How Data Variety Makes Programmatic More Effective: Part 3 of the “Programmatic Matters” Series
Advertising without data is like cooking without using the proper ingredients. Sure, the finished product may be edible, but it won’t necessarily taste good. Just like a delectable home-cooked meal, an effective advertising strategy requires the use of a diverse assortment of elements to make it a success.
When it comes to programmatic marketing, data is the essential ingredient that makes it so effective. It helps marketers hone in on their ideal customers, no matter where they are across the Web. The more high-quality data a business can harness, the better their results will be. Sure, you can still deliver an ad programmatically, but without baking in data-driven insights, the results would be pretty rotten.
According to eMarketer, digital marketers have a “love affair with data” and are planning to increase spending and usage in data-driven marketing strategies. But like the ingredients in your kitchen cabinet, data comes in many flavors, and marketers need to understand how and when to leverage different varieties of data to their advantage. Despite what my three year-old son may think, just because you love chocolate chips doesn’t mean you should use it in every meal. The same goes for data. You have to mix it up.
In Part 2 of our "Programmatic Matters" series, we briefly discussed the role data plays in making programmatic advertising campaigns smarter and more valuable. Before programmatic enabled the use of real-time data to inform decisions, advertising was essentially executed on a whim and a prayer. Today, data helps marketers understand their audience and plan messages around the most receptive customers. In the programmatic ecosystem, this information is accessible and managed in data management platforms (DMPs) and can be used to fine-tune the digital ad strategy in real time.
But for all the intelligence available today, data is still a very broad term, and one can use many varieties of data to make programmatic work. Here is a quick rundown of the sources and types of data in the modern marketer's cupboard:
Data Types Based on Collection, Methodology & Source
This is fundamentally information marketers and publishers create, collect and own. It is comprised of data collected through website behaviors, search queries and purchase activity (e.g., you went to Amazon and logged in). It’s also information users explicitly provide, such as their name, address, demographic information and any other personally identifiable information. This tends to be the most accurate compared to other forms of data – and the cheapest, because there is no cost in obtaining it.
Of the three sources of data, this tends to be the least talked-about in the industry, though it is still important. Second-party data is, for example, when Amazon uses your browsing activity to infer something about you. You did not know you were giving it to Amazon, but by clicking on some products on the site, Amazon collected that data and inferred you had certain types of interests. You do not know exactly what interest Amazon is going to infer, but you knew you were clicking on products on Amazon pages. The owner of this type of data is often makes it available for others to use.
The most common source of data and a constant in the programmatic world, third-party data is just that: anonymized data that’s been bought or collected from a third party such as a data broker. Again, let’s use Amazon as the example. Third-party data is collected by an entity the user is not aware of. On Amazon pages, there is a tag from another data company. As a user, you cannot openly see the tag but when you click through the site, this data company collects information about you and builds it into its audience data assets. You did not actively decide to interact with this data provider. As far as you are concerned, you were interacting with Amazon. That’s third-party data in a nutshell. No matter where you source your data, it comes in different forms that range from factual to extrapolative in nature. It can often be broken down into three distinct types:
Demographic: Often collected from a combination of online and offline registration sources, demographic data identifies users by key traits including age and gender.
Interest: Typically based on user-generated online actions, this data captures a consumer’s affinity for a particular topic or hobby.
Intent: This type of data is based on actions like web behavior and helps paint a picture of customers who may be in the market to make a purchase decision.
Probabilistic vs. Deterministic
Knowing what type of data to rely on can often be complicated. What’s accurate? What can you trust? Many marketers face this challenge, especially with a "probabilistic" approach to data generation. They use proxy models to define targetable prospects that they then feed into their marketing engines. Unfortunately, probabilistic can be problematic if proxies are based on incorrect assumptions. Conversely, there’s "deterministic" data like demographics – data gathered from verified sources and vetted for quality. It’s not derived using models and assumptions; it is real user information sourced from real people, collected, aggregated, edited and verified.
A good example of deterministic data is firmographic-level data. This particularly helps B2B marketers who need to reach a specific prospect within an organization. Firmographics are business-related criteria that can help you narrow down your audience to focus in on those organizations most likely to represent potential clients or customers. The data sets can include everything from a company’s geography and employee size to the organization’s annual revenue and total assets. These insights can help you understand where to spend time and money prospecting and upselling. And best of all, this data can be layered on your media buy in a programmatic environment.
For example, Dun & Bradstreet leverages proprietary matching methods and brings this data online in a secure and anonymous fashion that maintains the highest standards in PII protection and user privacy. With an agnostic approach to syndication and distribution, this data is available for use in programmatic ad campaigns.
All of the above types of data can enrich your programmatic advertising strategy, pre- and post-campaign. Just like preparing that perfect meal, it takes knowing the right ingredients to use to get the recipe just right. In Part 4 of the "Programmatic Matters" series, we will talk about putting this data in the various types of ad targeting you can do programmatically.