This article covers one of the main concepts in my recent research report Marketing at an Inflection Point: A Structural Diagnosis, and the Strategic Shifts Required Now.
I used to be able to feel momentum in marketing.
Not the kind of synthetic momentum you see in dashboards. The kind you feel in the room and in the work that’s being done.
There used to be a sense that when you were on an effective path, things had a kind of colour and vibrancy to them that’s hard to put into words. I think most people who have worked in marketing for a decent amount of time likely get what I’m saying, or even have their own version of this sense.
It was often the case that doing the right things was instinct more than science because we all had our finger on an unquantifiable pulse that we knew would lead to momentum.
And then, somewhere over the last few years, that feeling got harder to find.
Not because marketers got worse. If anything, the average competence level is higher than it has ever been, I personally feel like I’m constantly surrounded by people far more competent than I am. Tools are better. Execution is faster. We can produce more, more quickly, across more formats, with more polish.
But I keep noticing the same uncomfortable pattern across teams and organisations…
Output is going up. Impact per unit is going down.
And I hate to refer to the work as “output” or a “unit”. But that’s the world we’re in now. We mechanised marketing through the integration of big data, trackability and reducing everything to performance metrics.
But the more pressured the organisation has become, the worse this reality gets. Because under pressure, “output” feels like the safest lever to pull.
If you are a marketer reading this, you already know what this looks like.
It’s Monday morning. Your calendar is full. The content pipeline is stacked. Slack is already loud. Everyone is “shipping”. And now, with AI, you can ship at a pace that would have felt absurd two years ago.
And yet, you hit publish and you feel almost nothing.
Not because the work is awful. It is usually fine. It is competent, polished, useful in a generic way.
But it lands like another raindrop in a storm.
This is what I mean by output inflation.
When the cost of producing content collapses, supply floods the system. In an abundance environment, volume becomes a weaker lever. It’s because the market adapts. The baseline rises. And the marginal unit value of content collapses.
AI did not create this problem. It just turned it from a slow drift into a hard reality by pushing it further up the exponential curve.
The supply shock: what changed
The most important shift is not that “content is harder” or “attention spans are shorter”.
It’s more boring than that… production got cheap.
When something becomes cheap, it stops being a differentiator. It becomes a baseline for everything. And this is the bit that is so hard to stomach for so many people in the profession.
Ahrefs analysed 900,000 newly published web pages in April 2025 and found 74.2% contained AI-generated content.
Even if you want to debate the exact number, the point is clear… the world just experienced a supply shock.
Now I need to make clear, this isn’t me saying that AI content is bad or is “slop”. I use it a lot to create content myself. I’m purely pointing to the supply dynamics here.
And you can feel it.
LinkedIn feeds that read like they were written by the same person.
Blog posts that have the right headings but no heartbeat, personal stamp or soul to them.
“Thought leadership” that is technically correct but spiritually empty. Like it’s missing the lived experience and battle scars of the author.
And the really weird part is that much of it, is actually really good quality. But it’s interchangeable.
Which means the new scarcity is not output.
The scarcity is…
- attention
- trust
- recognition
- and memory
There’s another part of the supply shock that matters too. In many channels, content is being consumed, summarised, reformatted and re-published without ever reaching you.
Pew Research found that when people encountered a Google AI summary, they clicked on traditional results in 8% of visits, versus 15% when no AI summary appeared.
So the supply goes up, the clicks go down, and the marketing team is still asked the same questions…
“Why isn’t this working like it used to?”
That is the environment you are operating in. That is why this all feels so much harder and difficult to navigate. The environment is shifting rapidly.
Why volume feels rational (and why you end up doing it anyway)
Here’s the thing I have learned the hard way.
Nobody wakes up and thinks, let’s create a content treadmill that slowly drains the team’s soul. Most of us got into marketing because it’s a diverse career path that opens up opportunities to be creative, analytical and expressive all at the same time.
Output inflation forms because it is a rational response to pressure.
1) Brute-force signal hunting
When the journey gets messier and measurement gets weaker, teams try to find signal by increasing the number of shots on goal. It almost feels logical.
I have done this. I’m guilty of this. As I know many of you are.
You start publishing more because it feels like the only way to regain control. The output lever is irresistible because it almost feel virtuous to pull it.
If we just get more out, we will find what resonates. If we throw more at the wall, we’ll see what sticks.
And for a while, it can work. You get the occasional spike. A post that does unusually well. A campaign that performs above average. A “this one really landed” moment. And it reinforces your initial instinct.
