The Problem
Creating ad campaigns is time consuming. You need to write creative copy, get images, add related keywords, setup targeting, and you need to do this over and over again on lots of ad networks.
Once running you need to see what is profitable and what is not. To do this you need to manually pull spend and revenue stats and reconcile.
Then you you need to optimize your campaigns by pausing ads, adding new ads, changing targeting, adjusting when ads run, adjusting your buyside keywords, adjusting bids, adjusting budgets, pausing bad sites, filtering fraud, etc.
The bottom line is that to do this you would need a huge team dedicated to managing your campaigns 24 hours a day 7 days a week.
The Solution
Eva has been programmed to do all of that for you with no human intervention. Eva will create ads on all the major ad networks. You feed her a landing page and she does the rest. We have integrated chatgpt to create variations of ad copy, eva finds related images, and she finds the best keywords to target. She then creates all these ads using targeting settings based on what she has learned previously was successful.
Once campaigns are live she is constantly optimizing them in order to maximize profits. All her changes and the results of those changes are fed into her neural network. This allows her to get smarter over time and find trends that would otherwise go un-noticed. One such trend she recently found was that ads for business related products should not start before 5am pst. This change alone boosted profits by 5%.
Eva’s neural network takes her learning’s one step further by applying an audience mood. By applying different categories of ads to different peoples moods, she is able to further increase advertiser ROI’s. For example on cold days she would show more ads regarding tropical vacations and on days stocks are bad she would bid less on investment verticals.
Her neural network is setup with hundreds of input signals and tens of thousands of hidden neurons. Each input is mapped and scored against all possibilities. The selected path is then scored based on how successful it was. A 1 is the best possible income and a 0 is the worst. Anything > 0.5 is considered a success. Some sample inputs would be automotive category, running on bing, with a budget of $100, a bid of $0.50, with a schedule change from 4am start time to 5am start time. If this resulted in a little bit more profit it would be given an output score of 0.6 if it resulted in 100% more profit it would be given an output score of 1, if profit went down it would be scored under 0.5.
