Why I Tested 7 Thumbnails
I Need A Hero
We've all heard of A/B testing, right? What about A/B/C/D/E/F/G testing?
A while back, I launched a Hero Ad for my company that we both wrote and produced. A hero ad is simply an ad that sells what your business does, what your service/product is, and answers all the objections and questions potential clients might have.
Save A Few Bucks… A Few Million Bucks
A hero ad is your Super Bowl ad — without the $56,000,000 ad spend.
So, we launched the ad with our first goal to discover which thumbnail resonates with the target audience. This led to a debate over the value of split testing thumbnails. My brand strategist and I got into an animated debate over it (we always do, that's healthy balance in my company). He didn't see the value of testing 7 different thumbnails and wondered if it would deliver any substantive data.
With over a half million views, thousands of reactions, and hundred of shares, we had not only the social proof that the ad was well received, but we now had the data to move to Phase 2 of our ad testing.
Phase 1 consisted of the following:
1 Headline
1 Ad Body Copy
1 Video Ad (Long @ 3:59 — YES ALMOST 4 MINUTES!)
And the variable for this test is... 7 Different Thumbnails.
I know 7 Thumbnails seems a bit excessive, right? Maybe. Testing thumbnails lets you find the visual that gets the most engagement from your prospective audience.
Our thumbnails were labeled internally: KFC, Hair, Presidential, Marketing, E-commerce, Cocktail, and Dildo (yep).
The data was tied to one video. I’ll let you look at the thumbnails and play the match game.
Prove It To Me
A quick note…social proof stacks on the video (aka views go up no matter which thumbnail was presented), but the reactions, comments, and shares gave us the confidence to take us from 7 thumbnails down to 3 winners, the third being marginal at the time I share this.
Round 1 Data:
KFC: 551 reactions 31 comments 54 shares
Hair: 254 reactions 4 comments 8 shares
Presidential: 87 reactions 2 comments 4 shares
Marketing: 36 reactions 1 comment 1 share
E-commerce: 28 reactions 3 comment 2 shares
Cocktail: 1000 reactions 35 comments 45 shares
Dildo: 194 reactions 8 comments 14 shares
This tells us that we have solid thumbnails to use as we start split testing other variables in the ad including body copy and headlines. We completed all the testing in about a week for around $500, not including the cost of filming the ad.
Round 2 Data:
We then killed the 4 losing thumbnails and moved to Phase 2. Meanwhile, the social proof keeps growing.
KFC: 1K reactions 94 comments 149 shares
Cocktail: 1.8K reactions 44 comments 73 shares
Dildo: 254 reactions 12 comments 26 shares
OK, let’s kill the Dildo! I bet you never thought you’d read that in a blog, huh? The dildo thumbnail doesn’t have the reactions the others do, nor does it have the shares.
Now, some marketers will ask how many sales this ad has generated. That’s not the goal of the test phase at all.
The goal of this phase is: First, to discover which thumbnail has the most engagement with the audience. Next, to dial in the copy and headlines.
We are serving these ads to a global audience that consists of people who run ads on Facebook, Instagram, Digital Marketers, Ad Agencies, etc. This is a mostly COLD audience and if they have ever seen, heard, or touched our brand, it’s been at least 6 months. We excluded existing subscription clients and 180 day purchasers, as well as anyone who likes our page on Facebook or follows us on social.
The only variable we can’t control in this is the email they use to login to social media versus the email they use for business transactions, so we have to say there is a 5%-7% audience that has seen, touched, or done business with us over the last 6 months. It is what it is.
Round 3 Data:
With the 2 remaining thumbnails in rotation, one got drunk and couldn’t stand up. The cocktail ad was good, but it just couldn’t wing it.
Winner Winner, Chicken Dinner
Did I just give it away? Yep! The winner was the KFC thumbnail and that ad has gone on to drive lots of new, finger-lickin’ business our way.
With a little over 1.6 Million (Targeted Views) over 2,500 reactions, 481 comments, and 393 shares to date, this ad is my hero!
I Said “Show Me The Money”!
With a to-date spend of just over $7,000 ($7,463.42 for you precision nerds out there), this ad has generated over $51,000 in sales ($51,124.42 for you again). That represents a 6.85 ROAS.
While the thumbnail did not generate the ROAS, the right thumbnail changed the long-term success of the ad.
Next time you let a video ad into the wild, test your thumbnails. Then channel your best Cuba Gooding Jr. and shout “SHOW ME THE MONEY” at the top of your lungs. Just make sure you’re in the middle of some place really busy so you get strange looks.
Oh, and here’s the finger-lickin’ goooood ad —> Ad Zombies Hero Ad