Hello Crystal,
Here are some answers to your questions.
Deleting leads in mkto:
Step 1: Create smart lists to isolate the records you want to delete (e.g. "Email Address is Empty")--i recommend keeping these together in a folder in a data management program. CAUTION: Be sure you DO NOT delete anyone from your database who is unsubscribed--otherwise, if they re enter your database via web form or list upload, they won't be set as unsubscribed and you may email them, thus violating can-spam laws.
Step 2: Hit "Select All" in the leads view of your smart list and select "Delete Lead" (deep breaths :))--this will bring up a window that asks you if you ALSO want to delete them from the CRM (Delete from CRM = True/False). Set it to FALSE and hit delete. This will remove the leads from your MARKETO database, but not SFDC. The records in SFDC that you just deleted in Marketo will NOT sync back over into Marketo unless they are manually pushed into Marketo from Salesforce.
Duplicate management
Use the Marketo system smart list Possible Duplicates to view your duplicate records. Sort them by Email address (that's the field Marketo uses to define duplicate records). Rather than deleting the duplicates, it's better to merge them--select the records you'd like to merge, and hit Merge leads. This will bring you to a window that lets you pick what field values you keep and which you discard.
General best practices
1) Well begun is half done: Instate a data management process immmediately for all means of record creation in both Marketo and Salesforce. Define the key fields you need for audience segmentation, routing/ownership assignment, reporting and contacting and set up processes to protect and populate them. This means working with your sfdc admin to make key fields required and picklisted to keep data standardized, ensuring that all Marketo forms built require the appropriate fields and use pick lists (NOT open text fields) for your internal processes or trigger data management processes that do populate those key areas, AND making sure that anybody uploading lists into Marketo or Salesforce is doing so with the correct fields populated correctly.
2) Automate: if there are areas of data that are difficult to keep clean, set up smart campaigns to standardize or populate them. Example: Smart list: if Country is US, U.S. USA, U.S.A, (continue with varations on United states), then (FLOW) change Country value to "United States." Make these smart campaigns a part of a larger Data Management program.
3) Watch: Databases tend towards entropy, so be sure to regularly review your duplicates, white space, and data standardization. I like to create a set of smart lists geared to my key dirty areas, as well as a report to see the higher level. The report is a Lead Performance report, and under set up I group the records by either Original Source Type, Source, Lead owner or another meaningful and consistently populated field. Then, i add my dirty smart lists in as custom columns so that wen i hit "report" i can see which Source is responsible for creating the most dirty leads each week.
Hope this helps!
Elizabeth