rocket booster Rocket Fuel logo back to home page  
Remember those wonder kids who aced every test and went on to work for NASA? Well, now they’re building our stuff. Seriously.

The secret sauce really is the rocket science

Rocket Fuel provides a simple yet sophisticated way to manage, run and evaluate successful online campaigns. Our technology focuses on finding desirable audience characteristics rather than mere impressions. Through rapid automated testing and user-level targeting we find the individuals that matter. With “progressive optimization” we turn quantitative learning into actionable improvements for each campaign. The result is higher performance – against your metrics – over the run of the campaign.

But, we’re not just about automation. We also provide a dedicated analyst, real-time reporting and feedback so you'll never be caught off guard or unable to respond to clients or senior management with timely information.

Our core technical philosophy is simple

An ad server makes billions of decisions per day, tens of thousands of decisions per second, about which ad to serve for a given online impression.

  1. To make the best decisions, you need relevant up-to-date information and a solid decision-making framework or algorithm.
  2. To make the best decisions economically, you need rational computation – an architecture that understands the tradeoff between computational cost and net revenue.

It's about relevant information
There should be no contextual ad networks or behavioral ad networks. It's like debating whether you should have eyes or ears. If you can have both, it's a win and there's no point debating.

There should just be smart ad networks that use as much relevant information as they can to pick the best impressions for a given campaign. An ad server should just use whatever data is found to be valuable in selecting the best ad for an impression, or the best impression for an ad.

Eliminate the dumb mistakes and make smarter decisions
NASA's Mars rover is autonomous in the sense that it doesn't need to be driven from Earth every step of the way. For example, it can scan terrain and adaptively plan a route to get to a point to take a rock sample.

Shouldn't ad serving be autonomous as well, as long as it fits within advertiser and publisher constraints? For example, even if a campaign isn't geo-targeted, but the statistics show higher response rates in certain regions, the ad server should adapt its serving strategy and deliver relatively more impressions in higher-performing regions to boost the efficiency of the campaign.

Just as a Segway Personal Transporter can stay upright regardless of how hard a person leans forward or backward, an ad network should be able to not "fall over" when the environment changes, such as when a new publisher or advertiser comes or goes.

Rational computation means using common sense
Most businesses take efforts to treat their best customers a little better. Frequent-fliers get the seats with extra legroom. Restaurant regulars get the best tables, etc. But to most ad servers, every ad serving decision is the same. One minute the ad server might be looking for an ad to fill an impression on the worst-performing publisher in the network, with a user that hasn't clicked an ad all year. The next minute the ad server could be handling an impression on a premium publisher, with a user who has clicked ads for digital cameras and high-yield savings accounts recently.

Should the ad server really spend the same amount of time and money thinking about the right ad to serve in both of these contexts? In the new world of elastic computing, systems no longer need to be built to run on a fixed computational infrastructure, or at least an infrastructure that goes months between changes. If you can fire up a server, use it for the next 30 minutes of peak traffic, then shut it down, it leads to a new architectural approach.

back to top^

 
guy gesturing with hands in meeting