An often neglected optimization tactic in programmatic advertising can be performed at the exchange level. The ad tech ecosystem is littered with dozens of exchanges offering inventory of all kinds, from publishers to mobile apps and pre-roll video. There’s no shortage of companies out there highlighting the reach and access to these exchanges as part of their value proposition. However, as exchanges themselves have wide ranging levels of inventory quality, a statement such as “we reach 70+ Ad Exchanges” is only useful if the vendor is able to extract exchange data and interpret it to make efficient decisions. I use Pixalate’s trust radius to determine which exchanges to focus/buy based on the category of client. Pixalate’s ranks exchanges based on a number of factors including inventory quality, viewability and engagement.
This brings me to exchange level campaign data that I’m able to share from Index.Digital’s DSP. Reviewing this data got me thinking about the benefits of having a clear line of sight into performance by exchange. AOL Marketplace in this instance is 20+ times more efficient from an eCPA perspective vs. Google at the time I was writing this post (18 cents vs. $4.00).
A simple 50/50 segmentation analysis between top vs. bottom performing exchanges can have a significant impact on performance. Further analyze performance at the site level and have rules to turn off sites that are not delivering against eCPA goals. This can be done after a certain impression threshold has been reached and can be an automated. These are fairly simple steps a campaign manager can have in place to maximize a client’s budget.
To summarize here are 3 key takeaways that can help improve overall campaign performance.
1. Access to exchange level performance is key to driving optimizations. Review metrics across the board. CPMs on certain exchanges may cost more, but could also translate into lower overall eCPA’s.
2. Measure performance against campaign goals (conversions, eCPA) by exchange. Ensure there’s sufficient campaign data (spend, impression levels) before making wholesale changes.
3. Set rules to turn off publishers within exchanges that are affecting performance against campaign goals.