Search Visits Rose From 65 to 412. Were Inquiries Growing?

More search visits do not establish more business when inquiry tracking is broken. An agency's published account of its own site illustrates why measurement deserves attention before a company evaluates AI search optimization.

AI search optimization aims to make a business easier to understand and discover in AI answers. For teams publishing more articles or rebuilding a site, an early question is practical: when an inquiry arrives, can we find its record?

Visits rose from 65 to 412 a month; inquiry outcomes remained unknown

Evolve Media Agency reports average monthly Google and Bing organic visits rising from 65 in October 2025–March 2026 to 412 in June–August 2026, using data pulled September 15. It expanded guides, organized related resources, and changed its homepage. But inquiry tracking broke after site changes, and booking clicks were not tracked as outcomes. It could not calculate inquiry conversion rates or isolate the contribution of each change to traffic growth.[1]

Our focus here is a test: after updating a site, do the form's receipt record and the analytics record still refer to the same completed inquiry?

Decide what counts as one outcome

The following workflow is our recommendation. It requires access to the form's administration and the site's analytics. If different people control those systems, involve both.

Open the actual page containing the inquiry form. Clicking “Contact,” reaching the input screen, and completing a submission are different actions. Decide which one counts as the business outcome.

For a company taking quote requests, an accepted form submission can be one outcome. Counting a click on the quote button as an inquiry includes people who abandon the form.

Action	What the record means	What it does not establish
Reading an article	Someone reached the explanation	Interest in buying
Clicking a contact button	Someone tried to reach the form	A completed submission
An accepted submission	The inquiry arrived	A qualified opportunity or sale
Reviewing the inquiry	The team can assess suitability	Revenue already earned

Use the same names across the team. If a report uses “conversion,” specify the completed action. A conversion is a goal such as an inquiry or purchase, rather than a universally identical metric.

Match one test submission in both systems

Use a test environment if the form has one. If testing is possible only in production, tell the receiving team when the test will happen and avoid triggering real customer communications or orders. For payments or bookings, use the service's supported testing method.

1. Record the page and test time.
2. Enter clearly identifiable test content and submit once.
3. Confirm that the form's receipt list contains that submission.
4. Check whether analytics recorded completion of the same test.
5. Keep the test separate from real inquiries in reporting.

Names and email addresses do not need to be sent into analytics. Match the test through non-personal information such as the time, page, and an internal test reference.

Check whether records arrive in GA4

GA4 is Google's analytics service. Open the site's property in Google Analytics and use Realtime to inspect recent activity. DebugView supports closer inspection of test activity.[2][3]

DebugView requires debug settings for the device or implementation being tested. Merely opening the screen does not display every test. The person configuring analytics can follow Google's DebugView instructions.[3]

An event is a recorded action, such as a page view or completed submission. Names vary by implementation, so ask which event represents successful submission. Finding a button-click event alone is not a successful completion test.

Use these views to establish whether records arrive. Do not settle historical reporting or source attribution solely from a responsive real-time view.[2][3]

Use the mismatch to narrow the repair

Different totals do not immediately prove that AI referrals are missing. Start by locating the point at which the known test disappears.

Test result	Next place to investigate
No receipt in the form system	Validation errors, submission processing, or receipt failure
Receipt exists, analytics event does not	Success-event setup, consent settings, and tag coverage
An outcome appears without submission	Whether a click or page view is incorrectly counted as completion
One test appears several times	Duplicate handlers or events repeated after refreshing
Only the old form is tracked	A mismatch between the new form location and tracking coverage

Consent settings let visitors choose whether measurement is allowed. Some visits may remain unrecorded because consent was withheld. Investigate the actual configuration; do not override that choice simply to increase recorded totals.

Repeat the test after repairing the issue. The result to establish is that a known accepted test is recorded as intended, not merely that a tracking tag exists. This does not guarantee identical measurement for every device and visitor.

Track AI inclusion, visits, and inquiries separately

After checking inquiry measurement, investigate the entry point. An AI answer linking to a page is different from a person clicking through. Whether that person then contacts the business is another question.

AgentSignal provides AI citation checks and site-visit measurement. Follow the AI citation guide to register the site and questions. Check the credits required for the chosen AI services and analysis frequency.

Visit measurement requires installing the tracking tag. This supports reviewing observed citations and recorded visits separately. Installing AgentSignal does not automatically repair an existing GA4 completion event.

For comparisons with other search visits, see the AI traffic and outcome measurement guide. Tools can use different collection and reporting rules, so identical date ranges do not require identical totals.

Keep a short check record for future changes

Make this a repeatable check when a form or inquiry button changes. The following record helps a future editor understand what was tested.

Record	Information to retain
Changed location	Page and form name
Test conditions	Time, device, consent state, and environment
Receipt result	Whether the expected content arrived
Measurement result	Whether the right action was recorded without duplication
Retest	Repair made and subsequent result

Then compare acquisition results. Record identifiable AI referrals, accepted inquiries, and qualified inquiries with consistent definitions and periods. Include dates of advertising or pricing changes that may also affect results.

Before evaluating an article's return, trace one known submission through both systems. It is a small task that prevents a larger reporting problem: knowing that traffic grew while remaining unable to tell whether it produced business.