We live in a time when data has emerged as one of the most valuable resources available to businesses. Companies with good data are able to fuel their machine learning models, make critical predictions, and ultimately deliver greater value to customers. In the process, they can also optimize their operations, lower costs, and differentiate their business.
But to realize these benefits and compete with the likes of Amazon, Facebook, and Google, companies need access to huge datasets. And while there’s certainly no shortage of data available, the problem most companies face is that they only have access to data from their own customers. As a result, they often turn to third-party data to fill in the gaps. In 2018, for example, companies spent an estimated $19 billion on third-party data in the United States alone. Unfortunately, as we’ll see, using third-party data isn’t always a great approach.
The truth about third-party data
When you purchase third-party data, you’re buying data from providers with no direct connection to your customers. Instead, data providers simply collect demographic data from a variety of sources, such as websites and social media channels, among others, and aggregate it into the types of audience profiles they think you’re looking for. And while there’s nothing wrong with purchasing third-party data in theory, in practice doing so can be problematic for a number of reasons. These include:
Quality issues. The data you can buy often isn’t of very high quality and, in some cases, can be flat-out wrong. Part of the problem is that data providers aren’t always as concerned with the accuracy of the data they’re collecting, opting to instead focus on building massive datasets that they can sell to businesses. As a result, you can easily wind up with flawed data in your machine learning models that undermines the accuracy of your predictions.
Security and privacy concerns. Anytime you use third-party data, you could be exposing your business to serious security and privacy risks that have the potential to compromise your buyers’ trust. Such was the case earlier this year when Twitter announced that it would stop using third-party data in its ad-buying system in response to privacy concerns. Meanwhile Facebook famously faced a serious backlash in 2018 in the wake of the Cambridge Analytica scandal. Given that 91 percent of consumers agree that companies shouldn’t be accessing their data without their consent, maintaining privacy and security should be a very real concern for every business.
Regulatory considerations. A growing number of laws, including the General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA), dictate exactly how you can acquire and use data. While not yet widespread, such laws one day likely will be, adding significant complexities to using third-party data while remaining compliant.
Practically speaking, while using third-party data has long been viewed as a necessary way to get ahead, doing so is fraught with considerable challenges.
A better way forward
If you want your business to compete with the tech giants of the world, you’re going to need data. The trick, of course, is finding a more responsible way to source that data so that you know that it’s of the highest quality without raising security or privacy concerns, or putting you at risk of running afoul of regulatory requirements.
That’s where the Trusted Signals Exchange can help. It allows companies to enrich their data with cross-industry intelligence without relying on personally identifiable information. In that way, it provides an effective alternative to third-party data, while addressing pressing regulatory, ethical, and data privacy concerns. It uses machine learning techniques like transfer learning to generalize consumer preferences in one area and apply them to another, so that businesses more accurately predict their customers’ propensity to buy, and recommends how the customer experience could be optimized to achieve a desired outcome.
At a time when data-driven insights are critical to success, and the limits of third-party data are becoming increasingly obvious, our approach leads to better performance and better business outcomes.