Labs #1 – Upside Down

Steven Wang - July 18, 2019 - 0 comments

Welcome to happylane Labs! This is a weekly series in which we’ll be running a set of different, unique, or creative ads on Facebook and Instagram in order to gauge their effectiveness.

The reason we decided to run these unconventional ads is because people have become desensitized to ads. You’re bombarded with ads on your phone, when you look at a billboard, before you watch a Youtube video; even when you occasionally turn on your radio, you’re going to hear ads at some point. Eventually, you’re going to begin to tune them out. That’s why we decided to test ones that might break the mould and catch the viewer off guard, rather than to make the best possible “normal” looking ad.


This week, we ran a set of ads which had their images flipped. Each of them had the same copy, with just the image differing between ads.

We decided to run this particular type of ad because of an ad that we saw that involved an upside-down image of a man which received a lot of engagement. More importantly, it caught our attention and made us actually stop scrolling in order to rotate our phone and actually read the ad. Considering how you probably don’t even look twice at the vast majority of ads you see throughout your day, we’d say that being able to catch our attention like this is a pretty rare feat. Therefore, we wanted to see if we could replicate its success by creating similar ads. If the format worked, then it was a strategy we could build off of in the future.


We set the target audience to both men and women in Canada and the United States, ages 22 – 50 with an interest in marketing or SaaS. This was a pretty broad audience, but since our company is rather new we wanted more reach in order to get more initial exposure.

The objective was set to traffic and optimizing for landing page views, since our product is in open beta so we wanted the maximum number of signups possible. Although optimizing for lead generation was also an option and likely would have yielded better results, we wanted to see how much of a difference the creative advertising could have, instead of just showing the ad to the people most likely to click it.

The first ad we ran was a standard looking ad. The text across reads “Website personalization for different audiences” in order to emphasize our key product offering. The images behind the text were a man on a bike on one side, with a man sleeping on a chair on the other side.

The second ad we ran was a plain image of a woman smiling on a beige background.

The third ad we ran was an image of The Scream, a famous painting created by Edvard Munch in 1893.

For all three of these ads, the copy read:

“Happylane increases your engagement and conversion rates by allowing you to easily customize your website for different audiences.

Show the right perspective to convert visitors into customers.

No coding required.”

From left to right: standard, smiling lady, the scream

We let all three of them run concurrently, and Facebook would optimize for which ad was the most effective for whichever situation. The ads were set to run for 3 days (Friday – Sunday), with a daily budget of $10.


The ad reached 9640 people, with 55% being men and 43% being women. Predictably, the majority of interactions fell within the 25-34 age group. However, the landing page views were highly skewed towards the men, with 6 landing views compared to just one for the women.

While the reach and impressions were good, the unique link clicks and landing page views were a bit disappointing. This could be due to a variety of reasons; for example, two of the days that the ad was running on were on the weekend, when internet traffic generally dips a bit depending on the website, which definitely could have affected our results.


These ads could be misunderstood. While we tried to allude to the fact that the images were intentionally flipped upside down in the ad copy, there was a good chance that some viewers may have thought that the image was flipped by mistake. This could potentially paint us in a negative light, since it’s reasonable that a viewer might find it sloppy and unprofessional that we would use an incorrectly oriented picture and not notice.

Another limitation was that one of the images we used was of The Scream. This is a very well known painting, and chances are that most viewers have probably seen it at some point; even if you don’t know what it’s called or what it’s about, it would look familiar or recognizable to most people. However, when the impression size is this large, there are bound to be some people who have never seen it before. This is where the experiment doesn’t work as well.

The rationale behind using a well-recognized painting like The Scream was that people would see something they are accustomed to, but then recognize that something was off about the picture, which would prompt them to take a closer look at the ad. But this idea doesn’t work with people who have never seen The Scream – they might not realize anything is different at all and move on, thinking it’s just a random abstract painting. With this type of viewer, the main draw of the image doesn’t work in the intended way, making it ineffective.

What we could have done better

We neglected to create a control group to compare the original ad to for results. We went ahead and created each of the upside down ads, but we had nothing to base our results off of. There was no way to tell if these ads would actually be any better compared to a standard one, only relative to each other. This was a major oversight, and will be an essential part of the experiment for next week’s campaign.

We’ll be running these experiments weekly. For updates on this and any new features we’re adding to our app, make sure you subscribe to our blog and follow our social media pages on Facebook, Twitter, and Instagram!

Related posts