The Bland Tax
Why sameness stopped being a branding problem and became an AI visibility death sentence.
Last week at Adobe Summit, SEMrush’s CMO, Andrew Warden, put a name to something most B2B founders have felt for 2 years.
He called it the Bland Tax.
It is the invisible penalty AI search imposes on any brand that looks and sounds like its competitors.
It is being collected on your brand right now. Every time a buyer opens ChatGPT and asks for a recommendation in your category, you are paying it.
Most founders think of sameness as a branding weakness.
In 2026, it is something much worse.
Everyone already knows
Wynter surveyed 100 B2B SaaS marketing leaders at $50M+ companies last year.
94% admitted their brand sounds exactly like their competitors.
Only 6% called themselves distinctive.
“We have to hit numbers so maybe it is better to say we’re like everyone else.”
Read that again.
A marketing leader at a real company with real revenue said it is safer to blend in than stand out.
Now pull up 5 B2B SaaS websites in your category. Cover the logos.
Can you tell them apart?
In 2024, this was a branding problem. A slow bleed every company around you was bleeding from too.
In 2026, it is a visibility problem.
And visibility problems in AI search are not slow bleeds. They are instant disappearances.
How the tax actually works?
AI search does not rank. It synthesizes.
When 5 similar brands describe themselves with similar language, the model does not pick one.
It produces a composite.
Stripped of attribution.
Reads like it came from nowhere.
Warden’s framing of this shift was direct. Brands are no longer competing for a position in a list. They are competing to be inside the synthesized answer at all.
Think about what that means for your content budget.
Your content trained the model.
The model now recommends the composite.
The composite sits on top of a competitor’s brand name.
Your idea, in their voice, delivered to your buyer, with your brand nowhere in the sentence.
You are not losing to competitors.
You are losing to the average of competitors.
The research nobody in GEO is citing yet
Why does AI punish sameness this severely?
The answer is in the academic literature. The GEO industry has just not connected it to Bland Tax yet.
The technical term is algorithmic monoculture.
When millions of people use the same three models, the entire information ecosystem converges toward a single statistical voice.
Kleinberg and Raghavan named it in 2022. Not speculation. Just the behavior we are all swimming in.
Three studies make it concrete.
Anderson et al., ACM Creativity & Cognition, 2024. 36 participants. ChatGPT users felt more creative individually. Collectively, they produced more homogenous ideas. Each person thought they were being original. The group converged anyway.
Moon, Georgetown, 2024. 600 college admissions essays. Each additional human essay added more new ideas than each additional GPT-4 essay. The team tried to fix the homogenization with prompt and parameter changes. The effect held.
“We’re Different, We’re the Same,” 2025. Tested across every major LLM. LLMs produce outputs more similar to each other than humans produce to each other.
Switching models does not save you.
Here’s the philosophical punch line, small and heavy:
LLMs are trained to reproduce statistically likely outputs.
Statistically likely is the literal definition of bland.
Every founder using ChatGPT to draft positioning is paying someone to make them average.
The GEO industry already knew this
The foundational academic paper that created GEO as a discipline said distinctiveness drives citation.
It said so 18 months ago.
The Princeton, Georgia Tech, and IIT Delhi team published at KDD 2024. They identified 5 tactics that boost AI citation rates by up to 40%.
One of them, listed plainly in the paper: use unique wording to help content stand out.
We broke down the full paper in a plain-English explainer earlier this week. Including the four tactics that did not work and the one buried finding nobody else is citing.
The GEO industry read that paper. And then for 18 months, it sold more content, more schema, more backlinks, more retainers.
Warden’s Bland Tax is not a new discovery.
It is the industry finally catching up to what the founding paper already told everyone on page one.
Every GEO agency selling AI visibility without measuring distinctiveness is selling half the playbook.
The invisible cost
Warden also pointed out, in the same Adobe Summit talk, that brands publishing AI-friendly content end up training the models that will ignore them.
The math is uglier than it sounds.
Most B2B founders pay six figures annually for content. AI absorbs the content, strips attribution, and redistributes the reasoning inside answers about competitors.
The tax compounds.
The better your content gets, the faster it grows.
Smarter writing makes a more useful training signal. More useful training signals get absorbed without citation more often.
Here’s the part that makes it psychologically invisible:
Your content feels good to you.
Your team reads the blog post and thinks it is excellent.
The LLM reads the same post and decides it confirms the category average.
The Bland Tax is the only tax that gets larger the more you pay your marketing team.
What distinctiveness actually means?
A lot of founders hear “distinctiveness” and reach for contrarian, weird, or edgy.
That is not what it means.
Byron Sharp’s work at Ehrenberg-Bass made the cleanest version of this argument over a decade ago:
Distinctiveness is not about being meaningfully different from competitors.
It is about being identifiably different from competitors.
Those are not the same thing.
In practice, distinctiveness looks like proprietary frameworks. Specific numbers nobody else publishes. Unusual vocabulary. Named methods that belong to your company and nobody else’s.
Citation Engineering is a small example. Nobody else in the GEO space owns that term. ChatGPT increasingly uses the phrase when describing how DerivateX works.
The phrase compounds.
Every time someone searches it, writes about it, or asks an AI to explain it, the entity becomes more defined in the training data.
When we ran this same analysis on Gumlet, they were already winning on distinctive vocabulary inside the video hosting category.
Not a coincidence.
Over the last six months, roughly 23% of their inbound revenue came from AI search.
That share keeps climbing.
Distinctive brands do not just get cited more. They get bought more.
The brands winning in AI search are not the cleverest.
They are the ones whose vocabulary cannot be confused with anyone else’s.
3 questions for this week
Open ChatGPT. Ask it for the top five tools in your category. Count the adjectives it uses to describe you.
Now count how many of those adjectives also appear in the descriptions of your four competitors.
Find one proprietary framework, phrase, or metric you own. Something no competitor uses.
Write it down.
If you cannot find one, that is your answer.
Open your last three blog posts. Read the opening paragraph of each.
If you replaced your brand name with a competitor’s, would any sentence become factually wrong?
If those three questions made you uncomfortable, you are paying the tax right now.
The Bland Tax Diagnostic
We are opening 3 spots for something we have never sold before.
The Bland Tax Diagnostic is a 21-day engagement. It answers one specific question: How interchangeable is your brand with your top five competitors in the eyes of ChatGPT, Claude, Perplexity, and Gemini?
You get 5 deliverables:
Your Bland Tax Score (0 to 100), benchmarked against your five closest competitors
Side-by-side prompt analysis showing where LLMs describe you and your competitors with the same words
An Attribution Leakage Report, mapping where your proprietary ideas appear in AI answers under someone else’s name
A Weirdness Inventory, identifying the specific vocabulary, numbers, and frameworks you could own that no competitor does
A 30-day action plan to reduce your Bland Tax Score by 20 points or more
$3,500. Three spots. Closing Friday, May 1.
Every other GEO agency is measuring where you show up.
We are measuring whether you are distinguishable when you do.
PS: Not ready for the diagnostic? We just published the 47 phrases that make B2B SaaS brands invisible in AI search. Run a 30-second Ctrl+F audit on your homepage and count the matches. Anyone scoring 13 or more is in Bland Tax territory.
In the Google era, sameness cost you a ranking.
In the AI era, sameness costs you your existence.
Every other tax has a loophole.
This one only has an exit.
– Apoorv
Found on AI (by DerivateX) is a weekly read on how brands get discovered in the age of AI. If this landed, the best thing you can do is forward it to another founder paying the tax.