Your technical foundation is fine.
There is almost nothing standing on it.

This is a tier 2 audit. It is built entirely from what is publicly visible: a crawl of the site, its structured data, its robots and sitemap files, Google's own PageSpeed Insights API, and a public search index. We have no Search Console access, no Analytics access and no Google Business Profile access for Marketing Gardens, so this report carries no clicks, no impressions, no CTR, no sessions and no conversion figures. Any report that hands you those numbers without an account login invented them.

Everything below can be reproduced from outside. Where a check could not come back cleanly, the report says so instead of filling the gap.

Sources, named

• Crawl. A link-following crawl from the homepage, not a read of the sitemap. It reached 10 pages. The sitemap lists 11. The difference is /thank-you, which nothing on the site links to.
• Speed. Google PageSpeed Insights API, mobile, three samples, median reported. Not a local headless run, which reads several points low and swings by seconds between runs.
• Index coverage. A site: query against a Bing-backed public index, plus a targeted per-page query, plus a control query for a domain that does not exist.
• Rendering. The pages were opened in a real Chrome at a 390px mobile viewport and read from the accessibility tree, not inferred from raw HTML.
• Structured data. JSON-LD read directly from each page's source and parsed.

Word count is a crude measure and we do not treat it as a ranking factor. It is useful here as a proxy for something simpler: is there enough on the page for anyone, a search engine or a buyer, to work with. The answer across this site is no.

Gold bars are pages under 400 words. Median 300 words, whole site 3,532 words.

For scale: at 3,532 words the entire site holds about as much text as two 1,700-word articles. The blog that would hold those articles exists, is styled, is in the navigation, and is empty.

A page an index does not hold cannot rank in that index, so this check comes before any ranking question. Seven of the eleven URLs in the sitemap came back from a public site: query. The case study did not, and a targeted query for it confirmed the absence. Note the scope: this is a Bing-backed public index, not Google, because a cold audit has no Search Console access to ask Google directly.

The detector was proved both ways before this grid was drawn. A control query for a domain that does not exist returned nothing, and the live query returned seven URLs. A grid of absences from a broken checker is worthless, so we checked the checker.

The result to act on is the middle one. The deepest page on the site is the one confirmed absent, and every one of the seven pages the index does hold is shorter than it. Three more, the two legal pages and the thank-you page, were not returned either; we did not confirm those individually and they matter far less.

The first thing a visitor reads after the headline is a block of four figures. Three of them have nothing behind them anywhere on the site.

This is not a suggestion to delete them. It is a decision to make: either substantiate them with a named client, a dated case study and a stated method, or replace them with something you can show. A marketing firm's own homepage is the one place a buyer will test the numbers.

Speed is not your problem. Three PageSpeed Insights runs on mobile:

Median score 87, LCP 3.5s, CLS 0.001, TBT 47ms. The API itself varies by several points and by seconds of LCP between runs on an unchanged page, which is what the first sample is.

The more useful finding is what the API would not return. Google publishes field data, drawn from real Chrome visits, once a site has enough of them. For this domain it returns none. That is a floor, not a count: it says the traffic is below the threshold Google needs, independently of any analytics account.

Images, since they are the one speed item worth touching

Red is the eagerly loaded hero. All eight files are JPEG; none is a modern format.

Twelve findings, ordered by severity. Each carries a severity tag and one of four categories, and each states the evidence it rests on.

Everything in this report is yours to fix. We checked rather than assumed.

• The site is not WordPress. /wp-json/ returns 404, as do the WordPress sitemap paths. The pages are served as static HTML behind Cloudflare, so changes are a deploy, not a plugin.
• Cloudflare is in front of it. Security headers, caching and redirects can all be set there without touching the site itself.
• The contact form posts to forms.perfectcopier.com. A separate service you control, protected by Cloudflare Turnstile. Nothing about it blocks the fixes below.
• Analytics is installed and collecting. The data exists; this audit simply has no account access to it. Granting read access would let a follow-up report use your real numbers instead of public proxies.

There is no agency, platform or franchise template standing between you and any item in this report. Every fix is yours to make.

1. Publish the three blog posts that are already written as headlines. The titles, the categories and the bylines are on the page. They are the right questions. Three posts of 900 to 1,400 words each, every one opening with a direct answer in its first paragraph under a question-shaped heading, removes the worst signal on the site and creates the first content a search engine or an AI assistant can quote. Cost is writing time, nothing else.
2. Get the case study indexed. Link to it from the homepage and from the services page, submit the URL directly, and give it a paragraph a person would quote. It is already the best page you have.
3. Decide on the four homepage numbers. Substantiate or replace. One named client with a dated before-and-after beats four unsourced figures.
4. Fix the schema address reference and add geo, sameAs and an image. A ten-minute edit to one JSON-LD block.
5. Add noindex to /thank-you and drop it from the sitemap. Two lines.
6. Claim and fill the Google Business Profile, then build one Lynchburg page. This is the only route into the local pack, and it is free.
7. Thicken the four service sections into four real pages. Today they are anchors on one page. Four separate pages, each answering the pricing, comparison and problem questions for one service, is the structural change that makes the rest of the work compound.
8. Housekeeping: modern image formats, security headers at Cloudflare, a real founder photo, an email address. Each is small. Together they are the difference between a site that looks maintained and one that looks staged.