Google Analytics: Find Where Your Visitors Come From

You shared your app in a community, sent an email, and posted on LinkedIn. Visits went up. Which effort should you repeat?

Google Analytics 4, or GA4, can help you trace those visits. This guide has one goal: compare your traffic sources and choose one small marketing test. You do not need to build a dashboard or buy ads.

What question should you answer first?

Start with: “Which places bring visits to my public website?” A traffic source is the place someone came from. It could be a search engine, an email, or another website.

For now, compare sessions: groups of activity during a visit. Sessions are not unique people, and visits are not customers. One person can return several times without signing up.

Use these numbers to plan your next test, not to declare that your business works. Google explains the visit-based view in its Traffic acquisition guide.

How do you connect GA4 to your live website?

A GA4 property holds your website's reports. A web data stream is the connection that receives activity from that website.

1. Open Google Analytics with a Google account you control.
2. Create an account if needed, then create a property for your business. Choose your reporting time zone.
3. Under Admin → Data streams, add a Web stream for your final public address.
4. Open that stream and copy its measurement ID. For this GA4 website setup, look for the ID beginning with G-.
5. Check whether your website builder has a Google Analytics integration. If it does, follow that builder's instructions.

If AI built your app, ask it to inspect the existing tracking before adding anything. Installing through two different methods can count the same activity twice. Google's setup guide covers the available installation routes.

Give your coding AI a bounded task

Goal: measure visits to my public landing pages with GA4.
First inspect the app for existing Google tags and consent controls.
Explain where the tag belongs before changing files.
Use the measurement ID I provide; do not invent one.
Keep measurement off private account and admin pages.
Do not send names, emails, form text, or access tokens.
Check page URLs and titles for private data too.
Preserve consent choices and prevent duplicate page views.
If consent handling is missing, stop and explain what I must decide.
Keep sign-in, payments, and page content unchanged.
Show how to test the setup before I approve publishing it.

Before collecting data, review Google's consent guidance and the rules that apply to your visitors. A working tag is not proof of compliant consent. Google also prohibits sending identifying information such as email addresses, including through page URLs, titles, and campaign labels.

Check the published version

Publish the approved site change. Visit a public page and make the appropriate consent choice in your test browser. Then check GA4's Realtime report for activity.

Initial collection can take up to 30 minutes to begin. Many standard reports need 24–48 hours to process data. Use Realtime or DebugView to verify collection; do not reinstall the tag because the main report is still empty.

How do you tell two promotions apart?

Add labels called UTM parameters to the links you share. They describe the promotion, not the person clicking it.

For a first test, keep a small naming sheet:

Use Google's Campaign URL Builder to combine these values with your landing-page address. This example uses a placeholder domain; replace it with your own:

https://example.com/?utm_source=linkedin&utm_medium=social&utm_campaign=first-launch

For your newsletter, use founder-newsletter as the source and email as the medium. Keep first-launch as the campaign so both links belong to the same test.

Keep spelling and lowercase consistent. LinkedIn and linkedin become different values. See Google's UTM guide for the label rules.

Open each finished link before sharing it. Check that it reaches the right page and that redirects preserve the labels. Use campaign labels on incoming promotion links, not on the buttons between your own pages. Creating a labeled link does not require a new app release.

Where do you compare the results?

1. Open Reports → Acquisition → Traffic acquisition. Navigation can vary; an editor may need to add this report.
2. Choose a date range that includes your promotion and allows time for data to finish processing.
3. Set the table's main grouping to Session source / medium.
4. Compare Sessions for rows such as linkedin / social and founder-newsletter / email.
5. Use Session campaign to narrow the view to your named promotion.

Read the screenshot from left to right: google / organic in the search field filters the rows. Session source / medium names the grouping. Sessions is the visit count to compare. Clear the search filter when you want to see all sources. On a phone, open the full-size image; the steps above remain the source of instructions.

Use Traffic acquisition for this task. User acquisition answers a different question about how people first arrived. Keep the same report and date range when comparing results.

Do not expect every visit to have a useful source. Consent choices, blockers, missing labels, and links shared onward can limit what you learn. Treat the report as evidence with gaps, not a complete record of every person.

What should you do with the numbers?

Imagine an example week with 80 sessions from LinkedIn and 20 from your newsletter. Those are invented numbers, not a target. LinkedIn brought more measured visits; you still do not know which group wanted your product.

Repeat one small test with the same offer and page. Note the time you spent, the audience you reached, and any useful customer replies. A larger audience or more effort can explain more visits. Change one thing at a time so you can learn from the result.

When you need to compare completed signups, add and test an event for a successful signup. An event is a recorded action. Mark a verified business action as a key event to compare it across channels.

A click on “Sign up” is not a completed account. UTM labels alone do not measure signups, purchases, or profit. Keep that next measurement task separate from this first traffic check.

When is your first setup useful enough?

• Your public site sends test activity to the correct GA4 property.
• You checked consent handling, private data, and duplicate tracking.
• Two shared links use distinct sources and one consistent campaign name.
• After processing, you can find the source rows in Traffic acquisition.
• You wrote down one next marketing test, without treating visits as sales.

If a source is missing, recheck the final shared URL and the date range first. If Realtime is also empty, ask your AI to check the tag, consent state, and measurement ID. Do not “fix” a low count by collecting data against someone's choice.

What should you read next?

• Your App's First Customer Journey: What to Build First — choose the useful action visitors should reach.
• Before AI Writes Code, Ask It to Inspect Your App — check what exists before installing tracking.
• AI App Launch Checklist for Solo Founders — test the experience before promoting it.

Need to publish your approved app changes? Explore JustDeploy, then test the live page before sharing your campaign links.

Official Google guidance checked September 5, 2026. Menu labels can vary by account, permissions, and product updates. The examples are illustrative, not results from a customer account.