Website visitor identification: what it can and cannot tell you
Most B2B websites know how many people visited and almost nothing about which companies they represent. Visitor identification tries to close that gap. The useful question is not “can it identify everyone?” but “which decision can this signal support?”
The problem: traffic is not the same as demand
A company can visit your pricing page three times, compare two competitors and leave without filling a form. Analytics records a session. The sales team sees nothing. Visitor identification tries to turn that anonymous session into an account-level signal that someone can act on.
The signal is useful only when it is tied to a workflow: a company, a score, a contact, an owner and a next action. A dashboard full of unidentified traffic is not a sales system.
How identification works
- A script observes the visit. It sends the page, host, referrer, a local visitor identifier and request metadata to the server.
- The network is resolved. The server uses the IP and network data to estimate the company or network owner. This is where coverage is won or lost.
- The account is scored. Page weights, repeat visits, recency and confidence determine whether the account is hot, warm or cold.
- Contacts are suggested. Kiodo can search for relevant people at the company and reveal contact data when the plan and credits allow it.
- The signal is routed. Slack, CRM or the workspace queue gives the team a next action.
The person who appears in the contact list is a person at the company. That is different from saying the person visited the site.
Installation in practice
Kiodo uses one script in the global <head> of the website. The snippet contains a public pixel key, not a secret API key.
<script async
src="https://app.kiodo.dev/api/intent/px/YOUR_PIXEL_KEY">
</script>After publishing, open the website, wait for the event and verify it in the workspace. For single-page applications, make sure the script is loaded once and that route changes are captured. Do not put API credentials in client-side code.
Installation is not the same as coverage. The first event proves the tag works; it does not prove that the provider can identify the network behind it.
Why residential traffic is difficult
Corporate traffic is often easier to resolve because companies own address ranges and use stable egress networks. Residential traffic is different: the IP usually belongs to an ISP, not to the employer. The same is true for mobile carriers, VPNs, privacy proxies and CGNAT.
Those visits can still be recorded, but the system should mark them as low confidence or “network” rather than inventing a company. A conservative product should prefer “we do not know” to a false account.
The limits of identification
- Coverage is partial. No provider identifies every company visit.
- Confidence is not certainty. A company match can be strong and still be wrong at the edge.
- Company is not person. An IP usually identifies a network, not an employee.
- Intent is a model. A pricing visit is a signal, not a buying commitment.
- Consent and law still apply. Technical capability does not create a legal basis.
Concrete example
An illustrative account: a visit from a corporate network resolves to Example Co with a confidence score. Kiodo records the page, adds intent and shows the account. The team can then reveal relevant contacts at Example Co, choose an owner and send a contextual message.
The correct conclusion is: “Someone at Example Co is researching this topic.” The incorrect conclusion is: “Alex Morgan visited the site.” Kiodo is designed around the first statement.
What to do after identifying a company
Identification alone does not create pipeline. The next step is qualification, contact selection and a relevant commercial action. Read the operational playbook for a step-by-step workflow.
Privacy and compliance
Visitor identification processes IP addresses, page activity and sometimes professional contact data. The legal analysis depends on jurisdiction, legal basis, notices, retention and opt-out. See visitor identification and GDPR for primary sources and technical distinctions.
How to evaluate a provider
- Ask for measured coverage in your target countries, not a global claim.
- Run a controlled test on known corporate and residential traffic.
- Check how the provider handles VPN, hosting and CGNAT.
- Review the confidence model and the false-positive process.
- Confirm what contact data is included and how it is sourced.
- Review the DPA, retention and opt-out process before launch.
Website visitor identification FAQ
What is website visitor identification?
It is the process of using network and device signals from a website visit to estimate the company, and sometimes the person, behind that visit. It is probabilistic, not a complete visitor log.
Does it identify every visitor?
No. Coverage depends on the network, country, consent, device and the provider's data. Residential connections, mobile networks, VPNs and CGNAT are frequently not identifiable as a company.
Can it tell me which employee visited?
A company-level signal cannot prove which employee visited. Some vendors market person-level matching in supported cases, but the match remains probabilistic and should be treated carefully.
What does Kiodo identify?
Kiodo identifies visiting companies, scores intent and suggests relevant contacts at those companies. It does not claim that a suggested contact is the person who visited the site.
How do I install Kiodo?
Add the Kiodo script once in the global head of your website, publish it, open your site and verify the first event in the Kiodo workspace. Keep API keys server-side; the pixel key is public by design.
See the workflow on your own site.
Start freeThis page describes the current Kiodo workflow. Coverage, contact availability and legal requirements vary by site, market and configuration. It is not legal advice.