The True Cost of Low Advertising Render Rates
Among the many metrics that CTV/OTT publishers look at is render rate. Render rate is the ratio of the number of ads actually delivered divided by the number of ads decisioned for a particular stream. On the surface, low render rates would seem to represent an annoying illustration of opportunity cost—the ads (and therefore revenue) that could have been delivered but weren't, typically because a user stops watching in the middle of the stream.
However, there are more subtle and meaningful implications of low render rates that publishers need to be aware of. These fall into two main categories: cost and revenue impact. On the cost side, the process of making ad decisions and stitching those ads into a stream costs money. Depending on how a publisher's ad stitcher charges, these costs may or may not be completely apparent. For example, some stitchers charge a CPM fee for each ad stitched into a stream, where others will charge a CPM fee for only the ads delivered. Our world is not quite as simple as that. Oftentimes, the stitchers that charge based on delivered ads will have a fee structure that is 2-3x that of the stitchers charging for each ad. Either way, the publisher is paying for ads stitched, but not delivered.
In addition to stitching fees, publishers are also charged ad serving fees. Essentially, these are fees on every ad decision that their ad server makes, independent of whether or not the ad was actually delivered and viewed. It's not hard to see that with a render rate of 50%, a publisher's cost structure on delivered and viewed ads is actually 2x what they may think it is.
Unfortunately, there is more bad news. Low render rates also negatively impact fill and CPM rates. For most publishers, there are two main sources of ad fill; their primary ad server (servicing direct sold campaigns) and their programmatic partners. I'll start with the least worst case first, the primary ad server. Each campaign on an ad server has a budgeting and pacing line item. One of the ad server's jobs is to keep those campaigns trending along the theoretical ideal curve to satisfy the campaign's budget expenditure in line with the pacing requirements of said campaign. Low render rates throw a bit of a monkey wrench in this calculation. At scale, when an ad is committed to a particular stream, it cannot be committed to another stream without the risk of exceeding the ideal budgeting and pacing curve. As a result, the ad server will oftentimes take a conservative approach to determining where and when it will "buy" an impression for a given campaign. For example, there may be placements that are hard to fill because the ad server is "concerned" about the ability of that placement to deliver a valid impression.
This problem is compounded on the programmatic side of the house. In reality, it's the DSPs that are determining which advertisers participate in any given impression opportunity. These systems are incredibly capable of understanding and leveraging data flows to their advantage and are actively looking at render rates to determine the likelihood of a given opportunity actually running. Low render rates drive their decision logic and behavior. When you examine bidding behavior, you will see that there is a greater density of bidders against early ad pods as a result.
OK, that provides a look into fill, but you might be asking yourself, "How does this affect CPM rates since all of my rates are pre-negotiated?" Well, here's how. While you may have pre-negotiated rates for all direct and programmatic buyers, they are not all the same CPM rate. So, when you have scenarios where you have many advertisers competing for the same spot, the differential between the CPM rates is smaller than when you have fewer advertisers competing for spots. With fewer competing advertisers comes broder spread of CPM rates and therefore lower overall CPM when filling the inventory.
Admittedly, there are a number of approaches you can take to address render rates. Some of these are QoE related—in other words, make the user experience better and they will stay around longer. Other approaches are ad tech-related. I've been working on a solution that addresses the fundamental timing of when ad decisions are made for VOD which resolves the render rate problem and other associated issues.
[Editor's note: This is a contributed article from Penthera. Streaming Media accepts vendor bylines based solely on their value to our readers.]
In order for streaming services to offer a truly "real TV" experience, they need to match content, ads, and devices, as well as support rapid spikes in traffic, maintain high quality of experience, and more. In short, it all comes down to the manifest manipulator.
15 Jun 2022
NBCUniversal Senior Director, Ad Innovation Strategy, Advertising and Partnerships William Roberts-Foster discusses how considerations involving consumers and content play into the ever-evolving ad strategies of NBCUniversal and Peacock TV, based on assessments of the "mind states" consumers bring to different types of premium content in this clip from his keynote at Streaming Media East 2022.
13 Jun 2022
FAST has been the big story of the last two years, but there's tension growing between streaming services and TV/OS manufacturers over who owns the viewer and how to get those viewers to the content.
19 May 2022
Philo Director of Advertising Partnerships Aulden Kaye explains how addressability defines the value proposition for Connected TV advertising in this clip from Streaming Media Connect 2022.
22 Mar 2022
As CTV grows in popularity and programmatic becomes the primary source of ad revenue, problems around measurement and monitoring are only going to become more painful and expensive if action is taken early and robustly. The time to act is now.
08 Mar 2022