Hello HNers,<p>Posting ads on AdWords confused me at first, and arguably still does, and I found that other people shared the sentiment. So I used machine learning to make an automatic keyword selector and campaign manager, and put together a company around it because I think this would be useful for others beyond just myself.<p>Right now, the backend is useful for me as-is, but I wanted to know what you guys thought of the latest prototype frontend I have put together: silvos.com/captain-crunch. (I have written an account scheduler for ads which sorts each ad into an account, so that is why the ad is the only input. Also, I am not a skilled web developer, but this prototype should get the point across. A user buys the ad through the interface, and receives an email at their PayPal email address with the number of clicks the ad received, without any manual work on my part.<p>I am confused because this doesn't feel like the right design of simple inputs and useful outputs. I have not been able to figure out what is wrong on my own, so I decided I would get your qualitative perspectives.<p>If you're curious about how the backend algorithm works, email me (saketh@silvos.com) and I can run a test campaign for you.<p>I have a lot to improve here so I appreciate your feedback. Thanks, everyone!
This is a really cool idea! As someone who does a lot of SEM/social advertising, I think this has a great deal of potential. I think if someone did this for the Facebook Ads API (currently in Beta, but available if you apply) they could make a killing.
I do $100,000 per month in SEM between Google and MSN. So for me, I obviously don't see value.<p>But if you want insights into something in particular, I could help out!