HubSpot association evidence lab
Model the documented association path, distinguish an expected skip from an ambiguous match, and check why one Typeform or integration submission appears to create two companies before changing a setting.
Start with the symptom
Each points to a different evidence path. Model only categorical facts below before changing settings or production records.
A multi-object import does not trigger the automatic company-creation path for a contact without an imported association. Verify the import type, both objects' identifiers, and the explicit association columns before changing a portal-wide setting.
Model the import path →Two writers are only a hypothesis. Check the published Typeform company mapping and comparison fields, then compare each company's Record Source, source detail, association history, and timestamp. Timing alone does not prove which path wrote either record.
Trace the two-writer risk →If multiple companies share one Company domain name, HubSpot's automatic association chooses only one and does not let the user select which through that setting. Verify the match count, primary company, and association history before reassociating a record.
Check an ambiguous match →Typeform → HubSpot evidence path
If a Typeform mapping writes company data while HubSpot's separate domain-based company creation path is active, two company-writing paths may overlap. That is a diagnostic hypothesis—not proof that Typeform used an API path or caused either record.
Stop before merging, deleting, or disabling anything. HubSpot documents that API-created companies, including installed third-party sync apps, are not automatically deduplicated by Company domain name; the actual portal evidence must first establish whether that boundary applies.
Before an irreversible cleanup
A company merge is not a reversible cleanup step. HubSpot does not provide an unmerge action. The resulting record receives a new Record ID; activities and associations combine; primary-record property values are generally prioritized; and the secondary company's domain becomes an additional domain. Workflow enrollment and Salesforce sync behavior can also depend on the account's subscription, settings, and integration version.
Qualify the impact before anyone clicks Merge.
If any answer is unknown, stop before Merge. The $1,500 Pipeline Truth Audit is read-only, 100% upfront, and delivered in five business days within the published scope. It maps the relevant company-writing and downstream evidence; it does not merge, delete, or edit records.
The generic fit-check draft asks for the HubSpot plan or tier, approximate active-record and system counts, and the pipeline answer you trust least.
I will not send credentials, CRM records, record or portal IDs, exports, screenshots containing personal data, API payloads, or secrets by email.
See if the $1,500 read-only audit fits this merge →Prevent the next duplicate
Choose one future company writer before changing a setting. Preserve the association and qualified-freemail behavior the business still needs, then isolate every other confirmed company-creation path in fictional tests.
The planner returns DO_NOT_CHANGE, CANDIDATE_PATH, or PREVENTION_READY_FOR_TEST—never “fixed.” It checks writer evidence, the intended creation owner, Typeform comparison fields, import identifiers, preserved behavior, exact new-domain and freemail expected counts, four fictional cases, success evidence, and rollback conditions.
Record Source, source detail, association history, and same-timezone timestamps must establish the writer set. A mapped field, API payload, or close timestamp alone is not prevention proof.
Free local browser tool
One-click generic hypothesis
Loads fixed categories for automatic association, multiple same-domain companies, and a confirmed API or installed-app company writer. It does not claim that Typeform—or any named integration—used that path in your portal.
Use categories only. Never paste names, email addresses, domains, company names, contact data, record or portal IDs, screenshots, exports, integration payloads, passwords, MFA codes, API keys, or tokens.
Example only · generate before using the output
Generated output
Example output only. Generate the check or load the generic preset before exporting or requesting a diagnosis-specific fit check.
Need help tracing this association path?
Request a no-data fit check for the $1,500 Pipeline Truth Audit using only the generated result class and risk-flag IDs. The draft asks for approximate scope—not CRM records.
Documented risk flags
This does not prove the portal is clean. Complete the ordered verification and preserve unknowns.
Ordered verification
Read Record Source and the creation timestamp before applying a domain-matching explanation. Preserve only the generic source category and source detail in diagnostic notes.
Check whether Create and associate companies with contacts is enabled, whether existing contacts were included, and whether the relevant domain is excluded. If an exclusion was recently removed, follow HubSpot's documented off-and-on reset before expecting future association.
Compare the categorical contact email or Website URL domain with Company domain name values. Record whether zero, one, or multiple companies match and whether the value is a subdomain.
