Expanding Optimization Capabilities to Execute CTR Goals

Over the past several months we have been building out our machine learning algorithms to support a new goal type – click through rate (CTR). We know that many of our clients care about optimizing to CTR goals. While it was sometimes possible to use CPC optimization to achieve a CTR target, this wasn’t an optimization solution that worked automatically. Without the goal type, customers could not use line items to reflect their business requirements or achieve the advertiser’s desired outcomes and would instead resort to manual optimization.

New Feature in UI:


Behind the Scenes

The algorithms that power CTR optimization achieve advertiser goals by automatically determining 1) what to buy and 2) how much to pay.

What to Buy

When using CTR optimization, Augmented Line Items (ALIs) buy inventory that has performed well historically, as determined by platform-wide data. If the actual performance of the inventory meets the CTR goal, the line item will continue spending on it. Likewise, it will stop spending on inventory that does not meet the goal. For a deeper look into how we discover performant inventory, read our trader wiki.

How Much to Pay

Our system predicts the value of an impression based on your goal and six factors: the specific placement, device type, browser, URL, geo, and size. Our click prediction system values each factor independently in order to produce a bid price that will ensure the line item wins impressions most likely to result in a click.

Insights and Analytics


Users already have the ability to derive quick insights and monitor performance using the Analytics tab. Now, they’ll also be able to track insights against CTR goals. Depending on the line item’s revenue type and payment model, the most relevant metrics will be dynamically displayed to you, along with a performance graph and a chart that provides data on delivery health. The tab also presents inventory that the line item has stopped spending on and the associated click through rate. This will allow clients to continue to proactively identify problems and check performance against outcomes when using this new goal type. Users will continue to have access to standard reporting.

Release Timeline

We will be releasing this feature to all AppNexus Programmable Platform clients in the coming week. Watch our Release Notes page for the specific date.

In addition, there will continue to be optimization products released over the next couple of months. Sign up to get updated for new product releases here, and if you have questions or other ideas for optimization, please provide your feedback!

Natasha Harpalani

Read more posts by this author.

New York

Subscribe to AppNexus Product Blog

Get the latest posts delivered right to your inbox.

or subscribe via RSS with Feedly!