The main thrust of the project includes :
The possibility of integrating and combining different kinds of histories(audiences for Xiti-type sites, ad impressions via Ad Servers) site by site, section by section.
A user interface allowing integration of external data, which makes it easier to correlate histories (holiday calendar, current events, the site’s editorial planning...).
The possibility for the user to underweight or overweight certain periods in line with his/her knowledge of the business.
A robust algorithm that forecasts audiencesand thus the “stock” of available ad impressions, taking account of the interrelation between the different sections of each site.
Predictive accounting of Web-specific buzzes that amplify audience variations.
A self-learning function that measures forecasting errors compared to actuals.
Integration of impression volumes already sold (optional or firm) and which will decrease available stock.
Table or graph display of audience forecastsand available stock by site and by section over a given period.
Possible call from market software (Ad Servers, TBS’s WebResa...), all natively equipped with a Web user interface. This interface should be based on Web 2.0 technology, allowing throughput of input parameters (site, section, period…) and output parameters (total stock available...).