Read on Twitter
When it comes to creative testing at a tactical level, you don't need lots of videos. You just need lots of video openers.
Why is this a better approach?
Here's 4 of the top video ads (driving sales) from an advertiser we work with all targeting the same audience...
Why is this a better approach?
Here's 4 of the top video ads (driving sales) from an advertiser we work with all targeting the same audience...
Despite the various lengths of all four videos, viewership always looked similar:
- The avg watch time was ~5 seconds (range was 3 - 6 seconds)
- Roughly 1/4 people watched ~5 seconds (range was 16% to 25%)
- Viewership retention leveled off at ~5+ seconds
Notice a trend?
- The avg watch time was ~5 seconds (range was 3 - 6 seconds)
- Roughly 1/4 people watched ~5 seconds (range was 16% to 25%)
- Viewership retention leveled off at ~5+ seconds
Notice a trend?
It's not that one ad was inherently better than any other.
(Interestingly, even our worst performing ads still saw these similar trends.)
(Interestingly, even our worst performing ads still saw these similar trends.)
Instead, these trends comes down to user behavior in feed.
According to a 2017 Facebook study, video ads are watched in News Feed for 5.7 seconds on average. (given this is 2020, average watch time might be even lower now)
According to a 2017 Facebook study, video ads are watched in News Feed for 5.7 seconds on average. (given this is 2020, average watch time might be even lower now)
That mean the first ~5 seconds are the most crucial in capturing attention, driving brand awareness and persuading action to click.
~5 seconds to have someone stop scrolling, know who your brand is and what you sell and also persuade them to click in the span of 5 seconds is very difficult to do.
This is why you'll often see advertiser create ads that are 6+ seconds long.
This is why you'll often see advertiser create ads that are 6+ seconds long.
Instead of trying to accomplish all of this in ~5 seconds (which is also worth testing), if you're seeing good success with an existing video, an alternative method is to test different openings on the same video ad vs testing different video ads.
One approach we've taken with advertisers is testing about 5 - 10 different cold opens for any video we make.
While this may seem like a lot of different ads, it ends up being easier to build out than you might think.
While this may seem like a lot of different ads, it ends up being easier to build out than you might think.
How would this work?
Start by thinking through 2 to 5 unique openings and 2 to 5 unique headlines to accompany the opening of the video.
Then make each permutation into its own opening.
Start by thinking through 2 to 5 unique openings and 2 to 5 unique headlines to accompany the opening of the video.
Then make each permutation into its own opening.
So it might look like this:
Ad 1) Opening Visual A + Headline A
Ad 2) Opening Visual B + Headline A
Ad 3) Opening Visual A + Headline B
Ad 4) Opening Visual B + Headline B
Etc...
Ad 1) Opening Visual A + Headline A
Ad 2) Opening Visual B + Headline A
Ad 3) Opening Visual A + Headline B
Ad 4) Opening Visual B + Headline B
Etc...
We then take those 5-10 variations and run them as dynamic creative to better understand what opening combinations work best and iterate from there.
On the next batch of creative (both existing assets and future productions), we alter openings to test something new based on what did/didn't work well.
This approach is a good way to get some more creative bang for your buck.
Focusing on openings takes less time to edit and can stretch any creative production you do a bit further.
Focusing on openings takes less time to edit and can stretch any creative production you do a bit further.
By focusing on the lowest common denominator of video watchers, you'll also have the biggest impact on optimizing your ads vs trying to make new ads every time.
