@ Astounding * The timeperiod variable, week or month, is when the subscribtion is started. But since change2 describes the change between week one, and week two, it's really not nessesary. * It dosen't get added to the title, that's not really nessesary. * A active subscribtion is any cell that isen't blank, or has a "CLOSED from X" mark. * Optimally, something like the result that chang_y_chung@hotmail.com gets, but for all combinations and running from month to month (alternativliv week to week). I'll need to look at the Double-DoW and how the reshapeing to long looks. But the output looks promising. Basicly i need to find out how the flow in subscriptions is. So if X amount of people open to a 10 week campaign, how many of them open to a 13 week normal subscription anytime later.Saying that, having a fixed number of weeks that i need to look ahead (like from change4 to change9) is not a satisfactory solution. I might have a bunch of TRIAL_4W that opens in change2, then distributed over mabye 6-7 weeks, they will open to other subscribtions. I need to figure out a way to calculate the average number of weeks they are closed (blank) before they open to a new subscribtion, and then what subscribtion they open to. And then again, how many weeks they, on average, hold that subscribtion. @ chang_y_chung - Thanks a bunch! I think i understand the transformation to long. But i'll have to look at the Double-DoW method. The output you got is near perfect. But i still need to have a "dynamic" timerange, so that is isen't locked on to 4 weeks necessarily. @ MikeZdeb - Thanks for the input! Im not really sure that i fully understand your output? Im not really interested in how many subscribtions close, but rather how many of the ones who opened, are still active X weeks later.. I hope that my comments above clears up my point abit more. @ Ksharp - Just saw your post. I need to read some more about the _a and dim(_a)-8, as i'm not really sure what that does. Again, as commented in this post, i would like to have a "dynamic" timelimit, so i don't have to decide on a fixed number of weeks, but rather "Who opened in week1, how are they distributed in the weeks that i have info about now (weeks 1 to 16), how long were they blank before they opened, and what did they open to? Thanks alot for the input, it's really helpful. Finding help to SAS datastep on google, is a bit difficult when im not even sure where to start. T
... View more