Ideas
January 10, 2022

Stop Paying for Expensive Data Appends

Agencies have been acquiring their way to data competency, yet still overpaying for data appends with low utility. Head of Data Science and Analytics Glenn Waine unpacks how brands can better allocate budgets toward more meaningful data.

GALE

A Business Agency

It’s no secret that consumer data is required for effective modern marketing. However, the value should come from services sitting on top of the data, not the data appends itself. Agencies have been increasingly relying on data as part of their core competency, touting their ability to enrich client data with numerous (sometimes hundreds) of attributes they can match with a client’s first-party data. This, of course, will help compete against the walled gardens of Facebook and Google against a backdrop of increasing privacy controls (GDPR, cookies, iOS updates, etc.).

In the last few years alone: WPP launched Choreograph (a global data company), Dentsu acquired Merkle (a data-driven performance agency), IPG acquired Acxiom’s data-marketing division, Publicis acquired Epsilon, and Omni built its own system called Omni. These moves amount to several billions of dollars that deepen the data capability of these ad agencies. Yet, too often, marketers are buying expensive data appends with low utility. For example, a database of 10M consumers that could be $0.9M per year (match rate of 50%, updated bi-monthly, with a $30 cost per 1,000 matched). There are better uses for this spend because of these factors:

• The data attributes are not actionable or useful

• They are slow to update

• They are not reliably accurate

• There are a few core sources that all agencies purchase from (e.g., Credit Unions, syndicated databases, etc.)

Instead, this $900K could be reallocated to building a system to capture data from your customers through quarterly surveys, quizzes as part of a loyalty program, and more, that are specifically helpful to your business. Brands should seek partners that build their assets with meaningful data rather than buying expensive, unspecific and unactionable attributes from vendors.