Website concepts · prepared for Jimmy at Dirtcon
Each one is a full homepage, built from your own brand, your own service list and your own photographs. Open them on your phone as well as a laptop. Underneath the three there is a point by point look at what changes against your current site, and a real search check run against it.
Your logo is a chrome excavator arm on black. It has no colour in it at all, and neither does your current website, which runs on matte black #0A0A0A, chrome #C8CCD1 and steel #6E7277. Those exact values were read out of your own stylesheet rather than picked by eye, and all three concepts use them.
That leaves one real decision. Concept 01 introduces no colour whatsoever and stays as your brand is today. Concept 02 proposes a surveyor orange and concept 03 proposes a clay amber sampled from the ground in your own excavation photographs. Both are proposals, both are used only on markers and confirmations rather than spread across the page, and either can be dropped. That choice is probably the most useful thing on this page.
Same business, same facts, three ways of presenting them.
01 Brand flagship
Your own palette exactly as it is, and nothing else. Matte black, chrome and steel, read straight out of your current stylesheet, with your photographs full bleed and very large type. No colour is introduced at all, because your logo does not have any.
02 Qualifier funnel
The opposite direction. Paper rather than black, hairline rules, numbered sections and a spec table in the first screen. It reads like a job sheet, and it puts the access, the plant and the method in front of a builder immediately.
03 Live experience
Dark like the first, but warm rather than cold, and organised as a descent. Surface, cut, retain, crush, finish, with a depth scale filling down the page as you scroll. The sequence a site actually runs in, not a list of services.
So you can tell at a glance what has been taken from your own material and what has deliberately been left out.
The enquiry forms, the hero questions and the chat on all three are demonstrations. They log to the browser and send nowhere. Every concept ships hidden from search.
Side by side
Compared against www.dirtconau.com as it stood on 6 September 2026
I went through your current site properly before I built anything, so the concepts are not a fresh coat of paint. Every point below is a real difference between what is on your site today and what is on those three pages above.
The search side of it is measured separately in the check underneath this.
Nine of the ten pages your own sitemap gives Google return a 404. About, Fleet, Gallery, Contact, all five service pages, the privacy policy and the terms are all missing. Clicking About in your own header lands on a grey Page not found. So does the Get Your Quote button in the top right, because it points at /contact. The only page that exists is the homepage.
Everything in the menu on all three is a real section on the page, so nothing can 404. Get a quote goes straight to the form.
It is the most prominent thing in your header and it is the one link a ready customer presses. It points at /contact, and /contact does not exist. The only way to enquire is to scroll most of the way down the homepage and find the form.
The main button on all three scrolls to a form that is on the same page, and there is a second form in the first screen with your logo on it. On a phone the number in the header is one tap to call.
The homepage shows seven images and every single one is captioned Project Showcase. Not one suburb, not one client, not one date and not one job size appears anywhere on the site, so a builder cannot tell whether you have done work like theirs.
The same photographs are described by what is actually in the frame, shotcrete walls, soldier piles and timber lagging, a machine a metre off a house wall, a dig under transmission lines. The suburbs you have posted from, Balwyn North, Canterbury, Diamond Creek, Doreen and Eltham, are listed separately and are deliberately not matched to individual photos.
The Facebook, Instagram and LinkedIn icons in your footer link to facebook.com, instagram.com and linkedin.com. Those are the platforms' own home pages, not yours. Your real Instagram, dirtcon.au, has 1,127 followers and your best work on it, and your website does not point at it once.
Instagram is linked properly on all three, in the contact block and in the footer, going to dirtcon.au. Facebook and LinkedIn are left off rather than linked to nothing.
Outside the one form there is no way to ask anything. Someone wondering whether a machine fits through their gate, what happens to the spoil, or roughly how it works has to fill in the form and wait.
A chat on all three answers the twelve questions an excavation contractor actually gets asked, then takes a name, a mobile and an email. It never invents a price, a rating or a certification. On the concepts it is a demonstration and nothing is sent.
Two of the images on your homepage are bought iStock stock photographs, sitting among your real job photos. They are served from the site builder's stock library.
Nine of your own photographs and no stock at all. Where a page needed something behind it, it is drawn artwork in your own black and chrome rather than a bought photo.
You hand out jimmy@dirtcon.com.au. Anyone who types dirtcon.com.au into a browser gets a security warning and then an empty error page, because that domain has nothing on it and no working certificate. Your actual website is on a different address, dirtconau.com, which nobody would guess from your email.
