Getting found — on Google and inside AI answers
Search optimisation is usually aimed at the ranking, even though the gap sits elsewhere. We measured this on our own sites: decent positions, decent impressions — and still too few clicks, because the entry in the results list gives nobody a reason to click, and because the landing page offers no next step. On top of that there is now a second front: a growing share of answers is produced by AI systems that cite rather than link.
Technical SEO, structured data and content, plus our own rank tracking instead of an agency report. And the second front: ChatGPT, Perplexity and AI overviews should name you as a source too.
- ›Technical SEO: indexing, load time, structured data, sitemaps
- ›GEO: preparing content so AI systems can cite it
- ›Our own rank tracking and analytics on our own hardware
- ›Honest measurement: clicks and enquiries, not a visibility index
Businesses that already have a site and want to know why it produces no enquiries. If you are looking for a guarantee of first place, we are the wrong choice: there is no such thing, and anyone selling it is selling something else.
What we run ourselves
We run rank tracking and analytics for our own sites on our own hardware. Doing so taught us more about the gap between ranking and enquiry than any report ever sent to us.
- Our own rank tracking
- Daily position measurement with our own data storage. No tool subscription, no pre-filtered selection, every query auditable.
- AI traffic measured separately
- Visits originating from AI systems are reported separately. Across our measurement series that share has multiplied — which is what made us take GEO seriously.
- Analytics without a third party
- Matomo on our own hardware instead of an external analytics service. Visitor data stays in the building and the consent question eases.
- Immediate notification to search engines
- New posts are submitted automatically on publication instead of waiting for the next crawl.
- Machine-readable statement
- Our sites carry llms.txt and ai.txt: a clear statement of what AI systems may read, cite and summarise.
How a project runs
Measure first, fix second, write third. In the reverse order you produce content for a site that cannot deliver it.
- 1Assessment
A technical pass over indexing status, structured data and load times — plus an honest look at which search queries already bring clicks today.
1–2 weeks - 2Straighten out the technology
Fixing what costs visibility: broken addresses, missing markup, duplicate content, slow pages, unclear language assignment.
1–3 weeks - 3Content and GEO
Building pages for the questions your customers actually ask — stated unambiguously, marked up in a structured way, machine-readable.
ongoing - 4Measure and adjust
A monthly review: positions, clicks, AI traffic, enquiries. What produces nothing gets dropped rather than continued.
monthly
What it costs
The assessment is a fixed price and useful on its own — including if you then continue with somebody else.
- What drives the price
- The state of the existing site. A grown installation with old addresses and three language variants takes more clearing up than a clean site.
- What lowers the price
- Access to expertise inside the company. The content that genuinely gets found comes out of what your people know — not out of what somebody researches from outside.
- Running costs
- Rank tracking and analytics run on our own hardware, so there is no monthly tool subscription. You pay for the work, not for access to a dashboard.
What you get
- ✓A technical assessment prioritised by impact, not a defect list ordered by count
- ✓Technology that has been fixed, not merely documented
- ✓Structured data and machine-readable statements for AI systems
- ✓Measurement against clicks and enquiries rather than a visibility index
- ✓Figures that belong to you: your own tracking, your own analytics, exportable
Frequent questions
- What is GEO, exactly?
- Generative engine optimisation — the work of getting AI systems to find, understand and cite your content. In practice that means unambiguous statements instead of marketing language, structured markup, machine-readable information and a clear statement of what may be crawled. On our own sites we can see traffic from AI systems growing measurably. It is no longer a future topic.
- How long until something happens?
- Technical fixes take effect within days to weeks, once pages are re-indexed. Content work takes months. Anyone promising you success in six weeks means a report, not enquiries.
- Isn't an agency report enough?
- A report showing visibility and positions answers the wrong question. What matters is how many people clicked, and how many of those wrote to you. We measure with our own rank tracking and our own analytics — the same figures, without the selection a salesperson makes.
- Does this need a constant stream of new content?
- Not constant, but regular, and on real questions. A post that answers one concrete question properly is still found and cited years later. Twenty posts without substance only create maintenance.
senn-tech