Ah, there is so much to this question!!! Ok, let me see if I can get most of it out:
Tools: There is a manual in the Help section under Administration that covers Contract Import. There is also a .zip file with the 4 files you need for the load: contract data, documents, team members and then the load parameters. This last is available either from your deployment project manager or technical support. It might be available on-line as well.
Best practices:
The template you use in your load will have the tasks (like for amendments), documents and team members that will typically last for the duration of the contract's life. If you want tasks, particularly approvals associated with amendments, you should make sure that the legacy load template has your process embedded.
Unlike the supplier load, which can be run multiple times adding and correcting data, the legacy load can only be run once for each contract. Make sure you assemble all of the data and documents for the workspace before you do the load.
Review what data you want to load. There are only certain Ariba required fields. Are you configuring required custom fields? Do these need to be in the load template? If not, what happens when someone amends the contract? Will they need to fill in these fields?
It is recommended that you load your most active contracts first, then go back and build your repository of older contracts and, probably not even bother with contracts that have been closed for more than a couple of years. Some customers load the most important contracts in draft status so their content can be reviewed before publishing.
Decide on an amendment process. "Amendment" can mean: 1. editing attributes; 2. changes to language content; 3. changes to the value of the contract (either change order or by renewing); 4. Change orders or SOWs. Ariba can support several approaches, but consistency is preferred. Then when you load your contracts, you should be consistent with your chosen approach.
So, I hope all of this helps.