Florian S. Kluge

The S is silent. It stands for SEO.

Ex FDE. Technical SEO and AI search, for over a decade — mostly for technical B2B SaaS, the kind where the marketing site lives near the product and ships when engineering ships.

What I actually do

I built systems that replaced me, then built more. Now I spend most of my time finding out which ones were worth building.

That second half is the unusual part. Our field is generous with recommendations and stingy with follow-up — the audit gets delivered, the deck gets presented, the next engagement starts, and nobody goes back to check whether any of it moved. I got tired of being part of that, so I started building the boring infrastructure instead: baselines per page type, evidence rules fixed before the result arrives, and a written verdict afterwards.

Three outcomes: it worked, it didn't, or I couldn't tell. The third one is real and I say it out loud, which costs me something and is exactly why it's worth anything.

How I got here

Forward-deployed

Engineering work where you sit with the customer and own the outcome, not the ticket. It permanently changed what I think "delivered" means.

A decade in search

Technical SEO for companies where organic is a revenue line, through several eras of Google deciding to change its mind.

Then AI search

The same discipline pointed at answer engines: what gets cited, what gets ignored, and how little of it is under your control.

Then automation

I automated my own reporting, then my competitive analysis, then most of my analysis. Each system did the job better than I did by hand, and each one raised the same question.

Now, verdicts

Which of it was worth doing? Agents generate. Humans sign. Evidence closes the loop. That's the work.

Things I believe, stated so you can disagree

A baseline that flatters itself is worse than none. Most reporting is built to look good in a meeting. If the instrument can't produce bad news, it isn't an instrument.
"I couldn't tell" is a finding. Forcing a verdict out of insufficient evidence is how teams end up confidently wrong for a quarter.
Execution is commoditizing; judgment isn't. The tools that produce the work are getting cheap and fast. Deciding what deserves to ship, and whether it worked, is not.
An audit nobody acts on is theatre. The measure of advice is a changed decision. If nothing changed, it didn't land, whatever the document says.
The expensive damage comes from ordinary days. Not bad strategy — a Tuesday deploy nobody reviewed.

Practically

Independent. I work directly with the engineers who ship your site, in your stack, on your release schedule. Currently in Portugal, moving to Berlin at the end of 2026. Comfortable in German and English.

Two ways to work with me: a fixed 45-day release assurance pilot, or an ongoing fractional growth engagement.

Write to me — flor@floriankluge.com

I read everything and answer most things. If what you need isn't something I'd do well, I'll tell you that instead of taking the work.