Forecasting-Ways to Test Reasonableness and Success For New Products

I was working with a client last week and we were reviewing new product forecasts with their major accounts. In the process of this review I was testing for reasonableness of the weekly forecast based upon the intial roll out quantity and the annuallized weekly forecast. Typically a ratio of between 4 and 8 should be the answer received. If the answer is below four than the answer needs to be checked against other items being sold to this retailer to see what their turnover rates are, if they are all above four than the new item forecast shousl be questioned as to why and will there be a potential loser in this product. If the answer is over eight then the same test should be performed and depending on the answer the forecast should be reviewed. If for some reason the answer is over 20 turns as it was with one item we reviewed then other questions must be asked. They must be asked because turns in excess of 20 will mean a great many empty shelves and lost sales at the retail level unless there are multiple facings of the product and therefore the forecast maybe accurate but the number of turns needs to be reduced.
As a double check of the forecast and assuming your comapny can receive P-O-S data there should be a correlation between the retail sales after a few weeks and the "final" level reached for the product. Determining this correlation and quickly being able to determine the "final" sales figures will help in ordering more product sooner or the cancellation or delay of product already ordered. Having these models will go along way to having better inventory turns and less closeouts.