The short answer: connect visit data with a record of real enquiries and their quality. Clicking a contact button is not yet a conversation or a sale. Start with one weekly dashboard, shared definitions and someone who confirms what happened to each enquiry inside the business. More charts will not fix faulty measurement.
First define what counts as an enquiry
For a service business, an enquiry might be a received form, a telephone conversation or a message. Record each contact route separately. Viewing the contact page and clicking a phone number indicate interest, but do not prove that the customer reached the business. Tests, spam and repeat submissions of the same message should not increase the number of new opportunities.
Distinguish an enquiry from a qualified enquiry. The latter meets agreed criteria: it concerns a service you offer, an area you cover and a realistic project scope. You do not need to know the budget in the first message to record a worthwhile contact. Apply the same rules to every traffic source, including those receiving the largest marketing spend.
A sale is a later stage, confirmed in your customer system or business records. Do not assign every form the value of a signed contract. Revenue, a potential proposal value and a stated budget are different things. The dashboard should support decisions, not inflate performance by renaming a column.
Three data sources answer three different questions
Google Analytics can show measured sessions, channels and website events, depending on implementation and consent. Search Console supports analysis of your presence in Google Search, including clicks and impressions. An enquiry register or CRM answers whether the business actually received a contact and what happened afterwards. None of these sources replaces the others.
Do not expect Search Console clicks to equal organic sessions in analytics. Definitions, scope, recording time and measurement limitations differ. Do not add both figures into a “total traffic” number. When they disagree, first check dates, time zone and filters instead of immediately declaring the implementation broken.
A dashboard that fits on one screen
Start with a weekly review, but also assess a longer comparable period if traffic is low. Every number needs a source and a definition. The following table contains suggested fields, not actual results for your business. Label missing data as missing instead of entering zero, which would imply a confirmed absence of enquiries.
| Field | Source and meaning | Decision it supports |
|---|---|---|
| Measured sessions by channel | Analytics; keep the same definition and reporting period | Is the volume or origin of traffic changing? |
| Confirmed new enquiries | Inbox or CRM, excluding spam and duplicates | Does the business actually receive contacts? |
| Qualified enquiries | Business review against written criteria | Are the right customers getting in touch? |
| Time to first response | Enquiry receipt and first handling timestamps | Is the business's response process the problem? |
| Proposals and won projects | CRM; separate stages, not a sum of clicks | What happens after initial contact? |
| Measurement notes | Consent changes, incidents, campaigns and website releases | Are reporting periods genuinely comparable? |
If you calculate a rate, write down its numerator and denominator. The proportion of sessions containing an enquiry event is not the same as all enquiries divided by users. Comparing confirmed contacts from every channel with incomplete web sessions does not produce a clean website conversion rate either. Two clear numbers are better than one misleading percentage.
Trigger the success event after success
GA4 documentation provides generate_lead for submitting a form or requesting information. That does not mean every implementation records it automatically or correctly. Agree with your provider exactly what triggers the event. Clicking “Send” should be distinguished from the server accepting the enquiry, especially when the form works without reloading the page.
Server acceptance is still not proof of email delivery or lead qualification. Check the inbox or CRM separately. Exclude events following validation errors, and ensure that refreshing a thank-you page does not create another enquiry. If two tools send the same event in parallel, remove the duplicate trigger.
Test in a test environment first. A production test enquiry needs the owner's agreement and a clear label so it does not trigger real customer handling. Use an appropriate enquiry identifier to detect duplicates in the operational system; do not send form content, names, email addresses or phone numbers to standard analytics.
Phone calls, messages and acquisition sources
Call a tel: link click a phone-number click, not a conversation. Likewise, opening a messaging application does not confirm that a message was sent. Without an additional, appropriately implemented call-measurement system, check received contacts inside the business. Asking “How did you find us?” can supplement the register, but the answer is not an infallible contact history either.
Use consistent UTM labels for campaign links: source, medium and campaign. Do not put personal information in URLs. Do not add UTMs to ordinary internal links on your own website, as this complicates acquisition interpretation. If several people manage marketing, use a short naming dictionary rather than multiple versions of the same channel.
Attribution is a model for assigning credit, not a complete record of a person's motivation. A customer may see an advertisement, return through search and call from another device. Retain an “unknown source” category. Do not force it into other channels just to make the table's shares appear complete.
Read the result and choose the next change
During the review, ask in sequence: is measurement working, is the traffic relevant, is contact accessible, and does the business respond? If there are visits but no enquiries, start with a technical form and phone check. Then assess whether the offer matches visitors' needs. The guide to visits without enquiries can help.
If enquiry volume grows while quality falls, review advertising copy, service scope and qualification rules. If qualified contacts do not lead to proposals, inspect response time and handling. Do not immediately commission a website rebuild. The same visible outcome can originate in advertising, measurement or business organisation.
With few enquiries, one additional contact can substantially change percentages. Compare similar periods and consider seasonality. Record the date of each change; do not credit an improvement solely to a new button if a campaign launched at the same time. One sensible hypothesis and a planned review date are better than changing the entire site every week.
Limitations that belong next to the chart
Rejected consent, browser blocking and contact outside the website can limit measurement. Differences between analytics and the CRM do not necessarily mean a lead was lost. Mark incomplete periods after an incident and do not compare them with a full month without explanation. Work with a specialist to match privacy requirements to the implementation; this guide is not legal advice.
Before releasing a new website version, use the website launch checklist. After launch, maintain a shared contact register and a clear definition for every field. That is enough to begin discussing real enquiries instead of judging a website only by its visit count.
Sources
Event foundations: GA4 recommended events and Google Analytics event reference. Also see Search Console performance reporting and avoiding personally identifiable information in Analytics. Dashboard interpretation and register organisation are practical recommendations, not official performance thresholds or promises of results.
Planning a new website or a redesign?
Describe your project in the brief. It is not a substitute for technical acceptance or an analytics audit. The form is in Polish.
Open the website brief (in Polish) ↗