- Tim at Penfriend
- Posts
- I built a lie detector for boring content
I built a lie detector for boring content
20k micro-signals grade your draft in seconds—score 85 +, get confetti + ranking content

Day 119/100
Hey—It's Tim.
2018: a housing software company I was meeting with has lost 40 % of clicks in 48 hours.
Them: “Can we fix it by Friday?”
Me: “Can we invent time travel?”
It was a scene I’ve replayed in a dozen boardrooms since.
Any channel you rely on exclusively will betray you eventually.
Why Diversifying Feels Impossible
Because the advice is a Pinterest poster:
“Be everywhere.”
Cool. 👍
On top of writing, now I’m a video editor, audio engineer, shorts-whisperer, carousel artisan, and midnight community manager.
Right.
Most teams can’t even make one format sing, let alone five in harmony. They keep spraying half-good content across every feed, praying for compounding returns that never show.
The Real Bottleneck
It isn’t distribution.
It’s quality.
If your content can’t hold attention on home turf, shipping it to five other platforms just accelerates the embarrassment.
So—how do you prove a piece is actually good before you syndicate it everywhere?
VIBE Check:
Something I’ve always felt from great content. But never been able to prove it.
Until now.
Turns out there are a lot of research papers dedicated to working out what makes us pay attention to things.
With research comes process, and with process come formulas.
Now, I’m not gonna give you the formulas, at least not verbatim.
I’m literally building a product that figures this out for you But, I will give you overview.
And honestly, if you follow this, you’ll be diamond hand through any algorithm updates.
V – View
Plant a flag. If your piece could sit on the fence, it’s not a view.
Hot, sharp takes travel further. Defend them with empathy, not insults.
Example: “Zero-click SERPs aren’t the apocalypse; they’re free billboards. Fight me.*
I – Individual
Prove a human wrote it. Data you gathered, screenshots from your dashboard, mistakes you owned. If ChatGPT could fake it, you haven’t gone deep enough.
Show your fingerprints. “We tested 11 H1 variants on a page that sells shower heads” > “You should A/B test.” First-hand beats second-hand every time.
B – Beat
Readers tap dance through text. Give them rhythm:
punchy lines
roomy margins
sub-heads that double as tweets
E – Evidence
Receipts > opinions. Graphs, user quotes, Looms of you breaking things in staging - whatever shows the work.
V + I + B + E rolls up into a numeric VIBE Score (0-100).
Hit 85+ and we give you a little confetti burst in the editor ✨
How we’re actually building the VIBE Score
Forget “rubrics” and “heat-maps.” This thing is straight-up forensic linguistics:
Layer | What We Check | Why It Matters |
---|---|---|
20,000+ micro-signals | Syllable count, sentence length variance, paragraph cadence, uncommon-word ratio, conversational slang markers, even the shape of your punctuation. | Great writing sounds human before it says anything smart. |
Structural breathing room | Average characters per line, whitespace distribution, skim-path rhythm. | Your eyes need breaks the way your brain needs dopamine. |
Evidence integrity | Source credibility, citation format, method transparency, first-party vs. third-party data, “I-did-the-thing” verbs. | Claims without receipts nosedive the score. |
Insight quotient | Novelty detection against 10k+ published posts; flags recycled takes; rewards unique POVs. | If ChatGPT could have written it on default settings, you lose points. |
Quad-LLM ensemble | Four fine-tuned models trained on the web’s top-performing long-reads battle-vote every paragraph. | Consensus = confidence; disagreement triggers human review. |
End result: a single 0-100 VIBE Score delivered in seconds, with inline red flags that read like an editor who skipped brunch:
“Paragraph 3: 27 words, four commas, zero oxygen. Breathe, champ.”
Reality Check
Yes, this is extra work. But rebuilding traffic every quarter is worse.
Play infinite games, not quarterly ones.
See you tomorrow.

✌️ Tim "Not Financial Advice But Kinda Traffic Advice" Hanson
CMO @Penfriend.ai
Same brain, different platforms: X, Threads, LinkedIn.
P.S. Goal: make “What’s the VIBE Score?” as common a question as “Did we optimise H2s?”

Penfriend.ai
Made by content marketers. Used by better ones.
What to do next
Share This Update: Know someone who’d benefit? Forward this newsletter to your content team.
Get your First 3 Articles FREE EVERY MONTH! We just dropped the biggest update we’ve ever done to Penfriend a few weeks ago. Tone matching with Echo, Hub and Spoke models with Clusters, and BoFu posts.
Let Us Do It For You: We have a DFY service where we build out your next 150 articles. Let us handle your 2025 content strategy for you.