A true Marketing System of Record, known as the “golden record”, is the most authoritative data source on a person or device and delivers as close to a complete, 360-degree view as possible for marketing purposes.
It’s “as close as possible” because there are some regulatory constraints. Certain pieces of personally identifiable information (PII) are restricted by privacy laws, and HIPAA’s Personal Health Information (PHI) privacy rules (and penalties for noncompliance) that are in force regarding medical records are the most stringent. It doesn’t mean that the system can’t hold it, but certain pieces of data must be physically separated, tokenized or encrypted until they must come together for an approved reason.
I’ve noticed that there is some confusion in the market caused by “almost-but-not-quite” data profiles that are passed off as a true System of Record. Specifically, there are data management platform (DMP) companies that claim they can produce a complete 360-degree profile. However, DMPs have shortcomings on both the amount of data they can store and limitations on the type of data they can use, making it impossible for a DMP to ever provide a true System of Record to their users.
While their platform can accept first-, second-, and third-party data, the fact that they cannot store the full customer data set means their profiles will not provide a complete view, and therefore are not able to offer the insights that a System of Record can.
For most DMP and digital-onboarding solutions, it is largely a one-way trip: You push keys and attributes in, but really can’t directly manage the data like a System of Record must. DMPs generally assume that the data it processes is “clean and ready,” which would mean that there is an underlying System of Record that has done the work for it already. However, people (and their data) don’t remain in a static state, and external data not uniquely related to an individual (e.g. weather, inventory, ERP work-in-progress, etc.) are commonly needed to execute sophisticated marketing campaigns.
True System of Record
Before I go further, let’s review a true System of Record in more detail. On the surface, what constitutes a System of Record is pretty straightforward. To create a System of Record, you must take all of the information available about a person or a device, including all first-, second-, and third-party data. Once all available information on that person or device is gathered, it must be cleansed and resolved of conflicts through rigorous data quality and master data management processes. Matching then must be completed – the process of comparing incoming “potentially new” data to those that already exist in the system in order to try to “match” them. The matching is what creates new records or adds attributes to already-existing records in the system.
On the one hand, the amount of time available to make proper engagement decisions for marketing messages is being compressed, while on the other the number of sources needed to make these intelligent engagement decisions is increasing. Many companies are now fully engaged in creating customer data platforms that require near real time inputs of digital user information, other business unit information like inventory availability at shipping points to assure offer fulfillment within a promised time, IOT device information, and more. So even weather data in an area (predicting not only temperature, precipitation and like data, but also disasters) can be part of the golden record for a person even though it changes regularly.
DMPs and the Look-Alikes
DMPs, loyalty solutions and other similar point solutions can provide huge value to the marketing team; but providing a traceable, visible single record profile isn’t among the value attributes. Other point solutions fall short of creating a true System of Record as well, and many use acquisition-focused techniques that onboard certain attributes and then perform a look-alike search.
This is a misunderstanding of what a System of Record needs to be and what a Look-Alike match is designed to perform. As its name implies, a Look-Alike search begins with a small sample of customers with the desired attributes (location, income, known products he or she is actively shopping for) and uses the few data points to search (mostly third-party) data for others that have similar profiles. Kate Kay, writing in Ad Age, had a spot-on explanation:
“Look-Alike models are used to build larger audiences from smaller audience segments to create reach for advertisers. In theory, they reflect similar characteristics to a benchmark set of characteristics the original audience segment represents, such as in-market kitchen-appliance shoppers.” Kay goes on to identify the Look-Alike model’s most significant shortcoming: imprecision.
“Generally, the higher the marketer’s tolerance for loosening the model, the larger the segment can become,” she writes. When you come right down to it, Look-Alike models are fine for their purpose, that not fine to be used as a System of Record. They’re a checklist of attributes, but lack the depth of information that can provide a very well-rounded understanding of the target.
Data is Constantly Changing the System of Record
The final essential aspect of a System of Record to remember is that it has to respond to the dynamism of its information. Purchases rise dramatically over the holidays. Bank balances can vary day-to-day (or even hour-to-hour), as can credit card debt. New pieces of data might be added every day, or at any moment, and the System of Record requires constant updating. While the volume and variety of data sources for building a System of Record are well known and well understood, what has changed the most is the volume of data. Digital data collection (from web clickstreams, decision optimization, Internet of Things devices and applications, and more) has upset the balance and timing of processing. Digital data has a much shorter half-life than other types of data lineage sources.
Common use cases that increase the complexity of the profile and the maintenance of the System of Record that are being used now involve a mix of:
- IOT devices always add the possibility of great insight combined with increased complexity
- sentiment analysis (from social media),
- organic site behavior,
- real-time weather – triggering offers to purchase more of a certain product because the weather will cause greater consumption – water filters, coffee, makeup moisturizer, generators, and
- triggered contacts based on user searches on particular problems on your site
Given the growing complexity of data and the proliferation of communication channels, organizations are under greater pressure to have full access to all of their customer data when they need it and how they need it. Understanding what a true Marketing System of Record is brings marketers even closer to achieving that “Golden Record”, and it’s only a matter of time before intelligent customer engagement programs are the norm, not the exception.