From zero to a Fortune 500 deal in two years.
How Makula's organic and AI search channel was built from nothing, and what a falling click-through rate revealed about AI search changing the rules mid-engagement.
- Client
- Makula
- Industry
- Industrial B2B SaaS
- Engagement
- 24 months
- Role
- First inbound hire, solo
- Scope
- SEO, content, GEO
The starting point
Makula is a German industrial B2B SaaS company. I joined as its first dedicated inbound marketing hire, with no existing organic channel, content engine, or inbound pipeline to build from.
The mandate was to build the channel from the ground up and prove it with numbers, not vanity metrics. Here's what happened over two years.
Build the foundation by hand
No shortcuts. Blog content built around what buyers actually search, manual outreach for backlinks, guest posts on relevant industry sites, and structured (FAQ / comparison) schema so both search engines and, later, AI systems could parse the content correctly.
Referring domains, quarter by quarter
Authority built here didn't just help rankings, it became the foundation AI systems relied on to trust and cite the site once GEO work began in Phase 3. There's no shortcut to this step.
Turn visits into pipeline
As authority compounded, rankings moved. Average position across target keywords climbed from the high 30s into the 12–15 range, page two to the top of page one. Roughly one in five inbound leads booked a demo, and the channel helped land Makula's first Fortune 500 customer, its largest deal to date.
Average ranking position over time
The warning sign hidden inside a "good" number
Midway through year two, CTR started falling even as rankings and impressions both kept improving. On paper, visibility was up. In practice, fewer of those people were clicking through.
Click-through rate dropped roughly 40%, even while average position and impressions both improved. That combination, more visibility, fewer clicks, is the signature of AI Overviews and zero-click answers absorbing the query before it ever reaches the site. Ranking well was no longer enough on its own.
The response: restructure content to be directly citable by AI systems, not just rankable by traditional search, an approach usually called GEO.
None of this was a trick. It was the same fundamentals, structured so an AI system could lift the answer cleanly and cite the source. Google's own guidance is blunt about this: there's no secret markup or magic file that gets you into AI answers. The teams selling GEO as a hidden formula are mostly selling nothing.
AI referral traffic, quarter by quarter
This isn't a straight-line "4x growth" story. AI referral traffic surged after the shift, peaked, and then settled at a new baseline roughly 50% above where it started. That's still a real, durable gain, and a fair one to claim, but early-stage AI-search channels are volatile month to month. Anyone promising a clean exponential curve here isn't reading the data honestly.
Traffic share is one measure. A more direct one: towards the end of the engagement, roughly 90% of the demos that actually converted traced back to buyers who found us through AI, not a search results page. Traffic volume and demo share aren't the same number, and the second one is the one that paid rent.
The numbers, quarter by quarter
| Quarter | Referring domains (index) | Avg. position | AI referral traffic (index) |
|---|---|---|---|
| Q3 | 100 | 34 | 100 |
| Q4 | 188 | 18 | 141 |
| Q5 | 198 | 16 | 179 |
| Q6 | 360 | 14 | 151 |
Referring domains and AI referral traffic are indexed to Q3 = 100 to show growth without revealing absolute scale. Average position is a rank (1–100), not a volume metric, so it's shown as-is.
Results at a glance
What this proves, and what it does not
The channel didn't fail when rankings improved and clicks stopped following, that was the signal, not a setback. Most teams would have kept optimizing for position and missed it entirely. The fix wasn't more content, it was different content: built to be the direct, citable answer, not just the top blue link.
This is the playbook the AI visibility audit runs on your site. One week, fixed price. You'll know exactly where you stand on Google and inside AI answers, and what to fix first, whether or not we work together after.