“The Sell Sider” is a column written by the sell side of the digital media community.
Today’s column is written by Chris Kane, founder at Jounce Media. Chris will present “Server-Side Attribution” at AdExchanger’s upcoming PROGRAMMATIC I/O New York conference on October 25-26.
The growing power of walled gardens isn’t all bad for the rest of the media world.
Facebook, YouTube and their walled-garden peers are paving a path toward a pixel-free internet. Their obsession with user experience has made them allergic to third-party tracking pixels, and they are pioneering new methods of server-side measurement.
These server-side solutions provide the independent auditing that marketers demand without bogging down page-load times with third-party code. It’s a dramatic departure from the pixel-first culture that has invaded most marketing organizations.
Marketers, enabled by their ad tech partners, have a bad habit of defaulting to tracking pixels for all measurement challenges. Nielsen’s Digital Ad Ratings (DAR) product came to market as a pixel-based solution for validating age and gender targeting. Integral Ad Science (IAS) came to market as a pixel-based solution for measuring viewability and fraud. And Visual IQ came to market as a pixel-based solution for moving beyond last-touch attribution.
All three companies, and many of their competitors, now embrace server-side measurement integrations with a growing roster of media companies who restrict pixel-based data collection.
As advertisers get more comfortable with server-side measurement, the rest of the publisher world might be able to do something similar. That would be good for publishers and consumers alike.
Here’s how it works:
YouTube Demographic Verification
When big brands move their money out of broadcast television into digital video, they bring their TV metrics with them. Specifically, brands want independent third parties to measure gross rating points, target rating points, reach and frequency of digital video ad buys. And since 2012, that has meant firing a Nielsen DAR pixel or a comScore vCE pixel on every digital video impression – except YouTube. Google’s strength has allowed it to enforce strict “no pixel” policies on YouTube.
But that doesn’t mean Nielsen and comScore can’t measure YouTube demographics. Nielsen and comScore receive from YouTube anonymous user IDs for each impression and merge those with records of video impressions served on non-YouTube properties. While server-side data is transmitted more slowly than the classic pixel-based implementation, the richness of data is identical. The marketer achieves unified measurement while respecting Google’s pixel restrictions.
Twitter Viewability Measurement
Marketers and their agencies stay awake at night worrying about how much of their budgets are wasted on out-of-view impressions and bot traffic, and most of the ad-supported internet now supports independent viewability and fraud verification from companies such as MOAT and IAS.
But measuring inventory quality creates an even bigger user experience penalty than standard tracking pixels. It requires JavaScript that soaks up processing power and slows page-load times. Twitter refuses to load third-party JavaScript in its app and on its website, but it does provide rigorous server-side impression logs to viewability partners for independent auditing. The MRC hasn’t yet blessed this server-side measurement technique, but it is already vetting similar approaches used by Facebook and YouTube, and you can expect Twitter’s solution to be audited by the MRC in the coming months.
Facebook Multitouch Attribution
Attribution is the hot topic of 2017, and savvy marketers are embracing multitouch attribution models that move beyond the false positives and negatives of last-touch attribution. The classic multitouch attribution solution requires firing a tracking pixel on every impression, which is a non-starter for Facebook.
Instead, Facebook has developed a server-side integration with select multitouch attribution partners. When a user visits a brand’s website, the attribution vendor notifies Facebook, and Facebook provides logs of that user’s prior ad exposure on Facebook, Instagram and the Facebook Audience Network. It’s not quite as robust as marketers would desire (“What about the people who don’t visit my website?”), and it isn’t well suited to offline businesses (“I don’t even have a website”), but it is a massive step in the right direction.
There are more examples. Pinterest, Snap and Spotify are all live with server-side measurement solutions that look a lot like the ones mentioned above. Apple News and Pandora are moving down a similar path.
These server-side measurement techniques are new and largely unproven, and we are still in the earliest stages of the adoption curve. Measurement solutions that used to be the exclusive domain of $50 billion media companies are now accessible to $500 million media companies. And while it may still be a while before a $5 million media company can take advantage of server-side measurement, the trend is clear. Pixels are going out of style.
Follow Jounce Media (@jouncemedia) and AdExchanger (@adexchanger) on Twitter.
