The bigger the market, the less you exist: 697 businesses measured across seven countries

We measured 697 real businesses - not a sample built for the occasion, but the companies we were about to contact. Seven countries, nine trades, one question each, asked in the language their own customers would use: what do the assistants answer when someone looks for this trade in this city?

621 out of 697 appear nowhere. That is 89%, and the figure no longer surprises us. What stopped us was how it is distributed.

The same trade, two countries, two worlds

To remove the effect of the trade, we compare web agencies only - same question, same wording, same assistants. What is left depends on the country alone.

Three American agencies out of eighty-seven. One French agency in four. These are comparable companies: a website, a published contact address, an identifiable business. Nothing in their profile says they should be eight times less visible.

Intuition says a big market offers more room. The measurement says the opposite: it does not offer more places, it offers the same eight - and far more candidates.

Why, probably

An assistant does not return a directory, it returns an answer - a handful of names. That number does not grow with the market: you get five to ten names whether you ask about Lyon or about Chicago. But behind the American question stand tens of thousands of agencies, and behind the Lyon question, a few hundred.

In a small market a mid-sized firm can hold one of those places. In a large one, the same places go to national brands and everything else is nowhere. This is not a punishment. It is arithmetic.

We offer this as a hypothesis, not a demonstration. Our measurement establishes the gap; it does not establish its cause. A language effect is possible - our French and American questions are not the same sentence - and so is a selection effect, since we did not draw these companies at random.

The case that surprised us most

Two legal professions, one word in our system and two in real life: solicitor in the United Kingdom, lawyer in the United States.

• United Kingdom, 11 firms: 2 are named, and one scores 66/100 - the highest score in the entire campaign.
• United States, 26 firms: none. Not one, not once.

Eleven firms do not make a proof, and we will not pretend otherwise. But the gap runs the same way as the one measured on 263 agencies - and it is that agreement, not the eleven, that makes the lead worth following.

What this changes for a business

If you operate in a large market, your absence from AI answers says almost nothing about your quality and a great deal about the density around you. The useful conclusion is not “build national awareness” - it is to narrow the question.

An assistant asked for “a lawyer in Chicago” picks from thousands. Asked for “a construction law attorney in Evanston”, it picks from a few dozen. The seat is still there; what changes is the queue.

The limits

• This measurement queries two models that answer FROM MEMORY, with no web search. A zero means “these models do not know you by heart”, not “you are invisible everywhere”. The same businesses queried with web search get different answers.
• The sample is not random: these are businesses we had shortlisted for outreach, so all of them have a website and a published address. They are more online than the average of their trade, which makes the 89% figure optimistic rather than harsh.
• Group sizes are very uneven - 176 French agencies, 11 British firms. We print the count on every line so you can weigh each figure yourself.
• One question per business, one pass. We have shown elsewhere that the same question replayed does not always return the same list; these scores are snapshots, not averages.

You can run the same measurement on your own business, free and with no account. It also queries Google live, which this one did not - so your result will be closer to what your customers actually see.