Data: First Harvard Business Review, Then the ANA Trust Consortium

By Bill Duggan

Limeart/Shutterstock.com

At the end of last month the Harvard Business Review published the article “Buying Consumer Data? Tread Carefully.” The authors tested the accuracy of popular audience segments that a range of data brokers offer. This passage particularly stood out to me:

• Across all our tests, we found that the consumer information sold by data brokers varies greatly in quality. A lot of it is similar to or even worse than what you’d get if you used random chance to create a target list. Demographic data was particularly disappointing. For example, the average accuracy of gender segments classifying males was only 42.5 percent — which is lower than the 50 percent natural chance of identifying men.

Yikes! That’s scary. It seems that great minds think alike as the ANA has just released the white paper “Data Sources for Media: A Buyer’s Guide.” A tremendous amount of money is spent on data for media buying decisions, specifically third-party audience data. A new ecosystem has been created of companies selling data to use in media buying, promising the ability to deliver a more on-target message more efficiently. Advertisers can easily buy and deploy data allowing them to target consumers based on a wide range of criteria, ranging from simple demographic attributes all the way to detailed purchase intent data.

Third-party data can come from many different sources, and multiple methodologies can be used in its collection, structuring, and marketing. This new data ecosystem is deeply complex, and often opaque. This often makes it difficult for advertisers to understand exactly what they’re buying, leaving advertisers at a disadvantage and at risk of purchasing data that’s unsuitable for their purpose. Without transparency, a solid definition of data quality, and tools to evaluate the data quality, advertisers are at a significant disadvantage in navigating this new ecosystem. To address this problem, advertisers need to evaluate the quality of a dataset before purchasing it. This paper recommends five criteria to focus on when evaluating data:

1. Data Accuracy: Does the data actually mean what I think it does, e.g., are visitors to an auto website actually more likely to buy cars?
2. Data Precision: Are the data collection and modeling procedures sufficiently precise to avoid a large number of false positives, e.g., does the vendor use a lookalike model that assigns people to the audience who shouldn’t actually be included?
3. Data Recency:How regularly is the data refreshed? When was it last refreshed?
4. Data Coverage: Does the dataset cover enough of my intended campaign audience to provide necessary scale for my client’s campaign?
5. Data Deployability: Can I use the data with my chosen tech partners?

A checklist is also provided for advertisers to use when considering a new data partner.

This is an initiative from the Trust Consortium, which was launched by the ANA in 2019 in partnership with Reed Smith, the ANA’s outside legal counsel. The Consortium’s mission is to restore trust in the marketing ecosystem through transparency, integrity, and growth.

The ANA Trust Consortium is pleased to be in the good company of the Harvard Business Review on the very important, complex and opaque topic of buying data.