So you conclude… we need to do more of this.
And that is when the trap begins. That is when you take a path that is difficult to abandon. Because that path is where expectations of busyness and increasing output become baseline and you’re now battling a growing beast… diminishing returns.
2) Output becomes internal proof of value
There is an ugly truth inside many organisations.
Marketing is constantly asked to prove it deserves to exist.
Rory Sutherland puts it well when he discusses how marketing teams are increasingly being referred to as “the colouring in department”.
So output becomes a shield
A full content calendar looks like “a functioning department”.
A weekly newsletter looks like “momentum”.
A constant stream of social posts looks like “brand building”.
I have sat in too many meetings where the quiet subtext is… if we are not visibly producing, we will be questioned.
So we produce.
Because even if what is produced is not having the desired impact. We can at least point to the work and say “Ok, so performance isn’t where we want it to be, but at least you can’t question our effort or good intentions”.
3) AI makes “more” feel inevitable
This is the AI twist.
Once leadership sees AI producing assets instantly, the assumption becomes…
“If this is now easy, why aren’t we doing more?”
You can feel the expectation shift in real time.
The output target is never formally increased, but it is implied.
Suddenly the team is not just a marketing team. It is a content engine. A factory. A machine.
And this is where the mechanisation of marketing now reaches a new level. When you are inside a factory, it becomes very hard to stop and ask the most important question.
“Is this actually working?”
Why it fails: the noise floor rises and the learning dies
Output inflation fails for two reasons, and both are structural.
Failure mechanism A: the noise floor rises
Mark Schaefer’s “content shock” idea basically predicted this. Content supply grows faster than human consumption capacity, so reaching people becomes harder and more expensive over time.
AI has turned that into a daily experience.
When everyone produces more, the baseline rises.
When the baseline rises, “showing up” stops being a differentiator.
So your incremental content has to fight not just for attention, but for meaning.
And this is where a lot of marketing teams get stuck emotionally.
Because you are working hard. The output is real. The team is busy. The assets are clean.
But the market response is muted.
Personally, I’d argue this one of the leading contributors of burnout in our profession. Even if not explicitly stated, when you work hard to increase output, but the feedback doesn’t give a sense of momentum, it has an impact on you.
System1’s work on dull advertising makes the same point in a different context, dull creative is not neutral, it is expensive. You often have to spend more to get the same effect when the creative does not provoke a human reaction.
Generic content works the same way. It is not harmless. It is a cost.
Not just in effort, but in opportunity.
Failure mechanism B: evaluation overload kills learning
This is the one that tends to happen quietly.
When you produce too much, you stop learning properly.
I have watched teams ship at insane pace and then do a “performance review” that is basically…
- this did fine
- this was average
- this was below average
- hard to tell why
- we should test more
You end up with a fog of half-conclusions because too many variables are moving at once.
You cannot isolate what mattered.
You cannot build confidence.
You’ve probably experienced this yourself. When the idea of A/B or multivariate testing gets raised in a meeting, the very thought of it is daunting. Because where do you start? When you have a hosepipe of content being produced, on multiple channels, in multiple formats, for multiple audiences… how do you even begin to untangle that?
And the team starts to default to what feels safe… keep producing.
Meanwhile, distribution changes mean your ability to measure depth and intent gets weaker too. If people are satisfied by summaries and do not click through, behavioural data thins out.
So the team produces more to compensate.
The loop tightens.
The doom loop: motion without progress
Here is output inflation as a loop you can actually recognise:
- Performance pressure rises
- Measurement is unclear
- Output increases
- The noise floor rises
- The ability to interpret results drops
- Confidence drops
- Leadership asks for proof
- The team produces more
I have lived this loop.
It feels like being on a treadmill that speeds up every month.
And the most brutal part is that everyone involved is usually acting in good faith.
It’s just that the system rewards motion, not progress.
We don’t think to re-design the system, we instead look for more ways to increase output.
What to do instead: fewer bets, stronger standards, distinctiveness systems
This is the part that matters. Because diagnosis without a way out is just another kind of content.
If you want to escape output inflation, you have to change the operating posture. And believe me, this isn’t easy too do. It’s political, it requires a mindset shift and it means people have to confront uncomfortably realities.
A) Fewer bets (this is the hardest one)
Most teams are not short on ideas. They are short on focus.
