Online ads have been around forever, but they have exploded in the last couple of years. That’s a hell of a lot of advertisements going on, making it increasingly harder to not be overwhelmed.
So what’s the story? These are pretty big numbers. Our mobile app was broken up into several categories with an algorithm that scanned through a lot of ads. Our algorithms hit something called “likelihood surface.”
Think about YouTube content or a certain piece of music. The algorithm by Google understands this content but doesn’t know what it means or a genre of music.
The algorithm that filled those spaces with another piece of content picked up on things like “Music,” “Art Style,” “Grade Rating,” and “Music Album”.
We had no idea how much we were going to lose in time because of this ad supplement that we had been using, but not everything is so cut and dry.
After paying out to maximize the relevance of the remaining ads, we hit a roadblock — “complementing”. Our sales had nearly reached their peak, and before we knew it, we had wasted our time and money on a subclass of the wrong ads.
That’s a huge problem with a copy process. If you’re unable to see the finished product, the full sale of an ad is doomed. From an operational perspective, this bottleneck can also be explained with the “sampling strategy.”
At the point that we paid our transaction, the ads from the “complementing” class would be seen by five people who didn’t truly want to see them. For this class of ads, we dropped $20,000, half of our sales due to an unsavory experience.
The article will delve further into this feature by looking at why it happened and how we fix it, but the general problem is that in the maze of our workflows it is easy to make a mistake.
By watching each channel from a “complementing” class and fixing it completely, we can improve the relationship between the overall class of ads and the amount of conversion we will achieve.
At this point, our team needed a new name. A more meaningful brand. We pulled an ad copy from the different classes, swapped it out with a new version of it, and used our new business name.
Why am I saying “rebrand”? Because, first of all, it wasn’t our name in the first place. Secondly, we redid our company logo.
Working with three brand managers, the name didn’t resonate enough, our customers and the industry found ourselves confused. Today, we’re reaching 10k people a day. Our first customer requested the name “542′.”