(And the advent of the blank ad revolution.)Like any good company looking to promote a product online, we recently started running online advertisements. We have been experimenting with multiple different ad platforms and methods, and have had varying success.
There's been a fair amount of press over the last few months regarding the current state of online advertising. The current widely adopted metric to measure the effectiveness of an online ad is the
click through rate of an ad. The click through rates of online ads
have been declining for years, and are horrendously low. Currently this rate is somewhere near 1%, and as low as .01% on some ad networks.
Something's not working here. On a good day, out of the 1% of people that actually click the ad, how many actually buy the product (or in our case, download it)? In our experience, that rate is somewhere near 2%. So if 1% of people actually click an ad, and 2% of those people actually buy or download the product, the overall conversion rate is a lowly .02% (yes, two hundredths of 1 percent).
Some argue that the click through rate
shouldn't be used to measure the effectiveness of online ads. Even Microsoft thinks there's got to be a better way to measure the real effectiveness of online ads
by applying traditional offline advertising metrics.
But what if all this arguing over increasing click through or conversion rates by a tenth of percent doesn't really matter? What if there was a bigger influence that made these click through statistics insignificant? Introducing: accidental clicks. How many ads are actually clicked in error? Being good engineers, we did a little experiment in an attempt to do a bit of error analysis.
Which one of these ads would you be more likely to click?
This one (PK ad):
Or this (Blank ad):
Yes, that is a blank ad. Our ad rep was a bit confused when we tried to start running the ad. Here's an excerpt from the email conversation:
Them: "Can you please resend the creative? I am seeing it as a blank image."
Us: "That should be the correct image. We are trying to step outside of the box and show the viewers a mystifying and curiosity invoking ad. We are hoping their curiosity will lead them to be delightfully surprised by our app description and the depth of our game."
Them: "I downloaded the ad again and I am still only getting a blank ad for the test ad. Can you please double check and ensure this is the right ad?"
Us: "Yes, this is the right ad. It is a completely white ad."
Now things get interesting. Both of these ads were run on the same mobile ad network appearing inside of various mobile apps. Both ads received around 100,000 impressions and were run in the US over approximately the same time period.
The PK ad click through rate:
0.35%The Blank ad click through rate:
1.50%That's right, our blank ad (the
"mystifying and curiosity invoking" one) had a 4x higher click through rate than the real ad. Now granted, the conversion rate for the real ad (aka the people who clicked then downloaded our app) was much higher than the blank ad. But how is it possible that a blank ad has a higher click through rate than a real ad?
Maybe we just stumbled on a new paradigm in display advertising. Or maybe there's another explanation.
Do people
dislike ads to the point that they will actively avoid going anywhere near them? In an ad on a mobile app, this would mean people have to avoid clicking (with their finger) on the banner ad. Solution: people don't put their finger anywhere near the ad. If people are so scared of the ad and are avoiding the possibility of clicking the ad in the first place, they are most likely not paying attention to the content of the ad.
So why would someone click on a blank ad? Our conclusion is that almost all of these clicks must be accidental. Granted, probably a few adventurous people were intrigued by the possibility of some white space and wanted to find out more, but it's highly unlikely that this would account for all these click throughs.
Most likely people were in an intense game of
Tap Tap Revolution when their finger got a little too close to the banner ad and tapped it. But if the normal "accidental" click through rate (essentially the error rate) is
higher than the average click through rates themselves, how can these click through rates be statistically relevant (let alone the methods for "improving" your click through rate by .1%)? Why bother with online ads at all?
This little experiment is by no means conclusive. Mobile advertising obviously works to an extent, but this whole blank ad thing brings up some interesting questions worth pondering. So
put that in your milk, ponder away, and try out running a blank ad for yourself.
Posts
