Sharethrough Intros Algorithmic 'Quality Scoring' Into Bid Decisioning

Sharethrough, a supply-side platform (SSP) for native advertising, has announced the introduction of a new algorithm through which every piece of content will run before being delivered.

The “quality filtering” algorithm -- dubbed the Content Quality Score -- is meant to grade the quality of the content of the native ad before it’s served to the publisher’s site.

In theory, says Curt Larson, VP of product at Sharethrough, the quality score plays a role in the bid decisioning. If two advertisers bid $1.50 for an ad, for example, the advertiser whose content receives a higher “quality score” will win the space.

“But ultimately the goal is to put the publisher in the driver’s seat,” said Larson.

That makes sense, but why would a publisher do anything other than take the highest bidder, assuming the ads are of relatively equal quality?



What balance is right for a publisher will vary,” acknowledged Larson. “For example, a well-known content publisher with a native placement in the middle of their homepage feed is likely to have a high priority on the quality of each piece of advertiser content that fills that placement, but smaller publishers may not have the same priorities. We're early in the programmatic native ecosystem, and we'll know more about how it develops over time.”

Larson said Sharethrough’s quality score differs from others, namely Google’s, in that its grading is based on unique signals based on the type of content.

“For example, for YouTube videos an important signal is the ratio of likes to views. For articles, we look at the sentiment of tweets about the article as well as related articles and topics discussed in the article. So each category of content is unique,” he explained. "We explicitly do not consider anything related to the ad creative -- i.e. the click-through rate, the headline, etc. -- since these do not reflect the value the user actually gets from the underlying content.”

The idea of a quality scoring system automatically “grading” the worth of an advertisement -- and that grade subsequently playing a factor in whether or not the advertiser will win the space, all without their knowledge because the entire process is automated -- raises the existential question of under-the-table deals, in the same vein of kickbacks or rebates, which are a hot topic right now.

I could see technology like Sharethrough’s raising an eyebrow or two. To be fully clear, I am not accusing Sharethrough of this type of shady behavior at all, or even suggesting that it might be happening. I simply think their new technology raises the question, especially in light of the discussion -- and repercussion of the discussion -- taking place on Madison Ave. right now.

I asked Larson how Sharethrough ensures buyers that DSP #1 isn’t getting preferential treatment over DSP #2 when it comes to the Content Quality Score given by their algorithm. He said that “given the real-time nature of RTB with new creatives and bids constantly coming in, it would be very hard to ‘game the system,’ and we definitely haven't built out any features surrounding that.”

Collusion potential aside, the more pre-bid or pre-serve data publishers and advertisers can get their hands on in the automated world, the better. Technology like Sharethrough's represents the race from the bottom, not to the bottom. And it's not just better for marketers and publishers, but it helps protect consumers as well.

Next story loading loading..