-4 illustrates possible sources of PII; these can range from demographics to location to medical information and more.
PII presents a number of logistical and ethical challenges for organizations. First, as we’ve discussed, data comes from an ever-increasing variety of sources. The same individual could easily use a different name or email address at one source or another. It takes considerable computing power to construct a clean profile for each individual across these disparate, dirty sources.
Perhaps even more critical is the need to keep this information safe and transparent. Consumers have revealed a preference for personalized brand experiences and understand what is needed to deliver it: the consulting firm Accenture found that 83% of retail consumers are happy to passively share personal data if it allows brands to create a better experience for them. At the same time, these consumers have experienced enough selling and use of their personal data across organizations to be wary of providing the PII that enables this personalized experience, a phenomenon that researchers refer to as the “privacy paradox.”
Organizations can resolve this paradox through transparency, security, and results. For example, Salesforce has found that 92% of customers are more likely to trust businesses with their data when they are given control over what is collected about them. Moreover, if customers can’t see that their data is actually being used to provide a personalized experience, they may question how or why it’s being collected in the first place.
Cookies or not, regulators have caught onto the continued demand for consumers to control how their data is collected and used. For example, the European Union (EU)’s General Data Protection Regulation (GDPR) establishes guidelines for the collection and processing of data about individuals who reside in the EU. Under the GDPR, website visitors must be notified about what data the site collects and explicitly consent to it. The California Consumer Privacy Act (CCPA) establishes similar guidelines about the PII gathered from citizens of the state of California.
Earlier technologies like CRMs and DMPs still have their place in the marketing stack. But from an explosion in the variety, volume, and velocity of data to the demand for personalized, omnichannel experiences to the end of browser cookies and the need for transparently collected and reliably cleaned PII, many organizations need a new system for building customer-centric marketing.
Marketing technology consultant David Raab developed the concept of a customer data platform (CDP) over a series of blog posts in 2013 and later went on to found the vendor-neutral CDP Institute, also developing the first definition of the term. By 2020, the CDP Institute identified 133 CDP vendors and estimated annual industry revenue at $1.3 billion. While it’s clear that it’s a nascent, growing marketing technology segment, there’s still an emerging consensus about what constitutes a CDP. After all, this chapter mentioned earlier that a CDP should augment rather than replace older technologies like a CRM.
What this does not mean is that a CRM and CDP are interchangeable, and marketers should be able to delineate the two. This hasn’t always been clear: for example, Salesforce executives Martin Kihn and Christopher O’Hara memorably recall in their 2020 book Customer Data Platforms that in one survey, 62% of marketing technology professionals reported using Salesforce’s CDP before the company had one on the market.2
As a young technology, best practices in adoption and implementation of a CDP are forming. You will learn more about them in Chapter 4. For now, let’s focus on what is essential from a CDP to meet the requirements for modern, customer-centric marketing, as discussed earlier in the chapter. The research and advisory firm Gartner has identified four such facets that a technology must feature to function as a CDP: data collection, profile unification, segmentation, and activation.
Earlier sections of this chapter discussed the wide variety of devices and channels with which customers and brands engage. CDPs must not only be able to capture this data, but process and serve it quickly to account for its velocity. Chapter 4 will touch on some of the technologies needed to accommodate for this, such as Hadoop clusters and application programming interfaces (APIs).
A CDP must also integrate this disparate data into a “golden record” or “single source of truth” for each customer, based on their PII. The basic task of profile unification is illustrated in Figure 3-5. Notice that in this example, two different email addresses are provided. A CDP must be able to integrate this divergent, even possibly conflicting, customer data.
Profile unification is logistically and computationally demanding. A CDP which unifies this data based on exact matches alone risks discarding huge numbers of customer profiles. Many CDPs use sophisticated statistical models to match disparate pieces of PII based on probabilistic evidence.
Customer segmentation is nothing new: in Chapter 2 you learned how database marketers wrote SQL queries to identify segments based on response to previous emails. The CDP serves to unify and simplify how marketers perform this segmentation. With all customer data integrated into a customer-level record, segments can be defined using information ranging from media engagement to purchase data to call center interaction data.
Finally, the CDP should facilitate driving customers to action through brand interaction. Done right, this goes far beyond the targeted advertisements of DMPs: customers should be driven to action by an omnichannel, cross-journey experience. For example, CDP activation looks like ecommerce showing users different offerings based on their last in-store or online purchases. Or it could be a call center and social media representatives using a particular script based on the identified needs of the customer’s segment.
The CDP Institute reported that the number of CDP vendors increased by 35% in 2020. How can marketers navigate this landscape? Gartner’s four facets offer one framework for evaluating whether a technology meets the requirements to serve as a CDP. But still, these features merely constitute the means of what a CDP does. Its ultimate end is to break down traditional marketing silos and establish customer-centric marketing.
Consumers today are looking for contextually relevant and highly personalized engagement across all touchpoints and channels. They are willing to have trusted brands work with their data in exchange for these experiences, so long as it is done transparently. Ultimately, the CDP offers a marked shift from “interruption marketing to “permission marketing.”
The technological benefits of a CDP are one thing. But how do you actually implement this tool in your organization so that it’s used successfully? To go further, how do you measure that success? In Chapter 4, you will learn about the people, processes, and technologies required to adopt a CDP and how to evaluate it. Then in Chapter 5, you will assess your organization’s marketing maturity and the marketing technology right for you.