In the changing audience behaviour situation, and when commercial television in Finland was moving over to a channel of its own, it was necessary to create a system, better than the mathematic models, that could forecast coming audience behaviour patterns. A particularly central issue was to make the forecasting system more accurate because, in Finland, the basis of media sales is the contact guarantee, offered to each client for each campaign. Further, the business usage of increasing the target group approaches also presented special requirements for the redesign of the forecasting system. For television companies, forecasting the coming campaigns, as accurately as possible, is economically a central issue. That is why the idea was to create a continually updated database, based on real observed data, and a forecasting system connected with it, and utilizing past audience behaviours, that was being offered to advertising agencies for both nationwide and regional television campaign planning purposes. The information system is operated as concentrated in the television company and it can be used in the agencies via their data communication networks. Using this system, the quality level of the planning services received by the advertiser in all of the offices is guaranteed, as well as the campaign planning to be carried out according to the latest observed data information. Compared with traditional formula based models for estimating reach and frequency, the described new database enables calculation of cumulative reach and frequency distribution without any mathematical formulas or modelling. As it utilizes the raw data from an individual viewing file, it automatically incorporates individual duplication of viewing and programme loyalty of different programmes in the campaign schedule. All standard and user defined breakdowns are available. Also the individual database enables automatic schedule optimisation procedures.