GA4 Tutorial for Beginners

A GA4 tutorial teaches how to set up and use Google Analytics 4, Google's web and app analytics tool. GA4 records what visitors do on a website as events, such as viewing a product, adding it to the cart or buying it, and turns them into reports that show which marketing brings sales.

• Event-based: Every action is an event, from a page view to a purchase, instead of the old sessions and hits model.
• Key events: You mark the events that matter to the business, such as purchases and leads, as key events.
• Cross-platform: One property can collect data from a website and a mobile app together.
• Built-in AI: GA4 flags unusual changes as insights and can predict purchase likelihood for some sites.
• Privacy-aware: It works with consent banners through Consent Mode and can model missing data.

This tutorial follows one example: a D2C tea brand from Assam that sells online, accepts UPI and cards through a payment gateway, and is preparing for its Diwali sale. The owner wants to know which channel, Instagram ads, Google search or WhatsApp broadcasts, brings the most orders.

Prerequisites for This GA4 Tutorial

• A Google account: Use a business account that more than one person in the company can access, not a personal one that leaves with an employee.
• Website access: Admin access to the site builder (such as Shopify or WordPress) or to the site's code.
• Checkout details: Know how the checkout works, including whether customers leave the site to pay through a payment gateway.
• A consent banner, if needed: If the site collects data from users in regions that require consent, plan it with the Consent Mode lesson.

Setup: Create a GA4 Property

1. Go to analytics.google.com and select Start measuring.
2. Account: Name it after the company, for example "Assam Tea Co". One account can hold several properties.
3. Property: Name it after the website, set the reporting time zone to India and the currency to Indian Rupee (INR). Revenue reports use this currency.
4. Business details: Choose the industry and size, then pick the business objectives closest to the goal, such as generating sales. These choices decide which reports appear by default.
5. Data stream: Choose Web, enter the site address and a stream name. GA4 creates a Measurement ID that starts with "G-".
6. Enhanced measurement: Leave it on. It records page views, scrolls, outbound clicks, site search, video engagement, file downloads and form interactions without extra code.

Step-by-Step: Track the Store Properly

Step 1: Install the Google Tag

There are three common ways to install the tag on every page:

• Platform integration: Shopify, WordPress plugins and most site builders ask only for the Measurement ID.
• Google Tag Manager: Best when several tags will be added later, covered in the Google Tag Manager tutorial.
• Direct code: Paste the Google tag snippet from the data stream page into the head of every page.

Step 2: Check Data Is Arriving

Open Reports > Realtime, visit the site from a phone, and confirm that the visit appears within a minute or two.

Step 3: Send the Purchase Event

A sale is recorded with the recommended purchase event. Platform integrations usually send it for you. On a custom checkout, the developer sends it on the order confirmation page with the order details:

gtag("event", "purchase", {
  transaction_id: "AT-10482",
  value: 1198.0,
  currency: "INR",
  items: [
    { item_id: "ASSAM-CTC-500", item_name: "Assam CTC Tea 500 g", price: 599.0, quantity: 2 }
  ]
});

• transaction_id: A unique order number. GA4 uses it to avoid counting the same order twice if the page is reloaded.
• value and currency: The order total and "INR", so revenue appears correctly in reports.
• items: The products in the order, used by the ecommerce reports.

Step 4: Mark Purchase as a Key Event

Go to Admin > Data display > Events, find purchase, and mark it as a key event. Do the same for other important actions, such as a wholesale enquiry form. The design of events is covered in GA4 events and conversions.

Step 5: Fix Payment Gateway Referrals

When a customer pays by UPI through a payment gateway and returns to the thank-you page, GA4 may credit the sale to the gateway's website instead of the ad that brought the customer. Add the gateway's domain under Admin > Data streams > Configure tag settings > List unwanted referrals.

Step 6: Set Data Retention and Internal Traffic

• Data retention: Under Admin > Data collection and modification > Data retention, change event data retention from 2 months to 14 months.
• Internal traffic: Define the office IP address as internal traffic and activate the filter, so the team's own visits do not inflate the numbers.

Step 7: Tag Campaign Links

Add UTM parameters to every Instagram, WhatsApp and email link for the Diwali sale, such as utm_source=whatsapp and utm_campaign=diwali_sale. Without them, broadcast clicks often appear as direct traffic.

Step 8: Read the First Reports

Common Mistakes

• Tag on some pages only: Missing the tag on the thank-you page means purchases are never recorded.
• Duplicate purchases: Sending the purchase event without a transaction ID, or both from the platform and custom code, counts orders twice.
• Wrong currency: Leaving the property in US dollars makes every revenue figure misleading.
• Short retention: Keeping the default 2 months limits detailed analysis of the previous festive season.
• Untagged links: WhatsApp and email clicks without UTM parameters are credited to the wrong source.
• Personal data in events: Never send names, phone numbers or email addresses to GA4; Google's policies forbid it and so does good privacy practice.

Next Steps

• Events in depth: Plan custom events such as add to wishlist in GA4 events and conversions.
• Dashboards: Share a weekly sales view with the team through the Looker Studio dashboard tutorial.
• Organic traffic: Link GA4 with Search Console to see which searches bring buyers, explained in what is SEO.

How AI Changes GA4 Analysis

What AI Automates Now

GA4 itself detects unusual changes, such as a sudden drop in purchases, and shows them as insights. For sites with enough purchase data it predicts purchase and churn probability, covered in GA4 predictive metrics. General AI assistants can also summarise exported reports and suggest explanations.

What Still Needs a Human

Deciding which events matter, checking that tracking is correct, and judging whether a pattern is real or a tracking fault all need a person who knows the business. An AI summary cannot know that the checkout was down for an hour on Diwali.

Risk to Watch

AI tools will confidently explain numbers that are wrong because the tracking is wrong. Check data quality before asking for insights, and never upload exports containing customers' personal data into AI tools, in line with the DPDP Act.

Do It with AI

Use this prompt to get a first read of a GA4 export. It works in ChatGPT, Claude or Gemini. Export the Traffic acquisition report as CSV first and remove any personal data.

1. Export the report for the sale period and a normal period for comparison.
2. Remove any column that could identify a person.
3. Run the prompt and read the tracking problems first.
4. Confirm each finding in GA4 before moving budget between channels.

Check Before You Use It

• Facts: Recalculate one or two figures by hand; AI tools can misread CSV columns.
• Brand fit: Judge the suggestions against what the team knows about the sale, stock and offers.
• Compliance: Share no personal data with AI tools, and follow the company's data policy for what may be uploaded.