The concepts are all on one address. Which domain the real site should live on is a decision for you, and it is worth making it before anything else is built.
At phone width the Melbourne, Victoria wide label prints on top of the Commercial and Civil Excavation Specialists line in your hero, so two pieces of text overlap. The paragraph underneath sits straight on the photograph with nothing behind it, which makes it hard to read in daylight.
All three were built to the phone standard and checked at 375 and 390 pixels. Nothing overlaps, every form field is large enough that an iPhone does not zoom in, and no text a visitor reads is under 12 pixels.
There is no Google Business Profile for Dirtcon at all, so there is no rating, no reviews, no map pin, no hours and no photos in the place every customer looks first. Your ServiceSeeking profile is set to hidden, so whatever is behind it cannot be seen either.
The concepts say this plainly rather than filling the space with wording nobody wrote. Getting a Google listing up is the single highest value thing on the list, and it is what a review would then have somewhere to live.
There is no structured data on your homepage at all, so nothing on the page tells Google or an AI assistant that you are a business, what you do or where you work. It is the single biggest reason you are not quoted in an AI answer.
The full stack goes on when the direction you pick gets built, the business itself, your services, your service area and your reviews, all linked together. The concept pages are demonstrations and are deliberately held out of search, so they do not carry it yet.
Your site has no llms file and does not welcome the AI crawlers by name. ChatGPT, Claude, Perplexity and Google AI each read a site with their own crawler, so today they largely go past you.
Every build ships an llms file that describes the business in its own words with a question and answer section, and a crawler list that names every current AI crawler as allowed. Detail is in the check below.
Nothing on your site is published as question and answer data. Wherever you do answer a question, it is text on a page and nothing more, so an assistant cannot lift the answer and hand it to the person who asked.
The concepts answer them in a question and answer block. It is the strongest single signal there is for being quoted by ChatGPT, Claude, Perplexity and Google AI overviews.
Your reviews are not published in a form a machine can read, so they cannot show as stars in a result or be quoted to someone asking for a recommendation.
The built site publishes them with review data attached, so the same reviews you already have start doing work in search as well as on the page.
When your link is sent in a message or posted to Facebook it shows no preview card, just the address.
The concepts carry a share image, so the link arrives looking like your business.
Every line above was read off the pages themselves on 6 September 2026, both yours and the three concepts, so you can check any of it while we talk. The search and AI search side is measured in the check below.
Your website today
Checked www.dirtconau.com on 6 September 2026
Before any of the concepts existed I ran your current website through the same check I run on every Scalus site. It scored 62 out of 100 for how ready it is for Google, and 0 out of 100 for how ready it is for AI search.
That second number is the one worth talking about. It is not a measure of how the site looks. It is whether ChatGPT, Claude, Perplexity and Google AI overviews can read the site, understand what Dirtcon does, and quote it when somebody asks them for a recommendation. At 0 out of 100 the answer today is mostly no.
In total 10 items came back needing attention. They are listed below in the order I would fix them, and every one of them is included in a Scalus build.
Without a canonical tag Google can treat two versions of the same page as two competing pages.
What I would do: Add a canonical tag to each page.
When someone shares your site in a message or on Facebook it will not show a preview card.
What I would do: Add social share tags with a real image of your work.
This is the big one. There is no structured data on the site, so nothing on the page tells a machine that you are a business, what you do or where you are. It is the single largest reason AI search cannot quote you.
What I would do: Add the full structured data stack for the business, the services and the reviews.
The AI crawlers are not welcomed by name. ChatGPT, Claude, Perplexity and Google AI each read the site with their own crawler, and if they are not named they may not read you at all.
What I would do: Name every current AI crawler as allowed.
There is no llms file. This is the file AI models read to understand a business in its own words, and almost no local business has one yet.
What I would do: Add an llms file with a summary, your services, your area and an FAQ.
The description is the grey text under that headline. It is what makes someone click you instead of the next result.
What I would do: Write a description for every page that says what you do and where.
The tags that tell Google it may show your images large and quote more of your text are missing.
What I would do: Add the full robots meta directives so your listing can show more.
Shared links will not render a large preview image on some platforms.
What I would do: Add the summary card tags.
Some standard security headers are missing. Not a ranking issue, but it is the sort of thing that gets a site flagged.
What I would do: Add the standard security headers.
The page code has validity errors, which can quietly break how a crawler reads it.
What I would do: Clean up the markup errors.
Checked on 6 September 2026 with the same free checker at scalus.app/seo-geo-checker. Scores move as a site changes, so this is a snapshot of that day, not a permanent grade.
