Company-level visitor identification tells you which organization is associated with eligible website activity. Person-level identification returns an individual record associated with eligible activity. The difference is what the data supports your team doing next.
I am George, founder of Leadpipe. We build person-level identification, but company-level data can be the right choice for account research and prioritization. Choose the identity level your workflow requires, then inspect the record and its provenance. A longer contact list is not automatically a better identification result.
For company-level input, Leadpipe also offers an IP to Company API: look up stored company information from IPv4 or IPv6 without treating the result as a named visitor. Use the firmographic data guide to define the account fields your workflow needs.
Three outputs that comparison tables often mix together
| Output | What it represents | What it does not establish |
|---|---|---|
| Matched company | An organization associated with the visit | Which employee was browsing |
| Matched person | An individual associated with eligible activity | A purchase decision or permission to contact |
| Suggested contact | Someone relevant at the matched company | That the suggested person visited |
A company-level provider can have an excellent contact database. Those contacts can help a rep research a buying group. They must still be described as contacts at an account rather than verified visitors unless the provider supports that association.
A person-level record can contain missing or stale fields. Evaluate the identity link and the enrichment separately. This is why a single “data accuracy” percentage is rarely enough to make a buying decision.
Visitor identification versus contact enrichment
Identification starts with activity; enrichment starts with a record or identifier. If a company visits and a database returns its VP of Sales, you have enriched an account. You have not established that the VP was the visitor.
| What you provide | What you receive | Appropriate label |
|---|---|---|
| Eligible website activity | A matched organization | Company-level visitor identification |
| Eligible website activity | An associated individual | Person-level visitor identification |
| A known email or supported profile identifier | Additional person fields | Contact enrichment |
| A company domain and role filter | People who might be relevant at that company | Prospect discovery |
Ask each vendor to walk one example from the original signal to the destination record. Keep the activity timestamp, entity level and source attached to the record, and distinguish fields the provider knows from fields it infers. For a broader shortlist, read our data enrichment tools guide.
A practical handoff rule
A matched company can become an account-research task. An accepted person match can update the appropriate contact or owner queue. A suggested contact can enter a separate research process. Keeping these paths separate prevents a rep from telling someone “you visited our pricing page” based only on their employer’s activity.
When company-level identification is enough
Use account-level output when your next action concerns the company: prioritize a target-account list, review campaign relevance, or give an existing account owner additional context.
Suppose an account team already knows the buying group at a target company. A new visit to a product page can justify reviewing the account without claiming a particular executive was there. The value comes from account context, not necessarily from a new person record.
Account-level data also lets you avoid buying detail you will not use. If nobody owns a contact follow-up workflow, a richer record can become another unused queue.
Our B2B software buying guide maps requirements to these operating decisions.
When person-level identification is useful
Evaluate person-level output when the workflow requires the person, their role, and suitable contact details. Examples include reviewing a potential buyer against an ICP, finding an existing CRM contact, or attaching relevant activity to the correct owner’s record.
For this workflow, inspect more than the name. Check current company, role, available contact channel, observation time and suppression state. A record can be correctly identified but still be unsuitable for the action you planned.
Leadpipe Identification serves this use case. Its product documentation describes the visitor fields and filtering controls; actual field availability depends on the record. Use a real sample instead of assuming every field is present for every visitor.
How to compare coverage without misleading yourself
Choose one identity unit and one traffic denominator for each calculation.
- Person coverage: unique matched people divided by eligible unique visitors in the stated cohort.
- Account output: unique matched companies reported separately from people.
- Contact completeness: share of the sampled matched people with the fields required for the action.
- Accepted output: records that meet fit, quality and suppression criteria.
Do not divide person records by sessions and compare that with a vendor’s company count. One company can have several visitors, and one person can create several sessions.
Also separate regions. Leadpipe’s documented operating boundary is company-level by default for EU/UK visitors, with person-level processing requiring affirmative consent. Review the actual scope and configuration with your privacy owner. A US person-level result is not evidence of equivalent international coverage.
A sample-record review you can run
Use approved records and label each result as accepted, rejected or unresolved. This example is a review template, not a claim about measured vendor performance.
| Field to inspect | Question | Record the result as |
|---|---|---|
| Identity level | Company, matched person or suggested contact? | Explicit category |
| Company association | Is the organization appropriate for the record? | Verified, disputed or unknown |
| Employment context | Is the role usable for this workflow? | Current enough, stale or missing |
| Contact channel | Is the necessary field present and appropriate? | Usable, rejected or unavailable |
| Visit context | Is the event recent and relevant? | Timestamp and page group |
| Existing relationship | Is this already a customer or active deal? | Destination owner or suppression reason |
Keep unknowns as unknowns. Do not silently count them as valid matches. A provider should be able to explain how its records are created and how disputed matches are handled.
Matching method is one part of the evaluation
Deterministic matching links records through specific identifiers or signals. Probabilistic approaches infer a relationship from available evidence. Those descriptions help explain a system, but they do not by themselves prove a particular record is current or appropriate for your use case.
A correct identity can have outdated employment details. A current company can have several plausible contacts. A valid visit can still be irrelevant to a buying decision. Test each link in the chain rather than turning an architecture label into a universal quality score.
See deterministic versus probabilistic matching for the concepts, then use the trial scorecard to measure the selected product on your traffic.
Buying a person record does not replace routing
Decide whether an accepted person creates a new contact, updates an existing contact or becomes a review task. Preserve the account owner’s relationship. Do not restart a sequence because the same person returns to the site.
For company output, the next action might be research instead of outreach. That is a valid workflow. Assign an owner and an expiry rule so old activity does not circulate indefinitely as a new signal.
For multiple websites, use the multi-domain guide to test entity joins and client separation. A shared organization name alone is not an adequate key for merging people.
Compare cost using accepted output
The cheapest company alert and the cheapest named contact are not equivalent purchases. Compare the cost of the action you actually need.
If a company match requires research time to find an appropriate contact, include that work. If a person match still requires manual review, include that too. Use current plan terms and observed accepted output, as described in the pricing guide.
FAQ
Does company-level identification provide contact details?
Some products enrich matched accounts with contact data. Ask whether each person is identified as the visitor or merely suggested as someone employed at that company.
Is person-level identification always better?
No. It supplies a different level of detail. Account prioritization may need only company data, while an individual contact workflow needs person-level evidence and suitable fields.
Can I assume a matched person wants a sales call?
No. Identification is an input to review. Fit, context, suppression and your approved contact policy determine what happens next.
How should I choose between providers?
Use the same eligible cohort, keep companies and people separate, review disputed records, and inspect the destination workflow. Compare observed accepted output rather than headline match rates alone.
Evaluate the level you need
Start a Leadpipe trial for 7 days or 500 profiles, whichever comes first, no card. Pick one workflow, inspect the records, and compare the plan options once you know what your team can use.
Sources and review basis
Published by Leadpipe. Vendor references support product descriptions; recommendations reflect our assessment of fit. Confirm current plan and contract terms before buying.
