GA4 for Local Businesses: Measure Enquiries Before You Measure Everything
A useful Google Analytics 4 setup for a local business answers two questions: where does website activity come from, and which visitors take a meaningful next step? Start with one dependable measure of an enquiry, a few supporting actions, and a habit of comparing the numbers with the requests the business actually receives.
You can measure more later. If a click on a phone link is currently being reported as a new customer, adding twenty more events will only make the dashboard bigger. It will not fix the central mistake.
Treat this guide as a measurement brief an owner can hand to a website supplier. The examples are illustrative. The right collection and consent arrangements depend on your website, your visitors and the rules that apply to you.
Know which system answers which question
| System | What it helps you see | What it cannot tell you alone |
|---|---|---|
| Search Console | How your website appears, and gets clicks, in Google Search | All your website visitors, or finished offline jobs |
| Google Business Profile performance | Supported interactions with the profile | The full journey through your website and CRM |
| GA4 | Collected website or app events, and the acquisition information available | Everyone who contacted you, or whether a job paid off |
| Enquiry inbox or application | Requests your business actually received | Every earlier marketing touch |
| CRM or job record | Qualification, next actions and outcomes the owner recorded | Independent proof that a marketing channel caused a sale |
If you have no website, do not set up website analytics just to look present in a map product. Settle the question you want answered, and which system answers it, first.
Audit what is installed before adding anything
Ask whoever looks after the website to identify the Analytics account, the GA4 property, the web data stream, the measurement ID and how it was installed. Write down who controls access and which domain is being measured.
A site can get tracking from a theme, a plugin, Google Tag Manager or a hard-coded tag. Adding a second installation without checking the first can duplicate events and muddy ownership.
Google keeps the account, property, data stream and website tag separate. The property also holds reporting settings, such as timezone and currency, that change how reports read. Google's GA4 setup instructions
Keep an installation register:
| Item | What to record |
|---|---|
| Business owner | The person accountable for the property and its access |
| Property and stream | Names or IDs enough for authorised staff to find the right setup |
| Website domain | The production domain and any relevant subdomains |
| Tag method | Where the live configuration is maintained |
| Reporting timezone | The calendar you use to compare with enquiry records |
| Event owner | The person who approves a change to a trigger |
| Last controlled test | Date, device, result and known limits |
Keep passwords and customer data out of the register. Give suppliers individual permissions, not a shared owner login.
Define the enquiry before you name an event
Write what an enquiry means for your business, in one sentence. For example: a website enquiry counts when the application has accepted and stored a valid request that staff can retrieve.
That is stronger than "a visitor pressed Send". A click can happen before validation, during a failed request, or several times during a network hiccup.
Use separate measures for separate levels of certainty:
| Action | Suggested treatment | Limit |
|---|---|---|
| Phone link selected | A supporting contact-intent event | Does not show a call connected |
| Email link selected | A supporting contact-intent event | Does not show a message was sent |
| Enquiry accepted and stored | The primary website lead event | Still needs a spam and fit check in business records |
| Appointment requested | A request event with a clear meaning | Does not show availability was confirmed |
| Staff accept suitable work | A CRM or job outcome | Usually beyond what a browser can observe |
Google publishes generate_lead as a recommended event for the first stage of lead generation. Use it where the interaction fits, with a documented trigger. Mark a properly collected event as a key event when it stands for an important business action. Marking it does nothing to fix a weak trigger. Recommended lead events, events and key events
Give the developer an event contract
A short contract stops assumptions hiding inside a tag configuration.
Event. A valid enquiry was accepted.
Trigger. The website reports that the request was successfully accepted. It is not an arbitrary button press, and it is not any visit to a thank-you URL.
Count. One collection for that accepted interaction. The application's own lead identity prevents duplicate records, but the analytics code must separately avoid sending the same event again on a refresh, a retry or a duplicate listener.
Safe context. A controlled service category or form identifier where that is justified. Never send the enquiry text, customer name, email or exact service address.
Failure behaviour. A rejected form does not count. A saved enquiry stays real even if the visitor declines measurement or the analytics request is blocked.
