When Data Goes Bad

Scott German is general manager of Adroit Digital, an online advertising company that owns a data cooperative.

It’s rare to go a day in the online ad world without hearing about data and its ever-growing impact on the universe of digital marketing. Data is a major part of online advertising, and with that comes additional investment. Marketers know they need data, but more importantly, they need data that will help drive their marketing goals, rather than add bloat to the budget.

Failing to understand third-party data and its source(s) means the marketer lacks real insight into how a campaign will perform when that data is aligned with media. Data investment typically isn’t a matter of millions of dollars, but it’s significant enough that marketers can’t afford any waste. Marketers should consider certain things before making data decisions.

Data goes bad in about three days (for most activities). Only fresh data helps marketers find qualified audiences to complete their digital objectives. Nowhere is this more important than in online commerce, where targeting recent site visitors and in-market consumers, and prospecting new customers rely heavily on the recency of data.

The problems with using stale third-party data to drive an ad campaign start with the time elapsed between cookie creation and ad call, depending on the advertiser and offer. Airline tickets may have longer timelines, but items like clothing and shoes are typically purchased within a short time frame, giving the data an expiration point. Further, some advertisers mistakenly target consumers who already made purchases on their site, wasting money and creating a bad customer experience. A Toluna survey confirms those sentiments, with 18 percent of online consumers saying they have negative feelings for a brand after seeing an ad months after initial purchase. Fresh data allows models and targeting systems to better understand consumer patterns and purchase life cycles, leading to smarter ad buying.

The source matters when determining data’s value. If it’s not coming from a first party, you can’t be sure of its accuracy. When audience data is derived from media insights rather than actual purchase behaviors, it’s difficult to tell if the consumer is actually in market and how long it will be until the purchase. Auto-intender data rarely comes from a manufacturer, but from another party tangentially related to auto sales (think Kelly Blue Book or Edmunds).

First-party purchase intent insights are hard to come by in today’s market. If an online retailer identifies a consumer who has been shopping and will likely return to make a purchase, then why sell that data? That intelligence is highly valuable to their marketing efforts, and they have little incentive to sell it to a data exchange for pennies on the dollar.

It’s safe to say that the bigger data exchanges likely have trouble maintaining the value of their data. These companies are incentivized to build big cookie pools, so they push to make them bigger, dropping the value of the audience segments they sell. At a 75-cent CPM, there’s a much bigger incentive to sell the biggest possible pool of 2 million people, even if that audience contains only 1 million consumers who are actively in market. And that is contrary to the marketer’s need, which is to identify the most qualified population.

In an exchange environment, an influx of buyers means that many advertisers are seeking the same audience. This competition for the same cookies drives up prices. If two competing companies buy the data, it further erodes the value. Now the advertiser has overpaid for outdated data from a less-effective source.

Advertisers who find themselves with lackluster results from campaigns powered by third-party data should take a deeper dive into the data sets these campaigns are based on. The next time an agency or tech partner pitches a prospecting or brand awareness campaign that combines data and media, ask about the source of the data, the number of other buyers, and the freshness. While third-party data can be a great way to target and extend audiences, it can only be productive if it fits an advertiser’s needs. Buying bad data isn’t going to bankrupt a marketing budget, but too many ineffective campaigns could cost a marketer their job.

Image via Shutterstock

Digiday Top Stories