With Google Barring DMP Pixels on GDN, Has It Thrown Baby Out With The Bath Water?

tom-chavezData-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Tom Chavez, CEO at Krux.

Google recently informed its data-management platform (DMP) partners that it would impose new restrictions on pixels fired by campaigns that run on the Google Display Network (GDN).

Beginning Jan. 1, DMP pixels will not be allowed to fire on GDN impressions.

“[C]ollecting impression-level data via cookies or other mechanisms for purposes of subsequent retargeting, interest category categorization or syndication to other parties on Google Display Network inventory is prohibited,” the company said.

Google cited concerns for potential data leakage and pixel loading as reasons for the change. I commend Google for taking concrete steps to stave off these problems. Data leakage is a serious threat to the publisher’s greatest asset – its audience data. And pixel loading is troubling since it gums up web pages, prompting readers to click away.

But as with many things in life, good intentions can result in unintended consequences. While it’s too early to know for sure, Google’s recent move may fall into that category and create blowback that’s bad for Google. That’s because the policy, if applied inflexibly, ignores the needs of one of Google’s most important constituents: marketers.

Marketers seek to create positive brand experiences for their customers. They don’t want to bludgeon people with heavy-handed, repetitive ads. Consumers, meanwhile, want to see ads that are relevant to them at just the right moment. Sadly, Google’s new policy forces marketers to fly blindly by stripping them of the instrumentation they need to manage and meter their messages. In fundamental terms, the policy subverts Google’s espoused objective to bring an end to marketing without accountability.

Without DMP pixels, marketers must revert back to carpet-bombing users. Given that nearly every brand that engages in programmatic marketing purchases inventory from multiple exchanges, sometimes as many as 10, Google’s approach makes it impossible for buyers to apply global frequency management to inventory purchased via GDN. If, say, leading CPG brand marketers can’t track which consumers saw an ad on GDN, then they won’t know if or when their brands’ frequency caps have been busted. In an advertising industry that’s hurtling toward real-time programmatic methods, it simply won’t work for marketers to receive this information in batch mode. They need to know it in real time so they can be smarter about the next impression they buy.

Brands must either accept the resulting risk to their brand experiences or eliminate GDN from their media buys altogether.

One might then argue that brands can always limit their media acquisitions to a single ad exchange, which will make the requirement of global frequency capping moot, but that isn’t likely either. Publishers offer different parts of their inventory to multiple exchanges. For brands seeking maximum reach, purchasing from more than one exchange is an absolute necessity.

Google’s decision to effectively wall off its garden also undercuts another critical marketer goal: gaining insight into its customers’ journey through attribution modeling and view-through measurement. If brands can’t track which consumers have seen their ads throughout the display universe, then their attribution efforts will be hobbled and they won’t get accurate insight into how to best allocate their marketing spending. Without precision, they lose confidence. Without confidence, they take their dollars elsewhere.

The dynamic reverberates on the publisher side of the equation as well. Attribution benefits publishers by giving them the cold hard data they need to demonstrate their contribution to outcomes of interest to marketers, such as conversions, clicks, downloads and engagement broadly conceived. But if an advertiser can’t understand whether an ad seen on a GDN-purchased impression ultimately led to an outcome that matters, then that publisher has little ammo to justify repeat business, much less premium CPMs for its inventory.

It’s peculiar to me that Google has chosen this route when, throughout its history, the company has consistently opted to break down barriers for advertisers, publishers and consumers alike. It has created easy insertion points for services, such as paid search and AdSense. By making it simple for publishers to send Google their inventory, Google achieved huge scale quickly and decisively. Removing friction and breaking down barriers helped it grow its business not by carving up the pie, but by baking a much bigger pie. Why the change in strategy now?

DMPs aren’t principals in this game. DMPs exist to serve the objectives of publishers, marketers and consumers. Their goals shape everything they do. By making their data portable and actionable, DMPs give brands the confidence to bring more of their dollars and inventory to digital channels. This, in turn, promotes a bigger, healthier ecosystem.

Criticism without craftsmanship is cheap, so here’s what I think Google should do: Design a simple protocol that allows DMPs to demonstrate they’re acting as agents, not principals, for marketers when they deploy pixels to measure real-time ads. Track and certify the pixels those DMPs deploy on behalf of marketers. The certification, if appropriately designed, will differentiate between DMPs that drop pixels to accumulate data that they’ll use to trade for their own account vs. DMPs that deploy them purely for measurement purposes.

I applaud the intention behind Google’s recent move. When you game it out, however, it bears negative consequences for marketers, publishers and, ultimately, Google.

We can clean the bath water, but save the baby.

Follow Tom Chavez (@tommychavez), Krux (@KruxDigital) and AdExchanger (@adexchanger) on Twitter.

 

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>