Test evidence. The accepted request reference in the protected system, the event as observed where collection is permitted, and a check for duplicate emission.
GA4 does not become your operational database because it receives a lead event. Keep the actual request in the business system, independently.
Keep personal information out of collection
Check URLs, page titles, search terms, event parameters and form integrations. Personal information can leak through a confirmation URL even when nobody added a custom analytics field on purpose.
For example, a URL containing the customer's email or a message preview can expose it to systems that receive page-location data. Use a neutral confirmation page and keep private request details behind proper access control.
Google's policy forbids sending information it can recognise as personally identifiable, including email addresses and personal phone numbers. Its guidance also covers URLs, campaign parameters and fine-grained location. Analytics PII guidance
Consent and privacy controls can mean a real enquiry has no matching analytics event. Treat that as a limit of the measurement. It is no reason to bypass a visitor's settings.
Test the setup with real cases
A controlled test should include both success and failure. The owner and the website supplier can use this brief.
| Test | What the business system should show | What analytics should show, where permitted |
|---|---|---|
| Submit one valid request | One retrievable enquiry | One intended lead event |
| Submit with a required field missing | No accepted enquiry | No accepted-lead event |
| Refresh the confirmation view | Still one record | No second lead from the refresh alone |
| Press Send repeatedly during a slow response | One accepted interaction, or a clear status recovery | No multiplying from repeated clicks |
| Decline optional measurement | The enquiry can still be handled | Collection follows the visitor's settings |
| Select a phone link, then cancel the call | No claim of a completed call | Only the defined click event |
| Lose the network after acceptance | The enquiry stays retrievable | No promise the event arrived |
Use the current Tag Assistant or permitted debug tools on the relevant device. GA4 DebugView shows collected events while you troubleshoot. It needs debug mode and respects the relevant privacy controls. DebugView documentation
Write down what you really tested. An event in DebugView is good evidence of collection. An inbox check is separate evidence of the business outcome.
Read acquisition reports at the right scope
The Traffic acquisition report describes where sessions came from. The User acquisition report describes how new users were first acquired. Choose the one that fits the question you are asking. GA4 Traffic acquisition documentation
For a first review, use a fixed date range and look at the session source and medium, or the campaign, next to the useful action. If a report is missing from the navigation, ask an authorised editor to check the reporting configuration before anyone rebuilds the property.
Use the UTM guide to make your listing links distinguishable. Keep literal campaign labels apart from any channel grouping. A channel label is a reporting rule. It is not proof of a paid contract or a completed sale.
Worked interpretation: an upholstery workshop
This fictional example shows why website numbers and business records belong side by side.
| Measure over one illustrative month | Count |
|---|---|
| Measured website sessions | 600 |
| Contact-button clicks | 48 |
| Measured accepted-enquiry events | 18 |
| Accepted requests in the business system | 22 |
| Suitable enquiries after staff review | 15 |
| Jobs accepted from that enquiry group | 6 |
It would be wrong to call the 48 clicks new leads, or to treat the 18 events as all the enquiries. In this example four accepted requests have no matching measured event. The team would look into the gap, while keeping the 22 operational records as the basis for handling requests.
The 15 suitable enquiries help judge customer fit. The six accepted jobs need a stable definition and enough time for that group to mature. Comparing this month's visits with jobs won from older enquiries gives a misleading conversion story.
The owner also checks whether calls and walk-ins are recorded separately. Those routes can be valuable even when they cannot be joined reliably to an individual website session.
Run a short weekly review
Choose the period, check that collection is healthy, and compare received enquiries with measured events. Then see which sources and pages go with suitable requests.
Ask what action the numbers support. A service page with many unsuitable enquiries may need clearer scope. A source with good-fit requests and slow replies may need operational attention before more advertising. A sudden drop in events during a tag change needs technical investigation before anyone draws a marketing conclusion.
Use the marketing economics guide to combine these observations with cost and contribution. Analytics becomes commercially useful when it informs a decision the business can act on.