A Publisher’s Responsibility is Not to Take on Risk
I read with interest details of the Google Attribution tool, which marries online and offline activity, leveraging Google’s near-ubiquitous user tracking:
"The Sell Sider" is a column written by the sell side of the digital media community. Today's column is written by Tom…adexchanger.com
At the end of the article, the suggestion is for publishers to “go on offense”:
… publishers can leverage their own low incremental costs to deliver outcomes that are of much higher value to ad buyers. Rather than defending the indefensible — premium CPMs in the face of virtually unlimited supply — the publishing narrative focuses on driving and getting paid for meaningful business outcomes for buyers.
Publishers view making revenue based on outcomes to an advertiser as extremely risky. They cannot control the user experience on the advertiser site. Also, the publisher CFO will not get the warm and fuzzy feeling if they do not have any guarantee of revenue for a given month.
If you look at the risk associated with different types of ad buys:
Publishers should not be optimizing their audience to buy advertisers’ products. Publishers have historically been focused on attracting an audience and learning about their audience so they can sell them to advertisers. It is the advertiser’s job to take it from there.
Advertisers, on the other hand, are getting more and more data and tools to equip themselves to handle risk. In fact, it is getting to the point that the only risk an advertiser has is that the data they are basing their decisions on is either sparse or wrong.
However, this is changing quickly:
- Anti-fraud technologies are filtering out bots at multiple levels in the ad stack.
- Site laundering is being handled by a push for transparancy as well asthe ads.txt initiative.
- User identification is being handled by cross device technologies as well as coalitions to support audience identification.
All of this ultimately will allow an advertiser to know a lot of information about a publisher’s audience before bidding in the ad auction.
The best advertisers will be armed with tons of proprietary information about their potential audience and will train bespoke machine learning models to predict outcomes and make decisions for them at scale. This will take the risk to an advertiser out of a CPM auction because models will know whether every single auction will be profitable before bidding.