Analytics & Tracking: Project Report

Deland Gibson Insurance

July 30, 2026

Project Complete

Prepared by Bear Hair Dev Consulting

Executive Summary

This project is complete as of July 30, 2026. At the start, Google Analytics and the Meta Pixel were each installed twice, so every visit counted double. Google Ads had one working conversion (a single form on one landing page), and phone calls were not tracked at all. The setup was rebuilt: every tracking code now exists exactly once, managed in Google Tag Manager, with conversions for every lead form and for phone calls through CallRail. Form submissions, phone clicks, the pixel, and a live test call were each verified on the live site. Beginning with July 29 and going forward, web traffic data will be clean and accurate.

Tracking Codes: Where Each One Was

Tracking code	Where it was	What changed
Google Analytics (GA4)G-56L0N3KLX7	Installed twice: once by the Site Kit plugin's Google tag (GT-KDDGXZM), once inside the Tag Manager container. Every visit counted twice.	Site Kit removed. The Tag Manager copy is now the only install.
Google Tag ManagerGTM-T8P867B	On the site only because the Site Kit plugin injected it.	Reinstalled through the WPCode plugin. Now the single home for all tracking.
Meta Pixel770438159121279	Pasted in twice: hardcoded in the theme (functions.php) and pasted into WPCode's header box. Fired twice per page.	Both pasted copies removed. Rebuilt as a single Tag Manager tag.
Google Ads tagAW-586138399	Bundled inside Site Kit's Google tag. Visit data only, no working conversions beyond one form.	Signal now flows through the Google Analytics link. Conversion actions rebuilt (listed below).
CallRail	Did not exist. No call tracking of any kind.	New account created for this project, owned by Deland Gibson. Script installed in WPCode alongside Tag Manager. Visitor number pool live.
Microsoft Clarity (heatmaps)	Its own plugin, installed once.	Unchanged.

Changes Made

Removed

The duplicate Google Analytics install (Site Kit's GT-KDDGXZM tag) and the Site Kit plugin itself.
Both hand-pasted Meta Pixel copies (theme functions.php and WPCode's header box).
GA4 conversion events that counted non-conversions: first visits (first_visit), form-field clicks (form_start), file downloads (file_download), and two unreliable auto-detected form events (form_submit, gtm.formSubmit).
Two GA4 conversion events that never recorded any data (conversion_event_submit_lead_form, form_submission_event).
The 3-year-old "Form Submit Event" tag in Tag Manager (source of the gtm.formSubmit junk).
The Google Ads conversion action wired to a dead event (SUBMIT_LEAD_FORM), retired.

Added

Form submission conversions: a GA4 event (generate_lead) fires when any lead form is actually submitted, with the specific form identified. Primary conversion in Google Ads.
Phone click tracking: a GA4 event (phone_click) fires when a phone number is clicked. Secondary (observation only) in Google Ads.
CallRail call tracking: each visitor sees a unique number, so calls attribute to source, ad, and keyword. First-time calls of 15 seconds or more are a primary conversion in Google Ads; repeat calls are secondary.
Campaign tagging (UTMs) added account-wide to Google Ads URLs. Auto-tagging confirmed on.
Annotation in Google Analytics marking the July 29 changeover.

Kept

The personal insurance form conversion (the one event that already worked). Still a primary conversion in Google Ads.
Microsoft Clarity heatmaps, unchanged.

Analytics & User Behavior Insights

The conversion drop earlier this year was a tracking gap, not a campaign problem

Source: Google Ads and GA4 reviewHigh confidence

Observation: Ads moved from landing page LP1 to LP2 about three months ago. Only LP1's form was ever tracked, so reported conversions fell to nearly zero.
Interpretation: The ads kept producing; the measurement didn't follow the move. Every form is now tracked, so this can't recur silently.
Recommendation: Judge that period's campaigns by the corrected setup, not the reported numbers.

How to read historical traffic

Source: Google AnalyticsHigh confidence

Observation: Traffic before July 29 was double-counted.
Interpretation: Any comparison across the changeover will show a drop. That drop is the correction, not lost traffic.
Recommendation: Treat July 29 as the baseline for before-and-after comparisons.

Recommendations for Future Improvement

Deeper Meta measurement

Medium priority Effort: Needs scoping

Why it matters: The pixel reports accurate visit data, but the Meta ads have no conversion events and no campaign tagging.

Suggested next step: Define pixel conversions and add campaign tagging to the Meta ads. Needs Meta Business access.

Call recording

Low priority Effort: Small

Why it matters: Shows lead quality beyond call length. Massachusetts requires all-party consent, so it needs a consent decision first.

Suggested next step: Decide on a consent approach, then enable recording in CallRail.

Ongoing measurement refinements

Low priority Effort: Small

Why it matters: Tracking stays accurate only while it matches how the campaigns run.

Suggested next step: Loop us in when campaigns or landing pages change so conversions keep matching reality.

Appendix

Where each tracking source now lives, for reference. All accounts listed below are owned by Deland Gibson.

Tracking source	Where it runs
Google Tag Manager	Snippet installed through the WPCode plugin; the single home for tracking
Google Analytics (GA4)	Inside Google Tag Manager
Meta Pixel	Inside Google Tag Manager
Form and phone click tracking	Inside Google Tag Manager
CallRail	Script installed in WPCode alongside Tag Manager
Microsoft Clarity (heatmaps)	Its own plugin, unchanged

Nothing tracking-related lives in the theme’s code, so future design changes cannot silently break measurement.