Deduping in Salesforce: The complete how-to

Don’t be duped by duplicate records. Deduping seems like one of those routine system maintenance issues, but it’s actually very tricky and there's no undo.

twins dual boot

Every enterprise system has the risk of data quality problems, but customer relationship management (CRM) and marketing automation systems are incredibly vulnerable to incomplete records, bad data, and duplicate records.  Data comes from too many flaky sources, and there aren’t enough incentives for organizations and people to really follow standards.  Fortunately, a number of services and tools have arisen to automagically improve data quality (in particular, Salesforce’s  When it comes to deduping, though, even the best  tools out there  need to be used with careful attention to detail and the prep work that can take days.  We’ll show you how to avoid the main pitfalls here.

The big deal about dupes

Duplicate records are dangerous to system credibility because users can’t find the updates that they’ve made (they’re looking at the wrong copy of the record) and each of the dupes represents an incomplete record with multiple data quality issues. The longer a dupe exists, the more the data quality issues metastasize, making rectification ever more costly.

[ Find out how CRM buyers can negotiate the best deal, when Salesforce certifications matter & what to look for, and how to screen a Salesforce consultant. | Get the latest insights by signing up for our CIO newsletter. ]

Duplicate records are indicators of process problems, either in user training or because there something that makes the original record invisible or unusable for that user in a specific stage of the process.  Dupes are also key indicators of system problems (Is the weekly upload not using fuzzy matching?  Is this integration not properly checking for existing records?  Is that trigger using the right logic to check for and resolve dupes before doing a record update?)

I hope I’ve scared you about this issue, because fixing it can require a surprising amount of time.

Prevention is the best cure

The starting and ending point of any deduping project is this:  finding defects in processes and systems to prevent the creation of dupes in the first place.  Step one is to identify which process and system elements are creating which flavor of dupe (and yes, you will have to do pattern recognition on the particular types of dupe you’re experiencing).  Once you’ve done this, you need to have a brief checklist of situations to flag, with solution strategies for each one.  For example:

  • What do you do when two people share the same email address?
  • What do you do when a person has 10 email addresses?
  • What do you do when one company you do business with acquires another one that’s a prospect?
  • What do you do with international divisions of conglomerates, some handled by your sales team and others handled by a channel partner?

These checklists of business rules need to go to your relevant developers and consultants so they know what to look for and build in handlers in their code.

To continue reading this article register now

Security vs. innovation: IT's trickiest balancing act