Data quality is among the most common pain points
associated with marketing initiatives. For teams engaged in email
marketing, programmatic marketing, or other big data-driven projects,
quality issues can significantly reduce results. If your organization’s
efforts to produce targeted, real-time messaging are generating poor
lift, it could be important to look towards your third-party data vendor
as a potential source of the problem.
In best case scenarios, third-party data can allow marketing teams to develop 360-degree understanding of their target customers. However, directing dollars towards the wrong third-party vendor can actually damage efforts to programmatically generate advertising messages. If your vendor’s insights are out-of-date, generated through poor data logic or clustering technique or inaccurate, your results could be worse than if you were solely reliant on first-party insights in your data management platform (DMP). In this blog, you’ll learn the differences between data types, and how the wrong vendor can lead your team astray.
Understanding the Classes of Big Data
While sources and volume can vary significantly, there are a few terms commonly used to describe the origin of data that may be applied to a big data-driven marketing campaign. Understanding the following classifications can allow marketers to understand sources of risk in their marketing campaigns, and make the right choices about data acquisition at a large scale.
1st Party Data: These insights are generated by your company’s web, mobile, and transactional records. Typically, these insights are the most accurate, and are housed in a data management platform (DMP), which is typically integrated with a CRM.
3rd Party Data: These insights are obtained through an external data provider. The data is generally anonymized, and may be matched with your contacts in a data management platform. Vendor sources can vary significantly, but purchasing from a large-scale vendor can result in insights that are out-of-date and suffer from quality issues.
2nd-Party Data: These insights are among the most rare. 2nd-party data could originate from long-term data sharing agreements between organizations to continually combine and match profiles.
many big data campaigns, the single biggest source of risk is 3rd-party
data. When completing audience profiles with old or inaccurate
insights, your audience profiles could be significantly diluted. Sources
of risk in 3rd-party data quality can originate from the following
1. Sourcing Methods
Third-party data vendors often have “mountains of information” available, according to Dunn & Bradstreet (D&B). However, their sourcing methods can be a bit of a mystery, even to some external representatives of the organization.
In one case study, a 3rd-party data vendors classification of “new parents” proved 10-20% inaccurate, per D&B, because it was based on individuals who’d recently purchased a certain magazine subscription. In other cases, vendor’s sourcing is based solely on online browsing cookies.
Regardless, your marketing results could be questionable if you’re not able to quickly establish each of the following with a prospective data vendor:
● Where does the data come from?
● Does the data represent online and offline behaviors?
● Do you rely on multiple data points to build audience groups?
2. Quality Assurance Methods
Quality assurance represents a major source of effort for data science teams. While purchasing third-party insights that are cleansed can provide convenience for marketing teams, your vendor’s quality standards need to be impeccable to yield gains.
Understanding your vendor’s approach to data verification, elimination of old data assets, and comparison is crucial. The best indication of data quality is results. Proof of recent conversions is the most objective way to measure third-party data assets.
3. Refreshing Methods
Generally, most data vendors “refresh” their data assets on a periodic basis, by pulling new insights into their data management platform. For vendors that source from a variety of sources, these “refreshes” may occur very occasionally, such as every several months.
In a world where consumers have access to immediate purchases via mobile devices, recent data is crucial. Insights that accurately reflected your audience’s behavior three months ago are not accurate today. Unless your vendor’s data is updated in real-time, it’s out of data.