How to set up Google Analytics 4 for a small business in Australia

Most small businesses have Google Analytics installed and almost none of them use it. The tag was added at some point, the property fills with data nobody opens, and when a real question arrives, such as whether the money spent on advertising last quarter actually did anything, the answer has to be guessed. The problem is rarely the tool. It is that the setup was treated as an IT task rather than a measurement decision, so the property records everything and clarifies nothing.

Google Analytics 4 is an event-based system. Every interaction, from a page view to a purchase, is an event carrying parameters. That design is more flexible than the old session-and-pageview model, and it is also why an unconfigured GA4 property feels like noise: it collects a great deal automatically and makes no judgement about what matters to your business. Supplying that judgement is the entire job. What follows is the order we work through it, and the handful of settings that cause the most trouble for Australian businesses when they are skipped.

What does setting up Google Analytics 4 actually involve?

Five stages, in this order. Doing them out of order is the usual reason a setup has to be redone, because the events you decide to track determine how you install the tag, not the other way around.

Stage	What you decide	Why it comes here
1. Measurement plan	The three to five questions the data must answer	Without it you track everything and learn nothing
2. Property and stream	Account structure, timezone, currency, retention	Some settings are not retrospective, so wrong values cost you history
3. Tag installation	Direct GA4 tag or Google Tag Manager	Depends on how many events and pixels you need
4. Key events	Which actions represent real commercial value	This is what turns traffic data into business data
5. Links and hygiene	Google Ads, Search Console, internal traffic filters	Determines whether channel reporting is trustworthy

The measurement plan is the stage owners skip, and it costs the most. Write down the questions first. Which channels bring enquiries that turn into customers? Which pages lose people before they act? Is the advertising producing anything beyond clicks? Three to five questions like these tell you exactly which events to configure and let you ignore everything else, which is the point.

How should I set up the property and data stream?

Create one property for the business and one web data stream for the site. Resist the instinct to split the property by section or subdomain: GA4 handles a whole site comfortably in a single property, and splitting it makes cross-site journeys invisible. Keep it simple until you have a concrete reason not to.

Four settings matter disproportionately, and three of them are not retrospective:

• Timezone. Set it to your local Australian timezone. Leave it on a United States default and your reporting days are offset from your trading days, which quietly corrupts every day-of-week and daily-trend comparison you make.
• Currency. Set reporting currency to AUD. Otherwise revenue arrives converted, and the numbers will never reconcile with your accounting.
• Data retention. The default retention window for user-level data is short. Raise it to the longest period your property allows, or your year-on-year comparisons will simply be unavailable when you want them.
• Internal traffic. Define and filter your own office and home IP addresses. For a business with modest traffic, the team refreshing its own site is a genuine distortion, not a rounding error.

On installation, use the direct GA4 tag if you need page views and a couple of built-in events. Move to Google Tag Manager once you are handling several custom events or multiple advertising pixels, because it lets you change tracking without a developer touching the site each time. Install the tag once, sitewide. Duplicate tags are the most common cause of halved bounce rates and doubled event counts, and they are tedious to detect after the fact.

Which key events should a small business track?

In GA4, the actions that matter commercially are marked as key events, the term that replaced conversions in Analytics reporting. Google Ads still calls them conversions on its side, which is a naming quirk worth knowing when the two interfaces disagree. Mark three to six. Marking everything inflates your apparent conversion rate and buries the signal.

Business type	Mark as key events	Track but do not mark
Ecommerce store	Purchase, begin checkout, add to cart	Product views, search, wishlist adds
Service business	Enquiry form submit, phone tap, booking confirmed	Pricing page views, brochure downloads
Local or trade business	Quote request, call from mobile, directions tap	Gallery views, service page views

The distinction in that third column is the useful part. Events that are worth having in the data are not necessarily worth counting as outcomes. A pricing page view is a strong intent signal and belongs in your funnel analysis, but treating it as a conversion tells you that hundreds of people converted when nobody actually enquired. Reserve key event status for actions with real commercial value, and make sure each one maps to something you would recognise on an invoice. Getting this right is what lets you calculate customer acquisition cost by channel instead of estimating it.

One hard rule: never send personal information into GA4. Names, email addresses and phone numbers in event parameters or URLs breach Google's own terms and create a privacy exposure you do not want. Your privacy policy should state plainly that you use analytics, what it collects and why, and the consent choices you present to visitors should be honoured by your tag configuration rather than displayed and ignored.

How do I connect GA4 to Google Ads and Search Console?

Two links do most of the work. Linking Google Ads lets you import key events as conversions for bidding, which materially changes what the platform optimises towards. Linking Search Console puts organic query data beside behaviour data, so you can see which searches bring people who actually act rather than just which searches bring clicks. That pairing is what makes keyword research an evidence exercise instead of a guessing exercise.

Expect the numbers to disagree, and do not spend weeks trying to reconcile them. Ad platforms credit conversions to their own click or view under their own attribution window. GA4 applies its own model across channels. Your store or invoicing system counts only completed, paid orders. Each is answering a different question. The workable convention is to treat your commerce or accounting system as the source of truth for revenue, and GA4 as the source of truth for behaviour and channel mix. Comparing the trend in each is far more useful than forcing the totals to match.

Which GA4 reports should I actually look at each month?

Four views answer most questions a small business has. Traffic acquisition shows which channels bring visitors and, more importantly, which bring key events. Landing pages shows which entry points earn attention and which lose it. Key events over time shows whether outcomes are trending up independent of traffic volume. Path or funnel exploration shows where people drop out of a defined journey.

Build a short dashboard from those four and review it monthly against the questions you wrote in stage one. The discipline that separates a useful analytics setup from a decorative one is not the sophistication of the configuration. It is that somebody looks at the same small set of numbers on a regular cadence and acts on what changed. If you are unsure which numbers belong on that list, our guide to the metrics an Australian small business should track monthly is a sensible companion, and for advertising specifically it is worth understanding what a good ROAS looks like before judging a campaign by its GA4 numbers alone.

What does a good setup look like six months later?

A well-configured property is quiet. Key events fire reliably, the channel report is trustworthy enough to move budget on, and when someone asks whether the new landing page is working, the answer takes two minutes rather than two days. You should be able to say which channel produces your best customers, not merely your cheapest clicks, and to notice a decline in month three rather than discovering it in month six.

That outcome comes from restraint at setup and consistency afterwards. Decide the questions, configure the minimum needed to answer them, protect the data quality with timezone, currency and internal traffic filters, and then read it on a schedule. Analytics rewards the businesses that treat it as a monthly management habit rather than a one-off installation. If you want the measurement layer designed properly around the decisions you are actually trying to make, see our data and analytics services or book a consultation and we will map it out with you.