Third-party data suppliers sprouted up a few years ago, looking to ride the wave of big data and offer wider audiences to advertisers. But as advertisers get wise to the fact that they're sitting atop valuable data of their own, third-party data companies have had to pivot, given their data's limited effectiveness and high cost.
Many now package first-party data or employ other companies' data in more helpful ways, but these offerings are still largely ineffective for the amount advertisers pay. If it is to maintain any value to marketers, the third-party data industry will need to change its economic model yet again.
Of course, not all third-party data are ineffective or bad. But it's so difficult to get scale out of good data that third-party data companies are forced to use proxies that dilute the value. A consumer who visits MotorTrend.com, for example, is often instantly "cookied" as intending to buy a car. Several companies tout 45 million such consumers in their cookie pool, which is hard to believe when only 15 million cars are sold in the United States. The attributes are forced, resulting in less effective data.
For any one advertiser, between one million and three million qualified customers might be in a cookie pool. Even if there were 10 million, and the data were sold at a $1 CPM, the data seller would make only $10,000. You can't build a business on that.
Any insight into this dilution is rarely passed on to the advertiser or agency buying the media. Data exchanges inflate the price of this suspect data. Any advertiser executing a cost-per-action campaign that goes this route is automatically behind the eight ball, even before launching.
But the data industry is getting better, with companies like Amazon and MasterCard sharing their customer and intent data. If a consumer looks at an item on Amazon, that's a very valuable insight. The consumer packaged-goods industry is bringing on customer-loyalty data to find its consumers online and track offline impact via various studies. This is valuable, but again, it's difficult to buy these impressions at scale.
Often unmentioned is that any marketer buying data is not the only one bidding for an impression. They bid not only against competitors in their segment, but also other advertisers looking for similar consumers. The most in-demand consumers are typically wealthier and web-savvy, using the web to conduct most of their research and make several purchases.
This competition makes it even more challenging for advertisers to get superior value out of third-party data. To survive, third-party data will need to link pricing to campaign performance results, rather than to a flat CPM model. Instead of charging for audience packages up front, data brokers should share revenue on the incremental lift they drive.
Brands would likely fall over if they saw their true CPM for reaching qualified customers; tiered deals where the data broker sees a percentage of the return on investment would be better for all parties. Brands could still leverage the benefits of third-party data, without having to spend exorbitant sums for modest gains.
Additionally, real-time bidding gives advertisers access to some important data for free — geographic information, retargeting cookies and URL domains are all passed to the advertiser before making a bid, all for free. Plus, the most valuable first-party data, the marketer's own, is also free.
The bottom line is that marketers are always better off looking to activate their own first-party data. Third-party data carries some value, but it's crucial to constantly test. Advertisers can control the data market by keeping close tabs on what kinds of data and insights actually drive performance. If third-party data continue to drive small lift, then advertisers need to push for new models that reflect this, ensuring they get only what they pay for.