Lately I've been working a lot with live action video ads. If you work in mobile gaming you probably know first hand that live action can be hit or miss, which is unfortunate considering production costs can get very expensive very fast. One way to help offset those costs is to stretch your live action ads as far as possible. If it doesn't work on Facebook, try YouTube or TikTok. If click-through is low, try re-editing. If conversion is low, try mixing in non-live action content.
There's nothing worse than spending a big chunk of your budget on a live action ad only for it to under perform your cheaper evergreen ads. Because of this risk, I've been pondering how to get the most out of live action content that isn't outperforming your non-live action ads. The first idea that came to mind was to present the live action ad in a mobile device. Video sharing apps are super popular these days, so it's reasonable to think this trick could help boost performance.
The First Test
I ran a Facebook split tests with 3 variants.
Variant #1 (the control) - "No Phone"
I found this weird public domain video of a guy running on a treadmill while a girl teases him with food. It kind of looks like something you might see on social media though, so it was good enough.
Variant #2 - "Phone (Full)"
For the second variant I simply placed the treadmill video into the screen of a static mobile device.
Variant #3 - "Phone (Half)"
Finally, in an attempt to meet somewhere in the middle, I combined the first two variants to create a third variant that starts with the video in a device but quickly zooms in so the video fills the frame.
Results and Learnings
It's pretty clear that "No Phone" was the winner. And based on the performance of the third variant "Phone (Half)" - which is mostly the same as "No Phone" - viewers definitely prefer seeing a live action ad in full frame. One possible reason for this is that the mobile device makes the video smaller and harder to see. What's more likely, however, is that the use of a mobile device made it immediately obvious that this was an ad and not just an organic video.
The Second Test
My first attempt failed so I went back to the drawing board and brewed up 3 more concepts - this time a little more outside the box.
Variant #4 - "No Phone (Border)"
I added a bright yellow border to the outside of the video. I've observed this technique used on YouTube thumbnails and Steam game search result images.
Variant #5 - "No Phone (Flipped)"
This one was a bit more extreme, but I flipped the entire video upside down.
Variant #6 - "No Phone (Shake)"
For the final variant, I added a violent shaking effect to the entire video.
Results and Learnings
Once again, the "No Phone" control ad drove the highest CTR, but interestingly the "No Phone (Shake)" variant drove more clicks and a lower CPM, thus causing Facebook to declare it the winner of the test (at 95% confidence level). Unsurprisingly, "No Phone (Flipped)" performed the worst (but still better than the device variants if we want to compare cross-tests)
The Third Test
My final idea was to capitalize on meme culture. Who doesn't love a good meme? There's something about seeing black text on a white banner that will always make me stop and read it.
Variant #7 - "No Phone (Meme Banner)"
What's more topical than 2020? For this variant, I added text above the video.
Variant #8 - "No Phone (Meme Text)"
I felt testing 2 meme formats was more thorough, so this variant puts the text directly on top of the video.
Results and Learnings
Unfortunately, "No Phone" was again the winner. I also gave "No Phone (Shake)" another chance in this test, but CTR still underperformed the control. Between the two meme variants, it would appear that the meme banner format is more engaging than text over the video, but ultimately both variants failed.
I started this mini journey looking for a way to improve performance of live action video ads, but ended with the conclusion that live action videos may perform the best on their own, without any additional fluff or effects. However, the more accurate conclusion is probably that it depends on the content of the live action ad. This treadmill video was weird and eye-catching, but ultimately it was still another stock video; it didn't advertise a product or send a clear message or call-to-action. So I would say the biggest takeaway here is to not rely on additional editing, but instead focus your effort on the concept, script, and production.