Follow the visit beyond the contact button

The example above follows three possible points in a visit. Opening the contact page creates a page-view observation. Requesting an email draft creates an interaction. Receiving an inquiry requires evidence from somewhere else, such as the mailbox or a customer-management record. Selecting the final example adds that separate record; it does not turn the earlier click into proof of delivery.

This distinction belongs in the event names. A label such as “email draft requested” is useful because it tells the person reading the report exactly what happened. “Lead generated” makes a larger claim. If the website has no submission backend, a click cannot establish that an email was sent or received.

A copied project brief has a similar limit. The visitor has prepared a message, but they may send it later from another account, or never. Count the copy as an action, not an inquiry, and keep the contents of their message out of your analytics.

The practical follow-up can be modest. When an inquiry arrives, record what the person needs, whether it fits the business and how they found you when that is known. Some people will return directly or contact a colleague. Leave that uncertainty visible instead of forcing every conversation into a perfectly attributed funnel.

Before improving a conversion chart, read the pages a prospective buyer lands on. Can they tell what you build? Can they see the judgment behind the work? Does the project story answer a question they recognize, or could its paragraphs belong to almost any company?

Pages that explain something specific give visitors a reason to continue: the tradeoff behind a decision, an awkward exception, a small example they can try. Google’s people-first content guidance asks the same question: does the page serve the person reading it?

If a good page gets no visits, check that it can be found: a stable URL, internal links to it, an accurate title. Search Console will show whether Google has indexed it.

A rebuild often changes the measurement at the same time as the design. That makes the first comparison surprisingly easy to misread. If the old site counted only full page loads and the new one also counts internal route changes, an increase in page views can partly reflect the new instrumentation.

For a site with client-side navigation, choose who owns each page view. Application code, automatic measurement and Google Tag Manager’s history-change trigger can overlap. Test direct entry, an internal link, back and forward, then inspect the receiving analytics system. Seeing an event fire in the browser does not show that Google Analytics received it, so check the reports as well.

Keep search activity, on-site activity and inquiry quality alongside one another. Search Console can describe search impressions and clicks. Analytics can describe recorded visits and actions. The inquiry record tells you what someone wanted and what happened next. Comparing the same periods helps, as does noting campaigns or other changes that brought a different audience.

Then choose a specific repair. A useful page that visitors cannot find needs a discovery check. A contact action that fails on a phone needs a working mobile path. An inbox full of unsuitable requests may call for a clearer explanation of the service.