Concept work · prepared for Neal Baird, Vision Infrastructure Group Services
Website audit and GTM plan for vigs.com.au
Prepared by Kishan Jotangiya · 2 Nov 2026 · Real Lighthouse tests (mobile and desktop) plus a manual review of every page. I audited the public site only. I have not seen your ad accounts, CRM or analytics.
Kishan JotangiyaFounder of AdBuddy · design, web and AI-assisted lead systems · Surat, India, working remotely
1. The short version
- The site is clean, honest and fast on desktop. SEO basics and accessibility are good.
- The biggest gap is not the design. It is measurement. The only tracking on the site is Microsoft Clarity. There is no GA4, no Tag Manager and no ad pixel, so you cannot yet see which channel produces which lead.
- The site is built for one audience: trade owners who might hand their back-office to VIGS. A GTM plan for the home-services side (customers who book a job) needs its own offer, pages and tracking. This site does not do that yet.
- On mobile, where most trade owners will land, the page takes about 6 seconds to show its main content. That is the first thing to fix.
2. What I measured
66Performance, mobile
95Performance, desktop
96Accessibility
100SEO
77Best practices
| Metric (homepage) | Mobile | Desktop |
|---|---|---|
| First Contentful Paint | 3.2 s | 0.9 s |
| Largest Contentful Paint | 6.2 s | 1.4 s |
| Speed Index | 6.9 s | 1.1 s |
| Total Blocking Time | 30 ms | 0 ms |
| Layout shift | 0 | 0.002 |
Lab tests run from my machine on a simulated mobile connection. Real-user numbers from Google Search Console would confirm them.
3. Findings
What is working
- One clear H1 on each page, a meta description, a canonical tag, a sitemap and a robots.txt. All 17 images have alt text.
- A direct promise: "Run the tools. We'll run the paperwork." The three-step "Have a chat → Try it out → Grow" path lowers the pressure to enquire.
- Real proof: the eight group businesses (Primeline Roofing, SCR Concrete, 247 Floors and others) are named on the homepage.
- Good security basics: HTTPS with HSTS, no content sniffing, strict referrer policy. Sitemap lists 5 pages.
What to fix, in priority order
| # | Finding | Why it matters | Fix |
|---|---|---|---|
| 1 | No attribution tracking. Only Microsoft Clarity is installed. No GA4, GTM, Meta pixel or Google Ads tag. | You cannot tie spend to leads. Ads cannot optimise for form submissions. | GTM + GA4 + form-submit event + UTM and click-ID capture into the CRM (section 4). |
| 2 | Slow on mobile (LCP 6.2 s). The page is a JavaScript app, so nothing shows until the script loads. The form script from a third-party domain also loads on every page. | A slow first screen loses people before they read anything. This is the usual cause of "good clicks, few forms". | Load the form script only when needed, preload the hero image and fonts, trim about 86 KiB of unused JavaScript, and consider pre-rendering pages. |
| 3 | Raw HTML is the same for every page. About, Contact, Privacy and Terms all return the homepage title and description before JavaScript runs. | Google usually renders JavaScript, but other crawlers and AI search tools often do not. Each page may be read as a copy of the homepage. | Pre-render, or serve per-page titles and descriptions from the server. |
| 4 | The form is short but asks for the wrong things. Name, Business, Trade, Suburb, Phone. No email, no "what do you need help with", no consent line. Submit goes through an embedded GoHighLevel-style form. | You cannot qualify a lead or follow up by email. Phone-only leads are easy to lose. | Add email and a one-line need selector. Pass source, campaign and click ID into hidden fields. Confirm the CRM sends an instant reply. |
| 5 | No phone number or email on the site, no click-to-call. | Trade owners often prefer to call. One path (the form) fits only some visitors. | Add a click-to-call button on mobile and track taps as a conversion. |
| 6 | Third-party cookies (9 found) and a few low-contrast labels. Best Practices is 77 because of these. | Browsers are restricting third-party cookies. Contrast lowers readability for some users. | Review which scripts set cookies and add a cookie notice if needed. Darken the small grey labels. |
| 7 | No structured data (schema). No local business or organisation markup, and no pages for individual trades or locations. | Weaker local search and AI-search visibility. Eight trades sit on one page. | Add Organization schema now. Add one page per trade, plus NSW area pages later. |
| 8 | The homepage speaks only to trade owners, not to the households who hire the trades. | The home-services lead engine has no landing page of its own. | One landing page per trade brand (see section 4). |
Not checked, because I have no access: ad accounts, CRM, Google Business Profiles, call tracking, email sequences, real-user speed data. I would audit these in week 1.
