Why Completion Rate Is Only Half The Battle In Video Advertising

The dominant metric in digital video is the almighty completion rate. Report after report points to increased completion rates as a sign that video advertising is hitting the mark. If the viewer watched my 30 second ad to completion, the brand message and impact was made, right?

Well, yes, assuming that the viewer did in fact watch it. But this is an old issue, one that also appears with regular TV. Did the viewer get up and use the restroom or go to the kitchen? Fast-forward through commercials on the DVR? Just because an ad was broadcast on the screen doesn’t mean a viewer actually saw or heard it.

In digital, there are even more ways to miss or avoid the ads: clicking on another tab or scrolling down the page. Also, the same kitchen and bathroom breaks apply through the multiple ad spots before and during the show. Now, if you throw in the added curve ball of digital fraud and viewability issues, you won’t even know if a viewer was there in the first place. Like the proverbial tree that falls in the forest, your ad can register 100% completion -- but does it mean anything if nobody was around to see it?

We recently took a look at performance data for a couple of ad placements, and found the following:

Site One – 93% Completion Rate, .01% Engagement Rate, 1 second Time Earned

Site Two – 87% Completion Rate, 4% Engagement Rate, 28 seconds Time Earned

Engagement can be a rollover to see more content, a click to visit a site, a share with friends: really anything that is an opt-in consumer action, proactively taken. And for engagement that occurs right there inside the video ad unit, Time Earned is a measure of how much opt-in time the viewer chose to spend with your brand content.

An advertiser or agency that only prioritizes completion rates or had no engagement would have no idea that site two dramatically outperforms site one. A campaign with high Engagement Rate and Time Earned provides significantly more brand impact, resulting in lots of opt-in earned media and consumer actions. This makes site two worthy of more budget and optimization vs. site one.

Without needing any additional viewability or verification technology implemented on the campaign, it would be plain to see that site one is a red flag. Engagement and Time Earned is so low that the most obvious assessment is that the ads on site one were predominantly non-viewable to begin with. Or it was in a position that was technically viewable, but well outside consumers’ focus and attention. The completion rate is high not because it was a great ad placement, but for exactly the opposite reason. There was simply no one there to stop the ad or click away.

What about click-through rates, you ask? That’s a core pre-roll performance metric, so can’t those rates provide the same insights about ad placement? No, and here’s why: CTR is low to begin with, and the difference between good and bad performance is generally negligible. More importantly, it has been proven that fraudulent bot activity can simulate clicks and drive CTR. But bots can’t roll over a call-to-action overlay that opens a branded canvas and engagement slate. Bots can’t play a product video embedded from YouTube, scroll through an image gallery, read a Twitter feed and post their own tweet, etc. Basic CTR can be fraudulent, while Engagement Rate and Time Earned cannot.

I’m certainly not the first to suggest that engagement-based measurement on video ads is a proxy for viewability. With more transparency coming to the industry and more publishers allowing for URL-level volume and performance exposure, things will continue to improve dramatically. Combining viewability with advanced video engagement metrics paints a full picture of how an ad is truly received. So let’s all embrace the very things that make digital advertising so compelling and effective.

Recommend (4) Print RSS
1 comment about "Why Completion Rate Is Only Half The Battle In Video Advertising".
  1. Ed Papazian from Media Dynamics Inc , June 3, 2014 at 4:49 p.m.
    It has long been obvious that purely electronic indicators of "engagement" need to be verified by other means. While some have suggested that sales results might do this job, such data is not available or, if it is, can't be attributed specifically to the effects of a particular media exposure. One way to get at this problem might be to match electronic measures with viewer ad recall and intent-to-buy or similar "impact" metrics. If a certain configuration of electronic indicators consistently produces good or bad results in recall/impact then such findings might give advertisers more confidence when trying to interpret electronic-only data.