When you apply for a loan, you provide a bank with your personal information, but they pull your credit history from a third-party company such as Experian, Hyperion or TransUnion. If you want to go to another bank to get a better rate on your loan, the second bank will again pull your credit history from one of these independent firms. In other words, you can buy financial products from any bank you like and they will use the same data to evaluate your creditworthiness.
Now imagine if your credit history was held by one bank and you could only buy the products they offered. The data that enabled you to buy the best product would be trapped within one company – and this company may not offer the best products in the marketplace, but you are stuck with them. It would make no sense and consumers would cry foul. Yet this is exactly what is happening in digital advertising.
Several companies who were independent, agnostic holders of customer data for large brands – often known as data management platforms or DMPs – have decided to add audience buying technology to their offerings. The reverse is also common, as former audience buying platforms (or DSPs) add proprietary data management to their offerings, with the same net result. The motivations for these moves are understandable – many see a profit opportunity and they convince their clients that there are efficiencies to be reaped from managing audience data and media buying within the same system.
Yet this setup has inherent conflicts that may be putting the interests of the technology vendor before their brand client. In the worst cases, a handful of companies are refusing to integrate with other media buying firms, thus restricting their clients to whatever audience buying capabilities they have themselves, which in many cases are not the best in the industry.
While this is the most shocking behavior, there are also ongoing conflicts that brands must guard against. A DMP has a vested interest in proving that the audience data it is warehousing is valuable for media buying, otherwise a brand would question why it is storing all this data. When the DMP is also the vendor executing the media buy, they naturally want to show the brand how integral their data was in driving the performance of a campaign. However, the reality is that sometimes audience data helps achieve a brand's objectives – and sometimes it does not. An independent media buying firm does not care what targeting package is used – they just want to use what works best to meet a brand's objectives. This is not the case with DSPs who are trying to be DMPs.
We obviously bring an agenda to this as a pure-play video buying platform. But working with a variety of data companies taught us that our incentives are not always aligned, a fact that advertisers might be interested in.
It is worth noting that while most marketers believe in the promise of big data, many remain skeptical about data quality issues. They are not universally wrong. According to an internal analysis focusing on four different providers of gender-based audience segments, for instance, we found that companies disagree on gender 20.7% of the time -- meaning that for two in ten viewers, one data provider says a viewer is female, while another says it is male. Many marketers are discovering this themselves by using Nielsen's Online Campaign Ratings, which provide independent verification that a target audience was reached, and are learning that other proxies work better.
What is clear is that a buying platform's motives are simple: find what worked -- and where -- and learn from what did not so that more money flows through your system. A pure-play data company's motives are similarly sanguine: give marketers insights and they will pay for what they find useful. Whether merged companies can do the same, balancing their own motive to boost data fees with adding value to media buys, is ultimately up to marketers. Whatever happens, the stakes have never been higher: programmatic digital video advertising will be a $667 million market in 2013, according to Forrester.