The problem: heavy HTML and uncertain lead measurements
Our marketing website combined page layout and large image data inside the homepage HTML. Separately, the analytics implementation risked counting both an automatic and manual Google Analytics 4 (GA4) pageview. Some internal funnel summaries also treated browser-generated assessment-completion events as equivalent to submissions stored by the server.
Those are three different problems. HTML transfer size affects how much markup a browser must download before it can parse the page. Duplicate analytics events distort interpretation. An assessment-start event or a browser-generated completion message is not proof of a saved, qualified customer inquiry.
Change 1: separate image delivery from homepage HTML
We preserved the original two homepage JPEGs and changed the markup to reference them as separate, publicly retrievable image files. This did not remove the photographs or replace them with lower-quality pictures.
| Measurement | Before | After |
|---|---|---|
| Homepage HTML response | 234,240 bytes | 29,408 bytes |
| Embedded base64 JPEGs in HTML | Two | Zero |
| Individual photo URLs | Not used for these images | Two JPEG files |
Result: the homepage HTML response became about 87% smaller. The separately delivered image files were 96,703 and 56,920 bytes. They can now be requested and cached independently of the markup where supported. This change does not establish a measured improvement to total transferred bytes, Largest Contentful Paint, Core Web Vitals or user conversion rate. Those require separate browser performance testing.
Change 2: count one intended GA4 pageview per loaded document
A GA4 configuration can automatically send a pageview when it loads. If a site also sends its own `page_view` event, both may be counted. We configured the website to suppress the default automatic pageview and retain one intentional, session-associated event. We also removed code that automatically labeled every event as external traffic regardless of whether it came from an owner, QA test or genuine prospect.
Why this matters: credible decisions require consistent definitions. Session starts, pageviews, button clicks, assessment starts and customers are distinct measurements; the numbers must not be presented as interchangeable. Internal testing still requires deliberate filtering or reconciliation; removing a misleading tag does not itself exclude all internal traffic.
Change 3: distinguish a real assessment submission from a browser event
Our form asks for contact consent, business information, service needs and marketing goals before an owner considers a private offer. The server creates a saved lead and records an assessment_submit event. Browser-side assessment_complete telemetry remains useful for diagnosing the user experience, but the reporting logic now counts server-recorded submissions for confirmed completions.
That prevents a client-side signal from being treated by the internal dashboard as proof that the lead was stored. Internal QA lead records are excluded from the business-conversion summary. An inquiry is still not automatically a qualified prospect or paying customer; the owner needs evidence of the next stage.
Change 4: protect the human approval and follow-up boundary
The marketing assessment is a qualification step, not an instant-price or payment form. An owner-approved offer, agreement and payment handoff are separate steps. The related automation implementation was subsequently given an authenticated daily recovery schedule for pending approved offers and eligible three-day reminders, with test-lead and completed-payment suppression. That work does not authorize arbitrary marketing emails or remove a customer's ability to decline further contact.
For the design principles behind this, see our marketing automation for service businesses guide.
How we checked the changes
- We confirmed the deployed public homepage and both separate image URLs returned HTTP 200, with valid JPEG responses.
- Agency search-identity checks verified the business is a digital marketing agency, not a construction contractor.
- The tracked repository changes passed 28 existing public-frontend checks and targeted attribution/integrity tests. The follow-up automation update also passed its backend test suite.
- We checked that the privacy-sensitive assessment remained outside the public search sitemap and retained its contact-consent requirement.
These checks establish technical behavior within their scope. They are not an external accessibility certification, a Google ranking guarantee or proof that a new client signed up.
What a service-business owner can learn from this
Before investing in another landing page or advertising campaign, ask your marketing provider:
- Which part of website performance has actually been measured: HTML weight, total transferred bytes, real-user Core Web Vitals, or all three?
- Does the analytics configuration avoid duplicate pageviews, and can it distinguish internal tests from customer activity?
- Are completed inquiry counts based on server-confirmed records rather than untrusted button clicks?
- Is the handoff from lead capture to the responsible person documented, and can a failed notification be detected and safely retried?
- What evidence is available for Google Search impressions, qualified inquiries, proposals and paying customers—without mixing these stages?
A credible report should show sources, time windows and exclusions. If the data is missing, it should say so.
What remains to be proved
Organic visibility for our Marketing Firm is still developing. The site's Search Console property had limited query-level evidence in the initial reporting window; sitemap submission and successful HTTP responses do not establish that every page is indexed or ranking. GA4 recorded traffic and assessment-start events, but those do not establish new external customers. We will measure that separately rather than presenting estimates as actual sales.
See the service-business SEO audit checklist for the process of reviewing search queries, indexing, site health and qualified inquiries over time.
Want a careful review of your website and lead system?
Betts Designs Marketing Firm is a U.S. digital marketing agency offering website and conversion improvement, local SEO, lead generation, and owner-controlled marketing automation. We work with businesses across the country; remodeling companies are one example of a client industry, not the services this marketing firm performs.
Tell us what you currently use, where inquiries are getting lost and what kind of customers you want to attract. The next step is an owner-reviewed assessment, not an automatic charge.
Request a marketing assessment →