I’m publishing a blog post every day for 100 days. Here is my day zero data, warts and all.

Most SEO case studies are written backwards. Someone gets a result, then reverse-engineers a tidy narrative that makes the result look inevitable. The failures never get written up. The half-successes get rounded upward. Nobody publishes the version where they were wrong.

I spent over a decade as an investigator at ASIC, and the habit that job leaves you with is simple: evidence is only worth anything if it was recorded before you knew what you wanted it to say. So I’m publishing my starting position first — the full, unflattering set of numbers — and then I’ll publish what actually happens to them.

Here is the deal. For the next 100 days I’m writing one blog post a day on this site. Websites, search, AI visibility, and what small businesses can actually do about all three. At the end I’ll publish the same numbers again and you can decide for yourself whether it was worth the hours.

Day zero: the numbers

The baseline below was captured on 28 and 29 July 2026. Every screenshot is dated, and nothing has been re-taken to look better since.

Those aren’t rounded-down numbers or a modest framing device. They are literal zeros. Here is the full set, with where each one came from.

The 404 is the one worth pausing on, because it is the single most common reason a new website goes nowhere. Google came to look on 8 July. The homepage returned a “not found” error. Google, reasonably, concluded there was nothing here and left. No amount of publishing fixes that — you have to fix the error first and ask Google to come back.

The one number that isn’t a zero

Analytics was already running on the domain, so unlike the search data I do have a traffic baseline. It is not a flattering one.

In the week to 28 July, the site recorded 13 sessions from 11 users. Six of those users were in Singapore, five in Australia. Average engagement time was five seconds. Across the entire week there was exactly one engaged session — and that was almost certainly me.

Six users in Singapore, not one of whom registered a single second of engagement, are not people considering hiring a web designer in Noosa. That is automated traffic, and it is what most brand-new websites’ analytics actually contains.

I’m spelling that out because “we had a few hundred visitors last month” is one of the most misleading sentences in small business marketing. If you can’t say how many of those visitors were human, in your country, and capable of becoming a customer, you don’t have a number. You have a decoration. For the next 100 days I’ll be reporting Australian traffic separately, with my own visits filtered out, for exactly that reason.

What I need to declare before we go further

This is not technically a brand-new domain, and it would be dishonest to run an experiment about starting from zero without saying so.

I registered catherineyork.com.au years ago, half-built something on it, and abandoned it. It carries 49 backlinks from 22 referring domains, and an authority score of 2 out of 100. That is a small head start I did not earn deliberately and have not yet audited. Anyone can verify it in about thirty seconds with a free Semrush account, so I would rather state it now than be corrected later.

Auditing those 22 domains is on my list, and if any of them turn out to be junk, that gets published too.

The part that’s worse than zero

Here’s what I didn’t expect to be writing on day one.

Search “catherine york digital” in a clean, signed-out browser and Google returns an AI Overview that reads, confidently and in full sentences, that Catherine York is a London-based voiceover artist, presenter and communications coach. It lists her services. It cites her LinkedIn. Google even strikes the word “digital” out of my query and offers to show me results without it.

She is a real person with a real, long-established business, and she has every right to that name. But it means I am not starting at zero. I am starting behind. There is an entity called “Catherine York” that Google’s Knowledge Graph already understands, and it isn’t me. That same entity is baked into the training data of every large language model that has read the open web.

A new website isn’t competing against silence. It’s competing against whatever the internet already decided your name means.

This is the problem almost nobody writes about, and it applies to far more businesses than you’d think. Shared names. A competitor in the same category one suburb over. A previous owner of your domain. A franchise you used to trade under. Search engines and AI models both need to resolve you to a distinct entity before they can recommend you, and if something else already occupies that slot, you have to build enough independent signal to earn your own.

I would rather test that in public on my own site than learn it on a client’s.

The rules of the experiment

An experiment with no constraints isn’t an experiment, it’s just working hard and hoping. So these are the boundaries, and I’ll hold to them.

The one-page constraint for the first month costs me something, and I’m doing it on purpose. Conventional advice — advice I give clients and stand behind — is that one page can’t rank for more than one thing, and a consultancy needs proper service pages. That’s still true. But if I build everything at once, I’ll never know which part moved the needle. Thirty days of restraint buys a month of clean data. After that, commercial sense wins.

Why your new website isn’t showing on Google either

If you’ve launched a site recently and can’t find it, the cause is almost always one of four things, and none of them require an agency to diagnose.

1. Google literally cannot reach it

A 404, a redirect loop, a mismatch between the http and https versions of your address, or a “discourage search engines” checkbox left ticked in WordPress. Mine was a 404, live for at least three weeks. Check it yourself in Google Search Console using the URL Inspection tool and the Test Live URL button.

2. Google doesn’t know it exists yet

Nothing links to a brand-new domain, so there’s no trail for a crawler to follow. Verify the site in Google Search Console and Bing Webmaster Tools, submit an XML sitemap to both, and request indexing manually for the pages that matter.

3. There is nothing on it to rank

A homepage with a logo, a photo and a contact form matches almost no search query. Search engines need text that corresponds to what people actually type. One page can realistically compete for one topic, and only one.

4. Something else already owns your name

The problem above. If a larger, older, better-linked entity shares your business name, you will not outrank it by existing. You outrank it by accumulating distinct signals — consistent business details across the web, a complete Google Business Profile, real reviews, links from businesses in your industry, and content that unmistakably belongs to you and nobody else.

What I expect to happen

Predictions made after the fact are worthless, so here are mine in advance, specific enough to be wrong.

My honest expectation is that the traffic numbers will look underwhelming at day 100 and that AI visibility will move faster than the Google rankings. There is far less competition for being the source an AI assistant cites, and the signals that earn it are cheaper to build. If that turns out to be right, it changes what I would advise a small business to spend its money on.

If it turns out to be wrong, that’s a more useful post than the one where I’m right.