Google Analytics Not Working: A Step-by-Step Fix Guide

When Google Analytics is not working, it usually means one of three things: reports are empty, traffic fell to zero overnight, or every visit suddenly shows up as direct. The cause is almost always technical, and a methodical check will find it faster than guessing.

First, confirm it is really broken

Standard GA4 reports lag behind real activity. Data can show up hours later, sometimes the next day. If the property was created today, empty reports are normal, not a fault.

For a quick test, open the Realtime report and visit your own site from a phone on mobile data. If your visit appears, collection works and the problem sits in reports, filters, or expectations. If it does not, move on to the tag.

Also compare against an independent source: leads in your CRM, Click and Payme payments, orders arriving in Telegram. If orders keep coming while analytics stays silent, the failure is in data collection.

Tag installation problems

Most cases of Google Analytics not collecting data come down to how and where the code was placed. Check this first: it is quick and needs no access to property settings.

  • The tag is missing. View the page source and look for a measurement ID starting with G-. Tags often vanish after a theme change, a disabled plugin, or a move to a new platform.
  • Wrong ID. The code sends data to an old or test property while you look at another one.
  • Double installation. The tag is both in the site code and in Tag Manager, so pageviews are duplicated.
  • Partial coverage. A blog, cart, or subdomain has no tag, and moves from those pages look like new sessions.
  • Old Universal Analytics code starting with UA- no longer collects anything.
  • In Tag Manager, changes are saved but the container was never published.

Settings, consent and browsers

If the tag is fine but numbers are too low, look at settings and visitor behavior. A misconfigured cookie consent banner can stop analytics from firing at all. Test the tag in two scenarios: a visitor who accepts and a visitor who declines.

An internal traffic filter with an overly broad rule can exclude real customers too. Switch it to testing mode and see whether data returns. Ad blockers and privacy-focused browsers create a steady gap against your CRM, but they do not wipe reports completely.

Single-page sites built with React or Vue change pages without reloading. If only the first load is recorded, later views disappear. A Content Security Policy header can also block requests to Google domains, and those errors show up in the browser console.

Data exists, but every visit is direct

A sudden jump in direct traffic usually means the source got lost. Typical culprits: a redirect that strips UTM tags, a switch from https to http, and links from messengers or apps that pass no referrer.

A common trap for online shops: the buyer goes to a Click or Payme payment page and comes back. If the payment domain is not on the referral exclusion list, the payment overwrites the source and revenue looks direct.

What website analytics cannot see

Sometimes analytics works perfectly and the picture is still incomplete. Many businesses in Uzbekistan close sales in Telegram rather than on the site: a person sees an Instagram ad, opens a bot or chat, and orders there. For GA4, that customer ends at a “Message us on Telegram” button.

The fix is to measure the whole path. A bot link can carry a start parameter that passes the source, and key events such as a lead or a payment can be sent to GA4 through the Measurement Protocol. Reports then show the outcome, not just a button click.

Without this, it is easy to reach the wrong conclusion that ads are not working, when orders are simply arriving through a messenger.

A checking order that saves time

Work top to bottom and do not skip steps. It makes separating a technical cause from an advertising problem much easier.

  • Realtime: is your own visit visible.
  • Page source: is the tag present and which ID does it use.
  • DebugView: which events are sent and are there duplicates.
  • Different page templates: home, catalog, cart, blog.
  • Property settings: filters, domains, consent.
  • Site change history for the days the data dropped.

After the fix

Record the dates of the failure and the repair in a change log. Reports for that period will stay incomplete, and you need to remember this when comparing months.

If you want leads from a Telegram bot to appear in the same analytics as your website, talk to the markdev team. We can review your funnel and suggest how to connect the site, Instagram, and the bot in one report.

FAQ

How long should I wait for data to appear in Google Analytics?

Standard GA4 reports can lag from a few hours to a day. For an instant check, use the Realtime report and DebugView.

Why did all my visits become direct traffic?

Usually the source is lost through redirects, stripped UTM tags, messenger links, or a return from a payment page whose domain is not excluded.

How can I tell the tag is installed twice?

Each visit produces duplicate pageviews, and DebugView shows repeated events. Keep one method only: site code or Tag Manager.

Can I track Telegram bot orders in GA4?

Yes. Pass the source with a start parameter in the bot link and send key events to GA4 through the Measurement Protocol.

Let’s look at your task for free

Leave your contact — in a 15-minute call we’ll show where leads get lost and what to fix first.