Omega Ad Earnings Optimizer: Dynamic Equilibrium

For StudentsReview, we had a problem — managing advertising inventory from a number of ad networks with fluctuating pricing, yield, and failure rates.  The largest & most serious problem was “default loss”, a condition arising from varying yield and failover to secondary ad networks lag.  So we wrote this algorithm that monitors and delivers ads to a particular network in a state of dynamic equilibrium — the less a network defaults (fails to serve), the more ads Omega delivers to it, and the more it defaults, the less ads it gets delivered.  Over an hour or so, the choice whether or not to serve to a particular network, and a network’s likelihood of default becomes “well matched” within 5% of the actual performance stats.  This has allowed us to recover about 20-25% of the lost 30% of ads from ad network chaining.

Of course simultaneouosly we compute hourly pricing statistics and decide a realtime order to get the highest price.  It will never be as good as selling your own ads directly, but it will get you all the revenue you can get while minimizing loss to the networks!

Unfortunately, the code for this is not free, and I am writing a paper for Management Science on this topic (in addition to the provisional patent I filed), but I do provide access to it via a web management service and subscription, which can be signed up for at