Output inflation is usually a symptom of too many parallel initiatives…
- too many messages
- too many audiences
- too many formats
- too many stakeholders making “small requests”
- too many “quick wins” that become permanent
I often find this is actually the result of a lack of a coherent and clear commercial/business strategy, but that’s for another article.
So, the first move is to reduce the number of concurrent bets.
Pick fewer themes per quarter. Pick fewer promises you are trying to make. Pick fewer channels you are trying to win in.
And be ruthless about “random acts of content”.
If you cannot clearly state what the next piece of content is supposed to do, do not ship it.
This one rule will save you endless amounts of energy and performative work if you apply it effectively.
B) Stronger standards: install quality gates
If AI increases throughput, your standards must rise to match it. Otherwise you just scale generic, low-impact content that delivers little to no value.
Here are four gates I have found useful.
1) The job gate
What is this piece supposed to do?
Educate, reframe, reduce risk, build trust, create desire, answer a question, provoke thought.
If you cannot name the job, it is filler. And you need to kill it before time is wasted on it.
2) The distinctiveness gate
Could a competitor publish this with minimal edits?
If yes, kill it.
This sounds harsh but it is a gift. It forces you away from the mean. It reduces the risk of entering the algorithmic middle where AI primarily leads.
3) The proof gate
Can we defend what we are claiming?
AI can generate convincing nonsense, and people tend to overtrust it. Nielsen Norman Group’s work on hallucinations is a useful reminder here.
4) The taste gate
Would a real human care?
Would you send this to someone you respect?
Would you feel proud if a client quoted it back to you?
Taste is underrated because it is not measurable, but it is one of the most important filters in an abundance environment. I can’t emphasise the importance of this enough. With where AI is taking content, this is your moat. This is your chance to still stand out and make an impact.
C) Distinctiveness as a system, not a hope
Most brands treat distinctiveness like a branding exercise.
In abundance, it is a survival mechanism.
The Ehrenberg-Bass view is basically that growth is tied to mental availability and recognisable cues.
Applied to content, that means building a recognisable creative grammar…
- recurring formats
- signature story types
- consistent points of view
- recognisable visual cues
- a few topics you become known for
You are not trying to say more things.
You are trying to become easier to recognise and remember. That’s the whole point of this. And if your focus is purely on volume of output, you’ll miss this entirely.
D) Build compounding assets
In my opinion, this is the move that changes everything. And this is what turns yesterday’s marketers into designers of ecosystems as opposed to mechanical conveyor belts of generic content.
Ask:
“What are we building that gets stronger over time?”
Examples:
- a category point of view that becomes a reference
- a benchmark / dataset updated quarterly
- a signature series people recognise instantly
- a library of customer stories that becomes proof
- a framework that becomes shorthand for how you think
Compounding assets do not rely on novelty. They rely on depth and repetition. They’re not easy to copy. And they provide significant competitive advantage.
They turn content from output into equity.
Believe me, this is extremely difficult to do, and the change required to make it happen is a painful one. But in my view, this is the game now. The difficulty level in marketing just got put up to god mode after the rise of AI, and that’s the standard if you want to have an impact.
Where AI helps and where it misleads
AI is extremely useful. But only when it sits behind judgement. If you’re reeling off content from single-prompts with little to no editing, or original thinking to start it off in the first place, you’re probably already drowning in a void of low-impact work and performance pressure.
Where AI helps…
- first drafts and structure
- summarising research
- translating ideas across formats
- generating variants once the core is strong
- removing blank page friction
Where AI misleads…
- pulls you toward the mean
- makes “good enough” feel acceptable
- inflates confidence
- increases output faster than learning
- scales sameness
So here is the rule I keep coming back to…
Use AI to accelerate craft and iteration, not to justify volume. Using AI in content creation, or anything else for that matter, is not a bad thing as long as you can pass through all the gates we covered.
And remember, speed is only valuable if you can learn at the same pace.
The reframe
In the AI era, more content is easy.
That is the trap.
The job is not to fill the calendar. The job is to earn attention, build memory, and create preference in a world where everyone is producing more than anyone can process.
Output inflation is what happens when the organisation responds to uncertainty with activity.
The alternative is to respond to uncertainty with clarity…
- fewer bets
- stronger standards
- distinctiveness as a system
- compounding assets
- AI used to amplify judgement, not replace it
The next era of marketing will not be won by the brands that publish the most.
It will be won by the brands that become the easiest to recognise, the easiest to trust, and the hardest to forget.
Funnily enough, that’s exactly as it was before, but that idea will become more entrenched as the tsunami of AI content hits us with force.
