Records Identified for Merge
Under the selected Phone + Postal Code rule
HUBSPOT CRM CASE STUDY
HubSpot Company Deduplication & Merge Automation Case Study
A controlled deduplication workflow combining full-dataset analysis, weighted primary-record selection, manual-review safeguards and scripted HubSpot API merging.
- Client
- Tim
- Company
- Not Publicly Named
- Project Period
- May to June 2026
- Platform
- HubSpot CRM
- Project Type
- CRM Data Quality + Automation
Executive Summary
A Clear Decision Behind Every Merge
Tim needed a duplicate-company cleanup that could work at scale without losing the records the business depended on. Tech Striker built a rule-led process: analyze the full dataset, choose the primary record, protect critical data, review exceptions and execute approved merges in controlled batches.
Validation Rules Tested
Compared against the actual company dataset
Records Routed Through Manual Review
A dedicated review workload, not an automatic merge count
Eligible Merges Pending
Under agreed criteria at project completion
The Starting Point
A Large Portal. A Company-Data Problem.
Duplicate company records had accumulated across the CRM, affecting ownership, reporting and confidence in customer data. The task was to identify genuine duplicates and decide which record should survive, not simply combine anything that looked similar.
185.2K
Contact Records at Assessment
Approximate portal scale. The cleanup in this case study concerned company records.
At the time of the project, the portal’s native duplicate-management view exposed approximately 10,000 potential duplicates, so the full company dataset was exported for analysis.
The Challenge
Matching Was Only the First Decision
Hidden Scale
The native duplicate view exposed only part of the problem during the project. Full-dataset analysis was needed.
Inconsistent Data
Empty fields, postal-code variations and non-standard domains could turn simple comparisons into false matches.
Irreversible Merges
Records could not simply be unmerged. Choosing the wrong pair could permanently combine separate companies.
Look-Alike Companies
Branches, franchises and regional entities could share contact details while remaining legitimate separate records.
Large Volume
More than 14,000 records required merge consideration. One-by-one execution was not practical.
Limited Client Sign-Off
The client could not inspect every edge case. Clear protection rules and a structured review process were essential.
See the Full Dataset
Analyze Beyond the Visible Queue
Exporting the company data made it possible to evaluate every record, compare matching criteria and build a review process around the client’s business rules. Empty and null values were excluded from comparisons, postal codes were normalized and domains were standardized.
01Export
Full company dataset
02Normalize
Usable, consistent values
03Compare
Tested matching rules
04Review
Explicit merge decisions
Match-Rule Testing
Choose the Rule That Fits the Data
| Validation Rule | Project Evaluation |
|---|---|
| Company Name + Phone | A name-and-phone comparison |
| Company Name + Domain + Phone + Postal Code | A multi-field validation rule |
| Company Name + Domain + Postal Code | An alternative without the phone field |
| Phone + Postal Code | 14,208 records identified for mergeSelected with client approval |
Primary Record Selection
Decide Which Record Should Survive
A weighted scoring model helped select the primary record within each duplicate group. The model considered completeness, ownership and activity alongside record age. Business-protection rules could override an otherwise straightforward match.
Record Age
Consider how long the record has existed.
Phone Availability
Check whether a phone value is available.
Domain Availability
Account for populated domain information.
Company Owner
Consider existing ownership of the record.
Recent Activity
Review activity associated with the company.
Industry
Include available industry information.
Protect Business-Critical Data
Keep the Records the Business Relies On
Active Customers
Protect customer and current/active-customer records from merge or deletion.
Partner-Program Data
Prioritize records carrying business-critical partner-program values.
Linked Service Records
Retain records with linked service activity under the client's protection rules.
Open Tasks
Consider outstanding work before deciding which record survives.
Multi-Location Companies
Handle branches and separate locations without assuming they are duplicates.
Overlapping Conditions
Route conflicting protection conditions to manual review.
Human Review
Automation Had a Deliberate Boundary
Duplicate groups were classified before execution. Ambiguous matches, branches and overlapping protection conditions stayed in review. A team of four shared the manual-review workload, with an approval column recording Approved to Merge or Do Not Merge.
14,111Records Routed Through Manual Review
SAFE AUTO MERGEMANUAL REVIEWREVIEWBRANCH REVIEWREVIEW BEFORE MERGE
Scripted Merge Automation
From Approved Decisions to Controlled Execution
Merging around 14,000 records one by one would have been slow and error-prone. The approved Google Apps Script tool called the HubSpot API, with dry runs, result logging and validation between batches. When the client revised the criteria, the merge conditions were rebuilt to reflect the new direction.
Canonical ID Chaining
Track the surviving HubSpot ID when a previously merged record is referenced later.
Exponential Backoff
Back off and retry when API rate-limit constraints are encountered.
Dry-Run Mode
Preview intended merge actions before executing them in HubSpot.
In-Sheet Result Logging
Record merge results, response codes and retry counts alongside the review data.
Custom Merge Menu
Run a batch or dry run, and start or stop a scheduled auto-run directly from Google Sheets.
Controlled Batching
Process smaller batches and validate each output before continuing.
The Workflow
A Repeatable Route From Export to Validation
- 01Full Company Export
- 02Normalize Data
- 03Apply Match Rule
- 04Group Duplicates
- 05Score Primary Record
- 06Apply Protection Rules
- 07Safe Auto Merge or Manual Review
- 08Dry Run
- 09HubSpot API Merge
- 10Log Result
- 11Validate Batch
- 12Continue
Execution and Traceability
Run, Inspect and Validate Each Batch
The execution capture includes both failed and merged attempts, with HTTP response codes and retry counts. It shows the logging process, not an all-success completion screen.
Technologies Used
Connected Through the Existing Toolset
- HubSpot CRM
- HubSpot API
- Google Apps Script
- Google Sheets
- CRM Data Export
- Operations Hub Trial
The Outcome
Resolved Under the Agreed Criteria
All company records eligible under the client-approved rules were processed. Duplicate groups resolved to selected primary records, while protected customer and business-critical records were retained. The final sheets preserved the primary/secondary mapping and merge results.
0
Eligible Merges Pending
At completion, under the agreed criteria.
- Selected primary records retained
- Protected records preserved
- Merge outcomes logged
- Final record mapping delivered
- Repeatable process documented
- Client handover completed
Project Summary
Cleaner Company Data, With a Process to Continue
The result was a controlled company-deduplication workflow, supported by review sheets, client-specific rules and API automation. The handover gave the next stakeholder a documented way to continue the work. Contact-level deduplication was identified as a possible follow-on phase, not part of this completed company cleanup.
Explore our HubSpot CRM Solutions, HubSpot Consulting and CRM & RevOps services, or view more Case Studies.
