Why your real estate website enquiry counts do not match

Your website report says there were twelve form submissions. Your enquiry inbox contains nine records. The sales team says five are genuine new conversations. Those numbers can all describe different things. Before calling the difference a tracking failure, write down what each number actually counts.

The figures below are fictional. They show how to investigate a mismatch, not HomePage.Studio’s results or an expected conversion rate for an agent.

Give each count a clear meaning

A button click shows that someone activated a control. It does not, by itself, establish that a form was accepted, an email arrived or a CRM record was created.

Google Analytics distinguishes form interaction events from lead outcomes. Its enhanced measurement documentation describes form_start and form_submit. Its recommended events include generate_lead and qualify_lead. Naming an event does not make it a verified business outcome. Ask your provider what triggers it on your actual form.

A saved enquiry is a different record again. It may contain a test, a repeat request or an unwanted sales pitch. A qualified conversation requires an agreed human or business definition; it is not simply another name for every form submission.

A fictional weekly reconciliation

Imagine one agent reviews a single seller-enquiry form for a seven-day period. The website and enquiry reports use the same start and end times and the same time zone. No buyer, valuation-tool or open-house forms are included.

• 12 recorded form events: the analytics report’s starting total. Their trigger and repeat-event behaviour still need checking.
• 9 saved enquiry records: the form system’s starting total. These are records, not necessarily nine different people.
• 2 authorized test records: identified from the team’s test log, not guessed from the message.
• 1 repeat request: reviewed as a second record relating to an already-counted enquiry. The original information is retained.
• 1 unwanted sales pitch: classified after review, not because of the sender’s name or email provider.
• 5 remaining records: nine minus two tests, one repeat and one pitch. They are candidates for follow-up, not automatically five qualified sellers.

That explains the difference between nine saved records and five records for review. It does not explain why the analytics report shows twelve events. Do not assign those three extra events to duplicates or tests without evidence linking the systems’ definitions.

If the five remaining records have not been assessed, write “qualification unknown.” If some people have already contacted the agent through another channel, “new conversation” may also be the wrong label. Keep the count’s meaning narrower than the evidence.

Copy this count-reconciliation worksheet

Use one copy for one form and reporting period. Keep it to aggregate figures and definitions; do not paste names, email addresses or enquiry messages into a shared report.

Scope: Form/page: ___ · Period start/end: ___ · Time zone: ___

Analytics: Report/event name: ___ · Exact trigger: ___ · Count: ___ · Known exclusions or consent limits: ___

Form records: Storage/report: ___ · Accepted-record definition: ___ · Count: ___ · Failed or pending records checked: ___

Review: Confirmed tests: ___ · Confirmed repeats: ___ · Reviewed spam/pitches: ___ · Remaining records: ___

Business outcome: Qualification definition: ___ · Verified count: ___ · Unknown/unreviewed: ___

Unresolved difference: ___ · Evidence needed: ___ · Person responsible: ___ · Next review date: ___

Use “unknown” rather than zero where a system has not supplied an answer. Only subtract categories that do not overlap: a repeated test must not be removed once as a test and again as a repeat. Keep the original totals alongside the reviewed figures.

Investigate the remaining difference

First compare the reporting period, time zone, form identity and filters. A report for all website forms will not reconcile cleanly with an inbox folder for seller enquiries alone. Some systems also record or display an item later than the initial visitor action.

Then ask the provider to demonstrate the event trigger and the accepted-record path with a controlled, authorized test. Check unsuccessful validation as well as an accepted request. The website-to-CRM connection checklist helps separate the page response from confirmed receipt at the destination.

Browser analytics may miss activity when visitors do not consent or measurement is blocked. It may also count an interaction more than once if the implementation triggers repeatedly. Those are possible explanations, not a diagnosis of your website. A lower analytics total alone does not prove that an enquiry was lost; a higher total does not prove that the agent received more leads.

Do not disable consent choices to make reports match. Do not send personal enquiry content into analytics to help join the totals. Agree a proportionate, privacy-aware investigation with the people responsible for the systems.

Report what you know, not a neat-looking percentage

In the fictional example, “nine saved records, five remaining after documented review, analytics difference unresolved” is a useful conclusion. “A 42% conversion rate” is not: dividing five reviewed records by twelve events mixes different units and says nothing reliable about visitors becoming qualified enquiries.

Choose a denominator that answers the actual business question, document its limits and keep the definition consistent between periods. If the necessary data is unavailable, say so. A transparent baseline is more useful than a precise number whose inputs do not match.

For help defining a reporting scope, see our SEO, AEO and marketing service. Bring the systems and definitions you currently use; this worksheet does not assume a new dashboard, a particular CRM or a fully connected attribution setup.