Nesta página
The measurement chain
Four different signals with four different meanings. Keep them separate and your next move gets clearer.
| Step | Question | What it means |
|---|---|---|
| 1. Permission | Can we measure? | A visitor chooses which optional tracking to allow. |
| 2. Activity | What happened? | A page opens, a guide is read or a button is clicked. |
| 3. Outcome | Did it finish? | A form is accepted, a booking is confirmed or an order is paid. |
| 4. Learning | What needs work? | A person reviews the evidence and tests one improvement. |
This is a concept, not a live dashboard and not a claim about results.
1. Connect the right source
Before you start, have access to the right HighLevel sub-account, permission to edit your website, a published test page and a clear consent choice. Decide who may see contact data, and check which features your account has before buying anything.
Built in HighLevel: native analytics
- Open the location, then Sites > Analytics.
- Choose the asset type, such as Websites or Funnels.
- Select one asset and a date range.
- Use the KPI cards and page breakdown to inspect activity.
Metrics vary by asset type. Keep the selection and reporting period consistent when you compare results. The official guide is the HighLevel help article "How to use site and funnel analytics in HighLevel".
Hosted somewhere else: External Tracking
- Find External Tracking in your sub-account and copy its account-specific script.
- The external form guide places it before the closing body tag, across the whole site.
- Install it once, keep the tracking ID you were given and publish your change.
- Test a supported form. Review Sites > Forms > Submissions > External Forms and the contact record.
Supported forms expose real HTML form fields with names. Include an email field so contacts can be matched. Iframe forms and widgets that do not expose their fields are not supported. The official guides are the HighLevel help articles on external form setup and "Analytics for external tracking".
For external reporting, open Sites > Analytics > External Tracking once the script connects. Inspect page views, form views, responses and completion rate, and filter by domain, page or form. Some activity takes time to appear.
2. Give each tool a job
One journey, three useful views.
- HighLevel: connect interest to follow-up. Use the CRM to inspect a submitted lead, booking or order. Contact attribution is recorded through supported actions, including supported external forms. An anonymous visit does not establish a person's identity. See the HighLevel help article "Understanding attribution and traffic sources".
- Google Analytics 4: understand acquisition. Use traffic and event reports to ask which sources bring useful visits. On a site with client-side navigation, choose automatic history measurement or explicit page views. Using both can count the same navigation twice. Verify route changes in DebugView. See Google's guide for single-page apps.
- PostHog: understand the steps. Use deliberate events to compare a sequence, such as viewing a resource, copying it and opening its tutorial. Keep autocapture and recording off unless you have reviewed what they collect. Set up consent and page views explicitly. See the PostHog configuration docs.
Do not add their visitor totals together. Each tool has its own definitions, consent coverage and blocking. A funnel is an ordered sequence of actions, not a collection of popular pages. Confirm completion in the system that owns the outcome.
3. Check before you trust
Run a small, repeatable verification.
- Decline first. Open a clean browser session and decline optional analytics. Inspect network requests and storage. Your core pages and forms should still work.
- Allow, then navigate. Allow analytics, open two pages and use Back. Check the provider's event view for one intended page view per transition. A request in the browser proves an attempted send, not that the dashboard received it.
- Complete one test outcome. Use clearly marked test details, in an isolated account where possible. Check the submission and its CRM record. Keep automated follow-up from reaching real customers during the test.
- Withdraw and refresh. Turn optional analytics off. Repeat navigation, refresh and inspect storage again. Confirm the saved choice survives. Deleting history that was already collected is a separate process.
If the numbers look wrong
| Symptom | What to check |
|---|---|
| Nothing appears | The account, the published page, the script request, consent, blockers, the date range and processing delay. |
| Views appear, but forms do not | The actual HTML form and its field names, iframe boundaries, and whether a custom app prevents a normal submit. |
| Every click looks doubled | The same tag in the page and in a tag manager, or automatic page views running next to manual events. |
| Numbers disagree across tools | Match the time window, timezone, site and metric definition. Investigate the difference before calling it a broken integration. |
Keep names, email addresses, free-text searches, tokens and IP addresses out of custom analytics events. Use your consent system to gate optional tools, and explain the collection in your privacy notice, including external form capture. Test your actual setup, because a banner alone does not control scripts.
4. Close the loop
A better website starts with one honest question. If people reach a guide but rarely open the next step, inspect that step. Is the link clear? Does it answer their next question? Does it work on a phone? Traffic shows where to investigate. It does not explain the cause by itself.
- Choose one useful outcome. Pick a completed form, a confirmed booking or a paid order. Write down the page and the record that proves completion.
- Choose native or external tracking. For HighLevel pages, open Sites, then Analytics. For another website, copy the External Tracking script from your account and follow its current installation instructions.
- Check consent and collection. Test a declined visit and an allowed visit. Check that the right tools respect the visitor's choice before collecting optional analytics.
- Verify a complete journey. Use clearly marked test data. Compare the page visit, the submission and the contact or booking record. A button click alone does not prove completion.
- Review one change. Record the reporting period, the suspected problem and one proposed improvement. Review and test the change before publishing, then compare equivalent periods.
For a repeatable review, keep a short proposal: source, date window, observation, hypothesis, confidence, suggested change and review status. Compare similar periods after release. Small samples and unrelated site-wide event counts cannot prove that a page caused an outcome.
Want to build your first measurement loop with others? Bring your question to RocketLauncher University, our community that is free to join, then apply it to your own site. Prefer a practical next build? Explore HighLevel workflows.
Source documentation checked on October 9, 2026. Account menus and availability can change.