Here are some guidelines we've found to be successful on this topic:
1. If the contract is still active, it belongs in the repository. Obviously, newer contracts are more likely to have more current information. Older contracts may need a lot of research: has the supplier been acquired? Do we have a contract on file with both parties' signatures? Sometimes you need to triage these contracts: Here's the list that the document's missing, here's the list where there are no signatures, here's the list where the master is missing, etc. (nothing's worse than having an amendment to a contract where the master is missing!) However, over time, your repository is best when all the active contracts are in it.
a. Possible exception: small dollar amount, low risk conracts due to expire soon. Then I would look at the process under 3.
2. If the contract is expired or cancelled and the statutory retention period is expired (usually 7 years), why bother? Clean out your file drawer, destroy the paper and don't add to your repository.
3. For those that are expired or cancelled but still within the retention period, it seems to me that these would be very low priority, but it might be worthwhile getting them on the system so that reminders could be set up for retention period expiration. In this case, if you would choose to do this (because you had nothing better to do?), I would send the paper off site and only enter the minimum amount of information possible.
So, for number 1, which ones first? There are several approaches, but certainly getting all high risk, high dollar contracts from strategic suppliers needs to be at the top of the list. After that, any method of organization seems to be satisfactory: all of 1 type, all for 1 department, by commodity, by region, etc. The key here is to have a process, that legacy loads are processed over time and that there's a status to indicate how much is done and how much is left.