To get the best results, you typically run batch jobs in a particular order, as follows:
Run a cleanse request and review the results.
Run a match request.
If matches are found and you determine that there are duplicate records in your data, run an auto merge request.
A batch process must finish running before you run the next batch process.
Warning: Records must be cleansed before being matched
or auto merged. The cleanse process creates the key on which matches are evaluated.
If you run a Match Only process on records that have not been cleansed, no matches
will be found.
Note: Standard batch requests are designed to process subsets of your data. Processing more than 10,000 records in a batch request is not recommended.