GA4 Setup for Marketers

GA4 setup gets rushed more than almost any other marketing task, because it feels like a one-time chore instead of the thing every other report depends on.

Get it wrong and you won't know for weeks — the account looks fine, it just quietly isn't collecting the data you'll need for the report due at month end. Here's the process, done properly, once.

What you need before you start

• A Google account with edit access to Google Analytics (or the ability to create a new GA4 property).
• Access to your site's code, or a Google Tag Manager container already installed — you need one or the other to add the tracking tag.
• A clear list of what actually counts as a conversion for your business — a form submit, a purchase, a call click. Decide this before you start; it changes what you mark as a key event in step 6.

1. Create the GA4 property

Go to Admin (bottom-left corner of Google Analytics), then under the Property column select Create Property. Enter the property name, your reporting time zone, and currency — for an India-based business, set the time zone to India Standard Time so your reports align with your actual business day.

When prompted for business objectives, "Get baseline reports" is a safe default if you're not sure which option applies.

2. Create a data stream

After the property is created, choose Web as the data stream type. Enter your website URL — the protocol (https) is usually auto-detected — and a name for the stream.

You'll be offered Enhanced Measurement toggles (scroll tracking, outbound clicks, site search, video engagement, file downloads) at this stage. Leave these on unless you have a specific reason to turn one off — they capture useful events automatically, with no extra setup.

3. Get your Measurement ID

Once the stream is created, you'll see a Measurement ID in the format G-XXXXXXXXXX. This is the unique identifier that routes tracking data to this specific property and stream — you'll need it in the next step, so keep the tab open or copy it somewhere safe.

4. Install the tag (Google Tag Manager method)

Google Tag Manager is the recommended install method because it lets you manage this and future tags without a developer touching site code every time.

1. In GTM, create a new tag. Tags > New > choose the Google Analytics: GA4 Configuration tag type.
2. Paste your Measurement ID from step 3 into the Tag ID field.
3. Set the trigger to "Initialization — All Pages" so it fires on every page load, sitewide.
4. Test in GTM's Preview mode before publishing — this opens a debug panel showing whether the tag actually fires.
5. Publish the container using the Submit button once Preview confirms it's working.

If you're not using Tag Manager, GA4 also supports a direct install by pasting the gtag.js snippet into your site's header — check with whoever manages your site template if that's the route you need instead.

5. Verify data is actually arriving

This is the step that gets skipped, and it's the one that actually matters.

Go to Admin > DebugView in GA4 while browsing your own site with debug mode active (GTM Preview mode does this automatically; alternatively, install the Google Analytics Debugger Chrome extension). You should see your own events appearing in near-real time.

Separately, check Reports > Realtime — this confirms data is flowing for real visitors, not just your own debug session. Give it a few minutes after publishing; don't assume failure if nothing shows up in the first sixty seconds.

6. Configure key events

GA4 tracks a lot automatically, but it doesn't know which of those events are the ones your business actually cares about — that's your job to tell it.

Go to Admin > Key events (previously called "conversions" in earlier GA4 releases — same concept, updated name) and mark the specific events that represent a real business outcome: a form submission, a completed purchase, a phone number click. Don't mark everything — a key events list padded with low-value actions makes your conversion rate and cost per acquisition numbers meaningless for actual decision-making.

Do I need Google Tag Manager, or can I skip it?

You can skip it — GA4 supports installing the gtag.js snippet directly in your site's header, and for a very simple site with no other tags to manage, that's a perfectly reasonable choice.

Where GTM earns its place is the moment you have more than one tag to manage — GA4, a Meta Pixel, a LinkedIn Insight Tag, a conversion tracking snippet for Google Ads. Managing all of those directly in site code means a developer touches the header every time marketing needs a change. GTM moves that control to a dashboard marketers can use directly, without a code deployment for every small change. If you're already running paid ads on more than one platform, install GTM now rather than migrating later.

Common GA4 setup mistakes

• Skipping DebugView entirely. The setup "looks" complete in every screen except the one that actually proves it's working.
• Installing the tag twice. A leftover gtag.js snippet from an old setup, plus a new one added via GTM, double-counts every event. Check for this specifically if your numbers look unusually high.
• Marking too many events as key events. Every button click and page scroll doesn't need to be a key event — reserve that label for outcomes that represent genuine business value.
• Never checking back after the initial setup. A site redesign, a new checkout flow or a moved contact form can silently break a key event without anyone noticing until the monthly report looks strange.

What if you already have GA4, but it's not set up properly?

This is more common than a from-scratch setup, honestly. Symptoms usually include: key events that clearly aren't firing (a form that gets real submissions but shows zero conversions in GA4), duplicate or inflated numbers, or a property that was created years ago with none of the current Enhanced Measurement options ever reviewed.

The fix is the same six-step process above, run as an audit rather than a fresh build: confirm the data stream and Measurement ID are correct, check DebugView for what's actually firing versus what you assume is firing, and rebuild the key events list based on what the business actually needs measured today rather than what someone guessed at eighteen months ago.

How long until GA4 data is reliable enough to act on?

Give it at least two to four weeks of real traffic before drawing conclusions from anything beyond the basic "is it tracking at all" check. GA4's own modelling (used to fill gaps from cookie consent choices and cross-device behaviour) becomes more accurate with more data, and a lot of early-stage numbers look odd simply because the sample size is still small. For a low-traffic site, give it longer — a month with thin data tells you less than two weeks on a high-traffic site.

Setting up value on your conversions, not just a count

A key event that only tells you "this happened 40 times" is less useful than one that tells you "this happened 40 times and was worth ₹2,80,000." If your business has a knowable value per conversion — an average order value for e-commerce, an average deal size for lead generation — attach it.

For e-commerce, this is usually automatic once your platform's purchase event is correctly configured to pass value and currency parameters. For a lead-generation business without a fixed price, a reasonable approach is assigning an estimated average value per qualified lead based on your actual close rate and average deal size — an honest estimate, clearly understood internally as an estimate, not a precise number GA4 somehow knows on its own. This is what turns a raw event count into a number your finance team will actually engage with, and it's the same figure a PPC audit uses to judge whether a campaign's cost per acquisition is actually worth it.

Consent mode, briefly

If your site shows a cookie consent banner — increasingly standard practice, and worth having regardless of GA4 — how a visitor answers it affects what GA4 can actually record. Someone who declines analytics cookies still generates modelled, aggregated data in GA4 rather than nothing at all, provided Google's consent mode is correctly wired into your consent banner. If your reported numbers seem to sit meaningfully below what other evidence (like server logs or ad platform click counts) suggests your real traffic is, an incorrectly configured consent banner is one of the first places to check — not a GA4 setup mistake on its own, but a common reason a technically correct GA4 setup still under-reports.

What to check a week later

Come back after a week of real traffic and confirm the key events you configured are actually firing at a volume that makes sense against your known traffic — not zero, and not suspiciously higher than your actual form submissions or sales. Both directions usually mean a tagging problem, not a business change.