About us
A measurement tool, written by someone who needed one
Etalo came out of a concrete search problem, not a market study. That explains both what it does very well and what it does not do.
Where the tool comes from
Etalo was first built to answer one question: do AI assistants cite our brand when a customer asks them where to book?
No search tool could answer. Google positions, traffic, impressions: none of that measures what an AI tells someone asking it for advice. So the questions had to be put to the assistants, several times, and counted — hence the tool.
What shaped it was the need to defend the figures in front of a management team. A percentage you cannot explain is a percentage you eventually pull from the deck. That is why every rate in Etalo traces back to the exact sentence that produced it, with the assistant, the model and the timestamp.
What we believe
Three convictions, which are also three design constraints. They explain choices that sometimes look austere.
- An unverifiable figure is worthless
- Any indicator in the report drills down to the raw answer. Answers are kept exactly as received, indefinitely, which makes it possible to recompute the whole history when a counting rule changes. Without that, comparing two months would mean comparing two methods.
- A measurement without a confidence interval is an opinion
- AI never answers the same way twice. Each question is therefore asked several times, and each rate is published with its margin. A gap smaller than that margin is not a movement, and saying so saves pointless meetings.
- A corpus that moves measures nothing
- Questions are frozen before the first campaign and do not change afterwards. It is constraining, and it is the only way to obtain a trajectory rather than a series of unrelated snapshots.
The size of the company
Etalo is published by Galaxy SEO, in Aix-en-Provence, France. It is a small operation, and there is no reason to hide that behind an imaginary team "we".
Concretely: the person building your corpus is the one who wrote the code, and the one who will answer when a figure looks suspicious to you. There is no sales department promising what engineering cannot deliver, and no first-line support sending you back to a help page.
What it also implies, and this deserves saying: no contractual availability commitment, no overnight on-call, and a response time counted in working days rather than minutes. If your organisation requires a vendor with 24/7 support, Etalo is not the right choice.
Our relationship with AI vendors
Etalo is a partner of none of the assistants it measures, nor a reseller, nor an affiliate. ChatGPT, Claude, Perplexity, Google AI Overviews, Google AI Mode and Bing Copilot are queried through their public or paid interfaces, exactly as any customer would.
That is a comfortable position for a measurement tool: we have no interest in one assistant looking better than another in your reports. API costs are included in the engagement, which also saves you opening six accounts.
Who the tool is for
Worth saying plainly: Etalo does not suit everyone.
- Brands whose market turns on recommendation
- Travel, insurance, software, health, training, business services — any sector where the buyer asks for advice before choosing. That is where an assistant's answer genuinely replaces a list of links.
- Teams that already have measurement discipline
- The tool produces confidence intervals and frozen corpora. Those notions speak immediately to an SEO or analytics team, less so to an organisation expecting a single score out of a hundred.
- Organisations able to act on their content
- The correction plan names pages to write and markup to add. Without the capacity to produce content, the diagnosis stays on the shelf — and paying for a diagnosis you cannot act on makes no sense.
Let us see where your brand stands.
A demonstration on your brand: what an agent can read of your site, and where you stand against your competitors.