Essential SOPs, Part 4: Duplicates, the Mess That Comes in the Front Door
Duplicates aren't a one-time cleanup. They're a steady drip from every form, import, and integration that touches your CRM. Here's the routine that keeps them in check.
This is the fourth post in our Essential SOPs for Salesforce Admins series, based on my session at Texas Dreamin'. Part 1 made the case for routine maintenance, Part 2 covered user management, and Part 3 covered backup and recovery. This week we stay on the data and deal with what walks in the front door.
Quick question: are Lizz Hellinga and Elizabeth Hellinga the same person?
What if one has a personal Gmail and the other has a work email? What if one is tied to a customer account and the oher to a partner account? I'd call those duplicates. Would you? More importantly, would your sales team, your marketing team, and your finance team all give the same answer? That's the real problem with duplicates. Everyone agrees they're bad, and almost nobody has agreed on what one actually is.
Start by Defining "Duplicate" (Yes, Really)
Before you touch a matching rule, get your stakeholders to agree on what counts. Same email? Same name and company? Same phone number? What about the same person at two different companies, or two people sharing an info@ inbox?
As with defining "customer," it's easier to give people something to react to than to ask an open-ended question. Bring a draft: "A contact is a duplicate if the email matches, or if the first name, last name, and account all match." Let them argue with it. The debate is the work, and it's the only way your matching rules end up reflecting how your business actually sees people.
Know Your Front Doors
Most orgs don't have a duplicate problem. They have an intake problem. Duplicates are a symptom of how data enters the system, so map every way it gets in:
Web forms that create a new lead every time someone downloads a whitepaper
List imports from events, purchased lists, or that spreadsheet a rep has been "meaning to load"
Integrations with your marketing automation, enrichment, or support tools, each with its own idea of what a match is
Auto-capture tools that sync every email contact into the CRM. (My opinion, for the record: you rarely need every email address in Salesforce. Enable your reps to log what matters instead.)
Manual entry by a rep who searched for "Bob" when the record says "Robert"
Each door needs its own answer. A matching rule that works for manual entry may do nothing for an inegration that bypasses duplicate rules entirely.
The Monthly Routine
Duplicates accumulate quietly, so check on them on a schedule rather than when someone complains.
Review what your duplicate rules caught. If your rules are set to allow and report, build a report on duplicate record sets and work through it. If nobody ever looks at that report, the rule is only half doing its job.
Merge in small, regular batches. A little every month beats a massive cleanup project every two years. Prioritize accounts and contacts tied to open opportunities or active customers first.
Spot the patterns. If the same source keeps creating the same kind of duplicate, that's not a cleanup task. That's a front door that needs fixing.
The Quarterly Tune-Up
Once a quarter, step back and make sure your rules still match reality.
Has a new door opened? New integration, new form, new event platform? Confirm your matching logic covers it.
Alert or block? Blocking stops duplicates cold but frustrates users when the rules are too broad. Alerting is friendlier but relies on people paying attention. Review which rules should move from one to the other based on what you've learned.
Check with the people entering data. If reps are working around a rule (adding a period to a name, anyone?), the rule needs adjusting, and a quick conversation will tell you more than any report.
Where to Start
If duplicates are the pillar you're picking up first, here's the single move: pull a list of every way data gets into your CRM and write it down. Just the list. Most teams discover at least one front door they'd forgotten about, and that door is usually where the mess is coming from.
What's Next
Next post, we wrap up our three posts on data with rotational field audits: how to find the fields nobody uses, the ones everybody uses differently, and why a little regular pruning keeps your org (and your AI) grounded.