4. Proposed GTM plan
Goal: one repeatable lead engine for the first home-services brand, written down so your doers can copy it to the next brand. I would plan, review and coach. Your team and AI engineer build.
Phase 1 · Audit and baseline (week 1)
- Review current ads, CRM pipeline, follow-up speed, Google Business Profiles and phone handling.
- Choose the first brand with you. A good test is a service with high ticket value and clear demand, such as solar, HVAC or roofing.
- Agree 3 numbers to track: cost per qualified lead, lead-to-booked-job rate, cost per won job.
Phase 2 · Tracking and attribution (weeks 1–2)
- Tool stack, kept small: Google Tag Manager, GA4, CRM source fields (source, campaign, UTM, GCLID/fbclid), Looker Studio dashboard. Add call tracking such as CallRail only if calls are a big share of leads.
- Why this stack: it is free or low cost, you own the data, and it works with the CRM you already use. Paid attribution suites cost more and only help once spend is large.
- Send CRM outcomes (qualified, booked, won, revenue) back to Meta and Google so the ad platforms learn from real jobs, not just clicks.
Phase 3 · Offer, landing page and funnel (weeks 2–4)
- One landing page for the first brand: one offer, local proof, a short form, click-to-call, fast on mobile.
- Diagnose any drop-off using the funnel: ad click → page view → form start → form submit → CRM lead → booked. Check speed, message match with the ad, and form errors.
- Instant response: text and email within 1 minute of a lead, plus a missed-call text-back.
Phase 4 · Channels (weeks 3–6)
- Google Ads search on high-intent local terms, Meta lead ads for retargeting and offers, and local SEO (Google Business Profile, service-area pages, reviews).
- Weekly review: spend, leads, bookings, cost per won job. Cut what does not produce jobs.
Phase 5 · AI and automation, with your engineer
- Use Claude to draft ad copy and test variants, summarise lead notes and flag leads that need a fast call.
- I would write the specs and test cases for lead scoring, routing and reporting agents. Your AI engineer builds them. I would not claim large-scale Grok experience. I would test where it helps before relying on it.
Phase 6 · Playbook and rollout
- Written playbook: offer template, page template, tracking checklist, ad structure, follow-up scripts, weekly report. One brand at a time, in the order you choose.
5. How I would work and what it costs
Rate: $25 per hour. Hours are tracked on Upwork, so you only pay for time worked.
| Stage | Estimated hours | Estimated cost |
|---|---|---|
| Phase 1–2: audit, baseline, tracking plan | 12–16 | $300–$400 |
| Phase 3: first landing page and funnel plan | 10–14 | $250–$350 |
| Phase 4–5: channel plan, AI workflow specs | 10–16 | $250–$400 |
| Phase 6: playbook for rollout | 6–10 | $150–$250 |
| First brand, planned end to end | 38–56 | $950–$1,400 |
These are estimates, not a fixed quote. The ongoing weekly review afterwards is about 3–5 hours a week.
Availability: I am available Monday to Friday, 11:00 AM to 5:00 PM Sydney time. Calls and weekly reviews happen in that window, and I plan to work at least 6 hours a day when the engagement is full time.
6. About me
I run AdBuddy, a digital, creative and AI systems business. I have 14 years in design and digital work: brand, websites, social, motion, SEO and lead-generation systems that connect landing pages, CRM workflows, automation and follow-up. I like turning a business goal into a clear plan that other people can carry out, then checking it against real numbers.
14years in design
47Upwork jobs
100%Job Success
5.0from 26 reviews
2,600+hours worked
- Door-to-door sales marketing: I have worked on Upwork with Sam Taggart of D2D Experts, a door-to-door sales business. That is close to your model, where jobs are sold to households and need to be tracked back to the source. [KISHAN: add one true line on what you made for them, if you want.]
- Meta Ads and lead generation: hands-on Meta Ads work, and live sites I built and run: PixPause and SwapSages.
- AI tools: Claude Code with VS Code is my daily tool for building and improving web projects. I would write the specs and tests for agent workflows with your AI engineer. I have not used Grok at scale, so I would test where it helps before relying on it. I build landing pages with Claude Code rather than Lovable.
Upwork profile and reviews · Contra · Portfolio
7. Questions that change the plan
- Who are your "doers" (in-house team, the AI engineer, an agency)? Who built vigs.com.au and runs the digital marketing agency whose form script the site loads?
- Which home-services brand goes first, and which CRM and ad accounts exist today?
- Is the goal leads for the trade brands, or also new trade businesses joining VIGS? Those are two different plans.