A form submission is an activity, not proof of a successful marketing result. It may be a genuine sales opportunity, an existing customer asking for help, a supplier message or spam. Useful measurement follows the enquiry far enough to learn whether it was relevant and what happened next.
Agree the definitions before installing more tracking
Marketing and sales should use the same small set of stages. A practical starting model is:
- Enquiry: a person has submitted a form, called, emailed or started another identifiable contact.
- Valid lead: the contact is genuine and relates to something the business could provide.
- Qualified opportunity: need, timing and fit justify a sales conversation or quotation.
- Won work: the customer has accepted the commercial agreement defined by the business.
- Revenue: money recognised under the business’s accounting rule, not merely a proposed deal value.
Write down who changes each stage and the evidence required. If “qualified” means something different to every salesperson, the reporting will never be stable.
Map every legitimate enquiry route
List forms, phone numbers, email addresses, booking tools, chat, WhatsApp and ecommerce checkout. Decide which routes can be tracked directly and which require a staff action. For example, a call may be captured by call-tracking software, but someone may still need to mark whether it concerned new work.
Each form should have a unique identifier and a clear success condition. Do not count a click on the submit button as a lead. Count the server-confirmed completion after validation and delivery have succeeded. Test that refreshing a confirmation page does not create another conversion.
Capture only attribution data you can use
For most service businesses, useful fields include landing page, referrer, campaign parameters, form name and submission time. Store the original source when a lead is first created, then keep later interactions separately. Overwriting the first source on every visit makes it impossible to distinguish discovery from the final return visit.
Campaign naming needs one documented convention. Decide how the team writes channel, source, campaign and creative values before links are distributed. “facebook”, “Facebook” and “fb-paid” becoming three rows is a process problem, not an analytics problem.
Connect the website event to the CRM record
Generate a submission or lead identifier on the server and pass it to the CRM with the enquiry. The same identifier can be attached to the analytics conversion without exposing the person’s name or email. That makes troubleshooting possible when totals do not match.
The integration should handle retries and duplicates. If the CRM is unavailable, keep the enquiry securely and retry; do not ask the customer to submit again. If the same retry arrives twice, the CRM should update the existing record rather than create two leads.
Build a weekly reconciliation
Compare four counts for the same period:
- Successful submissions recorded by the website.
- Notifications or integration deliveries accepted.
- Lead records created in the CRM.
- Conversions reported by the analytics or advertising platform.
The numbers will not always match exactly because of consent choices, blocked scripts and attribution windows. Large or sudden differences still deserve investigation. A website count higher than CRM records can indicate a failing integration; an advertising count higher than server completions can indicate a conversion firing too early.
Report quality alongside volume
A compact monthly view can show enquiries, valid leads, qualified opportunities, wins and revenue by source. Include the percentage moving from one stage to the next. A source generating fewer enquiries may be more valuable if those enquiries are consistently relevant.
Keep reasons for disqualification short and consistent: spam, service not offered, location not served, budget or timing mismatch, existing-customer support and unable to contact. Review the patterns. If many visitors request something the company does not provide, the advertising or page wording may be setting the wrong expectation.
Respect consent and data minimisation
Google Analytics policy says not to pass data that Google could recognise as personally identifiable information. That includes names, email addresses and personal phone numbers, which must not appear in event values, page URLs or titles sent to Analytics. Keep contact details in the systems that need them for the enquiry, and use a non-identifying server-generated lead reference for reconciliation.
Explain the purpose of the form and link to the relevant privacy information before submission. Consent requirements depend on the technology and visitor location. Use professional advice where the processing is complex or high risk.
A test to run every month
Submit one clearly labelled internal test through each important enquiry route. Confirm the customer-facing message, notification, CRM record, source data, consent record and analytics event. Then remove or exclude the test from commercial reports. This catches quiet failures that a dashboard cannot see.
Sources and further reading
- ICO direct marketing and electronic communications guidance — the UK regulator’s guidance for consent and direct-marketing responsibilities.
- Google Analytics event collection documentation — technical reference for event-based measurement.
- Google Analytics guidance on avoiding personally identifiable information — the policy reference for keeping names, email addresses and personal phone numbers out of Analytics data.
If your current forms need to feed a CRM or your reporting stops at button clicks, Xapner can review the journey and integration as a defined project.