100th Meridian Window Film: from one indexed page to the name AI gives first

Who is the client?

100th Meridian Window Film is a founder-run installer serving the Niagara region of Ontario: heat-control, UV, privacy, decorative, and security window film for homes and businesses. Joey, the founder, runs a consultative, no-pressure operation, and his old website was a five-page Wix template.

The business was real and the work was good; the web presence was not carrying its weight. When we ran our pre-sale analysis in July 2026, the pattern was one we see constantly with local businesses: a site search engines could barely enter and AI engines had no reason to cite. That analysis is the same onboarding audit we run for any prospect, and it set the baseline every number below is measured against.

Where did they start?

Near zero. The old site's own Wix analytics for the 12 months before relaunch show 366 sessions, 97 Google-organic sessions, and 9 form submissions for the entire year. 87% of visits landed on the homepage, because there was almost nowhere else to land.

Two details matter for reading everything that follows. First, Google Search Console and GA4 were only connected in July 2026, so the Wix dashboard is the honest historical baseline, and we recorded it before touching anything. Second, organic was already growing (+177% year over year on a tiny base), which told us demand existed in the market; the site just could not catch it.

What did we build?

A complete rebuild, launched July 11, 2026: a 137-URL static site with silo architecture covering every service and every service area, machine-readable from day one, with verified lead tracking so results could never be padded.

Nothing on that list is exotic. The result came from doing all of it, at once, correctly, and then measuring honestly. We wrote up the build decisions in detail in the companion post.

• Silo architecture: dedicated, interlinked pages for each service (heat control, privacy, security, decorative, UV) and each Niagara-area city, replacing a single homepage that had to do everything. This is the same structure behind our website builds.
• The machine-readable layer: per-page structured data (LocalBusiness, Service, FAQ), clean semantic headings, and an llms.txt file, our standard technical SEO work.
• Verified value tracking: a lead only counts when the server confirms a delivered inquiry email, and our own QA traffic is excluded by design. A phantom lead is worse than no lead.
• Performance: static pages tuned for Core Web Vitals, because engines crawl fast sites more willingly and people abandon slow ones.
• Launch-day plumbing: full sitemap submission to Google and Bing, IndexNow, and Google Business Profile alignment so every fact about the business matches everywhere AI checks.

What happened in classic search?

Google indexed 50 pages within 24 hours of launch and 122 by July 23; Bing reached 136. Search impressions ran to 4,704 in the first three weeks (the old site drew 97 organic visits in a year), and by August 3 the site ranked in Google's top three for 18 tracked queries.

The click number is small and we will not dress it up; 45 clicks is 45 clicks. What matters three weeks in is the shape underneath it: a site that effectively did not exist to Google was shown in results 4,704 times in 22 days and now surfaces for dozens of commercial queries ("100th meridian window film" and "window privacy film installation" already sit at position 1). Clicks follow impressions; we report both as they come.

What happened in AI search?

By August 2, 100th Meridian held 16.4% of all AI citations going to any named window film business in its market. That is the largest share of the 22 businesses cited, and the #1 spot on ChatGPT, Claude, Perplexity, and Google AI Mode simultaneously.

The launch-day row is the starting condition, not a harness run: the harness's first full run was July 20, and at launch the business had one indexed page and no citations to its name, so its share of the market's citations was zero. From that standing start it was #4 within nine days, and it has held #1 overall in every run since day 16.

On the August 2 run the site ranked #1 on ChatGPT (25.0% share), Claude (22.7%), Perplexity (15.2%), and Google AI Mode (14.3%), and #2 on Gemini (13.6%).

The crawler data explains why. Between July 11 and August 3 we verified 5,286 AI and search bot visits: 1,273 from Bingbot, 873 from ClaudeBot, 757 from ChatGPT-User (the live crawler ChatGPT sends while answering a user), 564 from Googlebot, 410 from GPTBot, 283 from OAI-SearchBot, and 254 from PerplexityBot. Every request is checked against the operator's published IP ranges; 2,830 requests that claimed to be AI bots and failed that check were thrown out, not counted.

And it is not only bots: GA4 recorded 39 human sessions arriving from ChatGPT since launch. People asked, ChatGPT answered with this client, and they clicked through.

Did any of it turn into business?

Six server-confirmed inquiries arrived in the first 24 days, against nine in the old site's entire prior year. Each one is a delivered email a real person sent; our tracking counts nothing else, and our own testing is excluded by design.

Alongside the six inquiries came three tracked calls from the site's phone links and one email click. We count these separately and more skeptically (a tapped phone number is a weaker signal than a delivered inquiry), which is exactly how we present them to the client in monthly reporting.

At the old site's pace, six inquiries took eight months. The honest framing is still the cautious one: 24 days is a short window, local demand is seasonal, and we will keep publishing the numbers as the record lengthens rather than annualizing three good weeks into a forecast.

What are the honest caveats?

Every result above is real, dated, and reproducible from the client's own accounts, and the window is 24 days: long enough to prove the mechanism worked, far too short to promise anyone else the same slope.

• This is one company in one market. Niagara window film is a winnable field; your market may be harder or easier. Existence proof, not forecast.
• Absolute click volume is still small (47 Google clicks in the first month). Indexing, rankings, and AI citations are the leading indicators at this stage; clicks and leads are the lagging ones.
• GA4 attributes most of the six inquiries to Direct rather than a search channel. We report what the analytics say instead of re-crediting traffic to flatter the story; some of those visitors likely heard the name in an AI answer and typed the address, but we cannot prove it, so we do not claim it.
• The AI share-of-voice figure comes from our own harness (method below), not a third-party panel. We publish the method so you can judge it, and we run the same 15 questions every week rather than picking flattering ones after the fact.
• The old site's baseline comes from Wix's own dashboard because GSC and GA4 were never connected before we arrived. It is the only 12-month record that exists, and we archived it before cutover.

How is all of this measured?

Four independent instruments, none of which we can quietly pad. The client's own Google Search Console and GA4, Bing Webmaster Tools, a bot tracker that verifies every claimed AI crawler against its operator's published IP ranges, and a weekly harness that asks the AI engines the same 15 buyer questions and records who gets cited.

The share-of-voice method in one paragraph: every week we put the same 15 questions a Niagara window film buyer would ask ("who installs privacy film in Niagara-on-the-Lake?") to ChatGPT, Claude, Perplexity, Gemini, and Google AI Mode, then log every business each answer names and links. A citation only counts when the engine links the business's own site. Share of voice is the client's fraction of all citations going to any named competitor in the market, so the number cannot be inflated by asking softer questions or counting vague mentions.

Lead counting is stricter still: a form submission counts only when the server confirms the inquiry email was actually delivered, phone taps are tracked as the weaker signal they are, and all our own QA visits carry a test flag that excludes them from every number on this page. This is the same reporting standard every Viziblty client gets monthly.