Yield Optimization

Yield optimization (YO) is a method used by Ad servers to improve performance of a particular ad creative. The way this works is that the ad server identifies publisher impressions that a performing within the campaign parameters. It distinguishes performing impressions from non-performing ones.

The end objective of YO is to ensure that all creatives run on performing impressions. YO could be take a fairly basic as keeping track of CTRs for a site and optimizing creatives based on it. It could take more sophisticated forms feeding campaign specific parameters like time of day, publisher, ad size etc into an automated system allowing the system to take the decision of where best to place the creatives.

Factors to be kept in mind regarding YO

YO Support

Not all ad servers support YO, though. Many ad servers assume that if you have bought certain inventory, it is an informed decision. The premise therefore is that if you have taken a considered decision to buy certain inventory, it has got to be performing inventory.

Ads exchange and YO

The average ad exchange supports high-level YO. This is so because you are buying inventory on the fly. It makes sense, therefore for you to know whether an impressions is going to give you the required mileage or not.

Learning Curve

Automated or machine learning based YO systems need time to learn. In such systems, in the initial stages some impressions will be lost to learning. In the initial stages of the campaign, the ad server will place creatives with a number of publishers and slowly, based on learning, weed out non-performing impressions. We would eventually get there, but there will be a period of non-performance or more likely, low performance.

To sum up YO is a wonderful tool that an advertiser has. The use an effectiveness of YO to a particular campaign is essentially about understanding an ad server – features and functionally – in depth and harness these features to get optimum results.