Artificial Intelligence is Helping to Eradicate Waste in Paid Media

Paid Media

For much of online paid media’s existence there were two main pricing schemes that media buyers have grown accustomed to – impression-based and click-based cost. Unfortunately, these two pricing schemes can be strife with waste, fraud and abuse by nepharious bot-peddlers.

As a result, the IAB Tech Lab introduced a project called Ads.txt. It’s a secure method for publishers to declare who is authorized to sell their inventory in order to improve transparency for programmatic buyers. While it has helped clean up some of the waste, fraud and abuse, it doesn’t go far enough. Just last year P&G eliminated more than $250 million in digital ad spending because of brand safety concerns, waste, fraud and abuse.

At the core of this problem is the way the ad inventory is priced – impression-based and click-based cost. What series of events must occur for an ad to deliver maximum value for an advertiser? First, the ad must be seen, then clicked on, and lastly, the content on the other end of the ad unit must be engaged with by a user. Charging advertisers based on the first two gives no incentive to advertising technology companies to be concerned with engagement, which is what many advertisers want (or should want).

Engagement can be defined a number of ways, but the two most reliable are scroll depth and dwell time. According to Chartbeat Analytics, two out of three clicks on a native ad unit bounce from the subsequent page within 15 seconds. They also figured out that the users who spend 15 or more seconds with the content post-click consume 80% or more of it. This means that it takes three clicks to get one engagement on content of 80% or more. In other words, an advertiser pays for 3x the cost of one click for one meaningful engagement. Two thirds of the budget is essentially wasted.

Paid Media Waste

This waste could be eliminated entirely if advertisers bought media on a cost per engagement model. inPowered recently announced its new AI-driven dynamic cost per engagement pricing model.

Last week I spoke to CEO, Peyman Nilforoush, about how he’s eliminating waste in paid media.

We’ve known for a long time that programmatic content amplification inherently has inefficiency built into the cost per click (CPC) model. We reintroduced our cost per engagement pricing model powered by AI today to permanently solve for this.

inPowered’s technology sits on top of over 35 native and social networks. Their solution creates thousands of ad permutations and tests them at scale. Rather than optimize towards clicks or impressions it looks for just those permutations that drive 15 seconds or more of engagement. Advertisers only pay for the engagement, not clicks or impressions. That way, the advertiser avoids waste, fraud and abuse.

In addition, as the machine learning gets smarter in its optimizations over time it drives the cost per engagement lower. The savings is then past back to the advertiser in the form of more engagements. It’s called dynamic cost per engagement (dCPE) and it’s the first of its kind. It’s innovative uses of AI described above that will help clean up paid media and maybe get P&G spending again.