Confirm whether the contact already has a company, which company is primary, when each association was created, and the association change source.
Confirm no form, multi-object import, API writer, or installed sync app created the record. For company creation, compare the generic Record Source with HubSpot's documented domain-deduplication boundaries.
The browser-local prediction is ASSOCIATE_EXISTING. Do not change, merge, delete, or reassociate records merely to make the portal match the prediction; document any mismatch for read-only investigation.
Five-case association plan
| Test | Scenario and setup | Expected | Evidence | Stop condition | Result |
|---|---|---|---|---|---|
| BAS-01Baseline | Capture the unchanged association state before any testWithout mutating a record, document the categorical creation path, automatic-association setting and scope, excluded-domain state, matching-company count, current associations, primary company, and relevant Record Sources. | The starting relationship, writer candidates, settings, timestamps, and unknowns are explicit. | Sanitized categorical notes, generic Record Sources, association history, and same-timezone timestamps | Stop if the creation path, in-scope records, matching-domain count, or primary-company state cannot be established without exposing CRM data. | Not run |
| POS-01Positive | One eligible business domain matches one existing companyOnly in a separately approved sandbox or with fictional records, use one new contact category, one exact matching company domain, no exclusion, and the documented automatic-association setting. | The fictional contact is associated with the one matching company; association history supplies the source and timestamp. | Setting snapshot, one-company match, before/after association state, generic source, and timestamp | Stop on an unexpected company creation, second match, external send, production record, or unrelated property change. | Not run |
| NEG-01Negative | A documented blocker prevents automatic associationUse only a separately approved fictional control representing one blocker: disabled automation, an excluded domain, a legacy contact outside prospective scope, or freemail without Website URL. | No automatic company creation or association occurs through the blocked path. | Selected blocker, setting snapshot, before/after association state, generic source, and timestamp | Stop if the blocker is not isolated, a form exception applies, a production record enters scope, or another writer runs. | Not run |
| EXC-01Exception | A subdomain and root domain are evaluated separatelyUsing sanitized configuration evidence, verify whether the contact domain is a subdomain and whether a company already exists for that exact domain. | The subdomain is not silently represented as the root company domain; any separately created company remains visible in the evidence. | Root-or-subdomain category, exact-match count, company creation source category, and association history | Stop if the exception requires credentials, production mutation, bulk import, deletion, merge, or an unisolated external writer. | Not run |
| REC-01Recovery | A separately approved controlled test produces an unexpected associationStop the confirmed in-scope test writer, restore its documented baseline, preserve association history, and repeat only on a new fictional record after the mismatch is explained. | The configuration returns to baseline and the new fictional record follows the documented expectation. The original browser-local prediction remains recorded as ASSOCIATE_EXISTING, not rewritten as proof. | Baseline, revision or configuration history, affected fictional labels, association history, recovery result, and residual-risk note | Keep the writer paused and escalate if scope, permissions, primary-company state, or recovery evidence is uncertain. | Not run |
Planning artifact only. Use fictional records, disable external sends, preserve a documented baseline and recovery path, and keep personal data and secrets out of the artifact. A generated expectation is not execution evidence. Do not merge, delete, create, or reassociate production records from this output.
Created with Cynosure Company Association Checker.
Recurrence-prevention planner
Start from a generated association result. Add only categorical evidence about the writer set, intended creation owner, published matching contract, import guard, behavior the business must preserve, and exact company and association counts for the two variable test cases.
Loads fixed generic risk categories, generates the prerequisite check, and keeps this planner in view.
Need a read-only association audit?
The read-only audit inventories association settings, excluded domains, primary-company state, imports, Salesforce, and company-writing integrations. Request a fit check with the generic result class and risk-flag IDs only, then add approximate plan, record-count, and writer-count scope—never CRM records, domains, names, IDs, exports, payloads, or screenshots containing personal data.
See the exact audit scope and limits →Product evidence
Product behavior, labels, permissions, availability, and integration rules can change. Verify the linked documentation and the client's actual portal before acting. Cynosure is independent and is not affiliated with or endorsed by HubSpot or Typeform.