skip to main content

Reimagining the Third-Party Data Ecosystem

Acxiom Last Updated June 9th, 2020
Reimagining the Third-Party Data Ecosystem

In the last 10 years, data has reshaped the media ecosystem. Sadly, that ecosystem is still broken: most marketing today is not very good. Consumers throw away the mail they receive, ignore banner ads, skip pre-rolls on videos, and delete emails from brands.

While most ads are spam, there are a handful of counterexamples where ads actually work. As a consumer, my favorite marketing comes from Amazon; the recommended products feature helps me pick out good books to read. When the recommendations are good, it’s a win-win – Amazon and the book supplier get more revenue, and I get a better book.

A few years from now, all ads will look much more like Amazon’s, and provide actual value to consumers by being extremely relevant. While the biggest factor driving the change is the explosion of first-party data, the third-party data ecosystem is simultaneously being disrupted, shaped by the following five trends:

Trend #1: Moving From Segment-Based to Signal-Based

Today, data is predominantly used in segments; media planners buy audiences (e.g., 35-45-year-old males in Milwaukee).

While this level of targeting made sense in 2010, it is outdated now. Instead of humans manually selecting a handful of key attributes to target, a campaign should algorithmically utilize millions of “signals” to determine how to best accomplish a marketing objective. Signals are raw bits of information, such as past purchase behavior, location data, or online browsing behavior. These are generally not very predictive in isolation, but are exceptionally predictive when aggregated.

In the signal-based targeting ecosystem, brands will say “I want to drive more purchases” and the power of millions of signals should be automatically brought to bear.

Trend #2: A World of Thousands of Data Providers

As the world transitions from segment to signal-oriented, more companies will be able to monetize their data and provide insightful bits of information to improve campaigns.

Between publishers with browsing behavior, apps with location and mobile behavior, retailers with in-store purchase data, data collected from the Internet of Things, TV viewership data, and brands’ own data, thousands of companies are collecting exhaust data. Alone, none of these segments are as valuable as a compiled “household income” segment is today, but together, these signals are much more predictive than a compiled segment.

In a segment-oriented world, companies that assemble the best and broadest segments provide the most valuable targeting. But in the signal-oriented world, the best campaigns will use as many rich signals as possible – leading to a data economy that attracts a long tail of small providers.

Trend #3: An Ecosystem of Data Curators

Aggregating millions of signals from thousands of data providers is a major technical challenge, but making sense of all the signals to optimize campaigns is even more challenging. This will be an area of massive opportunity and competition for martech in the next 10 years: DSPs will seek to incorporate signal information directly into bids, AI companies will claim algorithmic superiority, and new martech vendors will vie for recognition as the best option to handle the challenges of a signal-based world.

Trend #4: Omnichannel Consistency

Today, there are fundamentally different data ecosystems around direct mail, email, TV, programmatic display, and mobile. To create the most targeted experiences, brands will seek access to signals from consumer interactions that occur across channels and contexts, and demand omnichannel consistency of their campaigns. Single-channel data marketplaces and list brokers will be replaced with channel-agnostic exchanges.

Trend #5: The Rise of Permissions & Privacy Management

With thousands of data vendors creating millions of signals across dozens of channels, managing data permissions will take on exponential complexity. Brands seeking to monetize their data will have strict access restrictions, data providers will seek to control the pricing of their data, and everyone will be concerned about data fraud and theft. Managing notice, choice, and consumer transparency will be fundamental for the ecosystem. The data economy of the future will view permissions and privacy management as central design requirements, not an afterthought.

Combined, these five trends will disrupt the existing data economy, and have the potential to transform marketing from spam to relevant, curated content.

This article was first published in MarTech Advisor on July 17.