Case Studies

How APP Helped Greenhouse Group Achieve Lower CPCs

Much has been made of machine learning technology’s potential impact on digital advertising. The idea is that the right kind of algorithm can make many of the decisions associated with a programmatic campaign – incorporating variables like user segments, real-time bidding trends, and inventory performance – faster and more accurately than a trader equipped with today’s tools can. That way, us humans can focus on higher-level questions of campaign strategy and goals.

Blog Post
The AppNexus Team
Reading Time: ~5min

We’re certainly betting on such an outcome with the release of the AppNexus Programmable Platform (APP). And if you’re wondering what a machine learning-powered programmatic campaign looks like in practice, then look no further.

We’ve gotten APP into the hands of some of our key clients, including Greenhouse Group, a data-driven agency in the GroupM family. Greenhouse Group came to us looking for a way to lower CPCs on behalf of American Express – one of their biggest clients – while also decreasing the time traders spent manually optimizing campaigns. They didn’t come away disappointed. Read on to learn what happened!

The Challenge

Greenhouse Group had been delivering top-notch results for American Express for a long time. But they set two ambitious goals for their next campaign with the financial giant:

  • They wanted to drastically lower CPCs and
  • Reduce the amount of time traders spent manually adjusting line items

In order to meet those goals, Greenhouse Group knew it needed a powerful, customizable DSP to take advantage of the firm’s deep data science expertise. At the same time, they also needed a tool that was simple and intuitive to use in order to reduce setup time. The two concepts – performance and usability – seem at odds with one another. But luckily, AppNexus answered the call.

The Solution

Greenhouse Group chose to test two separate components of APP: Augmented Line Items (ALI) and the AppNexus Programmable Bidder (APB). Here’s a quick rundown on each:

  • ALI is our streamlined UI for campaign setup and delivery. With a simple, intuitive workflow, It allows traders to start a campaign by simply choosing a goal KPI and a few targeting parameters – without having to set up multiple line items. From there, APP uses machine learning to automatically hit the trader’s defined KPI goals by purchasing the optimal inventory and automatically adjusting bid strategy in real time.
  • APB is a tool that allows more sophisticated traders to quickly and easily import their own proprietary data onto the AppNexus platform. While APB’s use cases are typically a bit more complex, traders can operate it from the same simple UI as ALI.

By making both ALI and APB available to traders, the AppNexus Programmable Platform ensures that traders can meet each campaign’s unique goals, no matter how nuanced. 

As one of the industry’s most sophisticated buyers, Greenhouse Group chose to test both. They set up a month-long A/B test between an ALI line item, an APB line item, and a legacy line item created and manually maintained by one of their best traders.

The Results

By the end of the month, it was clear that APP could meet the goals set out by Greenhouse Group. The ALI line item achieved 13% lower CPCs with 73% less operation time spent when compared to the legacy line item. APB performed even better, with 16% lower CPCs and 81% less operation time. Read the full case study here to learn more!