Customer Duplicate Prevention

Overview

Stopping duplicate records before they start.

In the high-velocity sales environment of car dealerships, speed is everything — but that urgency has a cost. Car dealership salespeople were routinely creating new customer records without checking whether one already existed. The result: duplicate records piling up across the system, each one quietly fragmenting workflows, muddying reporting data, and creating real disputes over commission attribution.

The fix wasn't about discipline or training. It was about design. Customer Duplicate Prevention set out to build a check directly into the customer creation flow — friction in exactly the right place, at exactly the right moment, so salespeople encounter existing records before it's too late to matter.

The problem

One bad habit. Three real consequences.

Root cause

Salespeople were creating new customer records without checking if one already existed.

The system made it easy to start fresh and offered no prompt to pause. In a fast-moving sales environment, that path of least resistance became the default — and duplicates accumulated silently across the database.

Effect 01

Scattered workflows

Customer history, notes, and activity were split across multiple records. Salespeople working a deal might be missing context living on a duplicate they didn't know existed.

Effect 02

Commission disputes

When the same customer appeared under two records, it created genuine ambiguity about who owned the relationship — and who should be credited when a deal closed.

Effect 03

Unreliable reporting

Duplicate records corrupted the data layer that reporting depended on. Customer counts, pipeline data, and trend analysis all became harder to trust.

My process

Understanding the cause before designing the fix.

Research

Understanding why duplicates were being created and identifying the right moments to intervene.

Investigated root causes of duplicate creation behavior
Mapped the full customer creation flow to find intervention points
Evaluated which friction points would actually be seen and acted on

Ideation & feasibility

Explored design directions collaboratively with engineering to stay grounded in what was buildable.

Generated multiple approaches to surfacing existing records
Worked closely with engineering on technical constraints
Narrowed options based on feasibility and user impact

Validation

Pressure-tested the design direction with stakeholders before committing to full execution.

Confirmed the intervention point with research findings
Validated the proposed flow logic with key stakeholders
Refined messaging and presentation of duplicate matches

Testing

Usability testing with real salespeople to confirm the friction worked — without breaking their flow.

Tested recognition of the duplicate check during task completion
Observed how users responded to matched records
Refined based on moments of confusion or hesitation

Key decision

Where do you add friction to a salesperson's flow?

The central design challenge wasn't whether to add friction — it was precisely where. Add it too early and you slow down legitimate new customer creation. Add it too late and the duplicate is already committed. Get the moment wrong and salespeople learn to click past it without reading it.

The question

At what point in the creation flow does a salesperson have enough context — and enough motivation — to actually check?

Three candidate moments emerged from research. Each had genuine merits, and each carried tradeoffs around visibility, disruption, and the likelihood of the check being taken seriously.

Option 01

At the entry point

Prompt the salesperson the moment they click "Create New Customer" — before they've entered any information. Low disruption to the creation flow, but also low information: they haven't typed a name yet, so a match can't be shown.

Considered

Option 02 · Chosen

After name and contact info

Trigger the duplicate check once a name and at least one piece of contact information — email, phone, or address — has been entered. This combination gives the system enough signal to surface meaningful matches, while still catching the duplicate before the full record is committed.

Selected approach

Option 03

At form submission

Surface the duplicate warning only when the salesperson tries to save the record. Maximum information available, but maximum sunk cost too — by this point, they've filled out the whole form and are motivated to push through anyway.

Considered

The insight

The right moment isn't the earliest possible — it's the earliest moment where a real match can be shown. A name alone isn't enough; but a name paired with an email, phone number, or address gives the system enough to be confident. At that point, if the customer already exists, the case for creating a duplicate collapses.

Design work

From research to reality.

Customer creation

Simplified new customer form — streamlined fields to reduce friction and time to complete

Duplicate detection

Inline match notice surfaces a potential duplicate mid-flow — letting the salesperson act without leaving the form

Outcome

Shipped. Expecting duplicates to drop.

✓

New customer creation flow with built-in duplicate detection shipped

↓

Duplicate record creation — expected reduction once fully deployed

Downstream problems addressed — workflows, commissions, and reporting data

The redesigned customer creation flow shipped with a built-in check that surfaces matching records the moment a name is entered — before a duplicate can be committed. Salespeople now encounter existing customers at the exact point where redirecting them is still frictionless.

In progress

The project is not yet fully deployed across all environments. Once rollout is complete, the expectation is a meaningful reduction in the rate of duplicate record creation — removing the source of scattered workflows, contested commissions, and degraded reporting data.