The contact page has a section called “Note on timelines.” I wrote it to tell prospects I protect my process. The old copy read: I’m not taking on rushed work. Two weeks minimum. Probably not the right person if you need it faster.
I re-read it through a prospect’s eyes. It reads as a rejection. Not “this person has standards,” which was the intent, but “this person doesn’t want your work.” Same information. Opposite reader experience.
That was the fourth time in five weeks I’d found the same problem.
The shape of each fix
The same sequence every time. I re-read a section of the live site cold, not as the person who wrote it but as someone seeing it for the first time. Something doesn’t land. Not a typo or a broken link. A sentence doing work inside my head that it isn’t doing on the page.
Strip the sentence. Rewrite for what the reader actually decodes. Ship.
The timelines note
Old copy: “I’m not taking on rushed work — minimum engagement window is two weeks. If your project needs to move faster than that, I’m probably not the right person.”
Inside my head, that sentence signals selectivity. I have a process, I protect it, the work is better for it.
From outside: the first thing a prospect reads is “not the right person.” The framing is negative throughout. The note tells them what I won’t do before it tells them what they get. The quality signal lands, but it lands behind a gate.
The fix inverts the frame. The two-week minimum is still there; it’s a real constraint, not one to hide. But it leads with what that constraint enables: considered work, not reactive patches. The prospect reads a strength, not a rejection.
Same information. What changed was the order and the frame.
The album credit
The NAPA SEASON album blurb included: “Suno for the production and vocals.”
Technically accurate in a narrow sense. Suno renders vocals and instruments. The problem is what a reader decodes: Suno made the album. Not “Suno is a tool in the chain.” “Suno is the author.”
The reality: lyrics are written here, production is built here, melodies developed in Logic. Suno renders specific elements at the end of that chain. The creative authorship is not Suno’s.
“Suno for the production and vocals” was the inside-head version of that process. From my own perspective the split is obvious, because I know exactly which pieces Suno contributes and which it doesn’t. A reader sees “production and vocals” attributed to a generative AI tool and draws the obvious conclusion.
The fix names the split explicitly: what’s written and produced here on one side, what’s rendered via AI on the other. A reader can decode the actual authorship now.
This one is the most uncomfortable of the four. I wasn’t just writing for myself; I was accidentally crediting a tool with work that wasn’t the tool’s. The reader didn’t get wrong framing. They got the wrong author.
The sub-brand rename
The music content ran under “Creative AI Series” across seven places: single and album pages, the footer, the videos index, the series descriptions.
“Creative AI Series” doesn’t answer the question a new reader arrives with. AI did what, exactly? Generated everything? Assisted? Curated? The name sits in the middle of the authorship question without settling it.
“Creative AI Assisted Series” answers it. Assisted. The work exists; AI is in the chain; the human is in charge. One word closes the ambiguity.
Seven places updated. The name is longer and probably no one reads sub-brand labels in footers with much care. Neither point changes what the original name was doing wrong.
The learning hub headline
The /learning/ hub opened with a headline that described the catalogue: what courses were in it, what subjects they covered. Functional. Accurate.
The fix: “Engineering thinking.” Full stop. The sub-heading, meta description, and hub JSON-LD carry the same positioning: courses that teach how an engineer reasons, not just what engineering topics they contain.
Old headline: here is what is in the catalogue. New headline: here is what you get from spending time in it. Those are different things, and the difference matters to a reader who doesn’t already know what they want.
Inside my head, the catalogue was always about reasoning, not content. Outside on the page, it looked like a content list.
The pattern
Four fixes. One move.
Each piece of copy was written from inside a working model of the thing. That model is accurate. I know why the timelines note is a quality signal, know what Suno’s role is in the chain, know the learning catalogue is about reasoning. The copy expressed that model faithfully.
The reader doesn’t share the model. They decode from words on a page, cold, without any of the context I’d built while writing. “Not the right person” decodes as rejection. “Suno for the production and vocals” decodes as authorship. “Creative AI Series” decodes as ambiguity. “Here are the subjects in the catalogue” decodes as a list.
The fix is not finding better words for the same meaning. It’s switching from encoding your model to anticipating their decode. The test is not “does this say what I mean?” It’s “what does a reader with no prior context decode from this?” Those two tests fail in different places.
The sentence that flatters the writer is the one that accurately expresses the inside model. To me, “I’m not taking on rushed work” reads as honest; “Suno for the production and vocals” is technically true; “Creative AI Series” was the internal name that made sense from the inside. None of those sentences are lies. They’re written for the wrong reader.
What comes next
No clean endpoint. The copy doesn’t stop being inside-head once you’ve noticed the pattern.
The timelines note took one re-read to catch. The NAPA SEASON blurb took longer because the error was more uncomfortable: giving authorship to a tool when the work is mine is not a framing problem, it’s a credit problem. The rename took a count of seven instances across the site. The learning headline took one conversation about what the catalogue actually does.
None of the fixes were long. They were slow-to-see. The test is now visible: whose head is this sentence written for? If the answer is mine, then run the decode pass and ask what a cold reader actually gets from the words. If that’s not what I want them to get, fix it.
The next one should be faster to find.
