You Used AI to Produce More Marketing. The Market Didn't Notice.

A plumber in Brisbane used to publish two blog posts a month. Now, with AI, he publishes twelve. His organic traffic hasn't moved. His leads haven't changed. His phone rings the same number of times it did six months ago.

He's working harder than ever. Writing social captions, email sequences, blog articles, Google Ads variations. He bought three AI subscriptions. He feels productive. He's exhausted.

He's caught in a trap that has a name. Economists identified it 161 years ago. And it's eating small business marketing alive in 2026.

The 161-Year-Old Economic Law That Predicted Your AI Marketing Problem

In 1865, English economist William Stanley Jevons observed something counterintuitive about coal-powered steam engines. As the engines became more fuel-efficient, you'd expect coal consumption to fall. Instead, it skyrocketed. Cheaper coal per unit of work made MORE uses of coal economical. Factories expanded. Railways multiplied. New industries emerged that wouldn't have existed if coal were still expensive.

Total coal consumption didn't decrease with efficiency. It increased. The efficiency gain didn't save resources. It unleashed demand.

This became known as the Jevons Paradox, and it's the single most useful framework for understanding what's happening with AI and marketing right now.

AI made content production roughly 4x faster. Teams using AI content tools produce 4.1x more published content per marketer per month. Production costs dropped by an average of 42% across formats.

So what did businesses do? Exactly what the Jevons Paradox predicts. They produced more.

More coal. More steam. Same destination.

The Numbers Are Damning

Here's what's actually happened since AI content tools went mainstream:

Read those numbers together. Content volume is up. Cost is down. And 77% of marketers admit their own content feels generic. That last number is the Jevons Paradox in one statistic: efficiency gains didn't improve quality. They lowered the bar for what gets published.

Mark Schaefer called this "Content Shock" back in 2014 - the point where content supply exceeds the audience's ability to consume it. In 2026, with three-quarters of new web pages containing AI-generated content, we've blown past Content Shock into something worse: Content Invisibility. The content exists. Nobody sees it. The audience's attention hasn't grown to match the supply. It never will.

Why More Marketing Doesn't Mean More Growth

Here's where marketing science explains what the productivity dashboards can't.

Professor Byron Sharp at the Ehrenberg-Bass Institute has spent decades proving that brands grow by reaching more people, not by saying more things to the same people. His research across 130+ brands and 13 product categories shows that reach matters significantly more than frequency, and that additional exposures to the same person suffer steep diminishing returns.

Think about what most businesses do when AI makes content creation faster. Do they reach new people? No. They produce more content for the same audience.

They publish 12 blog posts instead of 2, but nobody new is reading their blog. They write 30 ad variations instead of 5, but they're targeting the same postcode. They send 4 emails a month instead of 1, to the exact same subscriber list.

The Jevons Paradox turned your AI efficiency into frequency. Sharp's research says frequency is the wrong lever. The gap between what AI made easier (production) and what actually drives growth (reach to new buyers) is where businesses are losing money right now.

The Exhaustion Nobody Talks About

Researchers at UC Berkeley's Haas School of Business tracked 200 employees at a technology company for eight months to study how AI tools affected their work. Their findings, published in the Harvard Business Review in February 2026, were blunt: AI doesn't reduce work. It intensifies it.

The pattern they observed maps precisely onto what's happening with small business owners and their marketing:

• AI accelerates a task (writing a blog post takes 20 minutes instead of 3 hours)
• The time saved creates an expectation to do more ("I should be publishing more often")
• Scope expands ("Now I can also do email, social, video scripts, and case studies")
• The marketer is now managing 4x the channels with the same brain
• Quality drops because attention is spread too thin
• Results don't improve, so the response is... to produce even more

If you've subscribed to an AI writing tool and find yourself working MORE hours on marketing than before, you're living inside the paradox. The tool works perfectly. It made production cheaper. And you responded by producing more, not by doing something more valuable with the freed-up time.

We've seen this pattern play out before, just at smaller scale. It's the same dynamic we described in why your marketing AI works when you do - and that's the problem. The tool amplifies whatever behaviour you feed it. Feed it volume, get volume. Feed it strategy, get strategy.

Rory Sutherland Called It: Efficiency Is the Enemy of Magic

Rory Sutherland, Vice Chairman of Ogilvy, has been warning about the dangers of efficiency-driven thinking for years. His core argument: the hunt for certainty is killing creativity in marketing.

When AI makes content cheap to produce, the instinct is to produce a lot of it and see what sticks. More variations. More A/B tests. More formats. More channels.

Sutherland calls the people driving these decisions the "arithmocracy" - decision-makers who believe that more data and more output equals better results. His counter-argument: marketing's biggest wins come from fat-tailed distributions. Rare, creative ideas that produce outsized returns. "If 10% of your work could produce 130% of your revenue gains, would you still be obsessing over dashboards?"

His favourite example: DoubleTree hotels didn't grow by producing more marketing materials. They grew by giving every guest a warm chocolate chip cookie at check-in. One idea. Tiny cost. Massive distinctiveness. Sutherland has been telling that story for over a decade because it proves his point: constraint breeds creativity, and abundance breeds mediocrity.

An SME with unlimited AI content production has no constraint. So nothing forces them to think about what would actually be memorable. They default to volume. They end up with 48 pieces of content that all feel the same, because producing "more" was easier than producing "different."

This connects to what we explored in why following every marketing rule makes you invisible. The throughline is the same: more isn't better. Different is better.

The Creative Quality Multiplier Most Businesses Are Ignoring

If the Jevons Paradox identifies the trap and Sharp explains the wrong lever, Les Binet and Peter Field point to the right one.

Their analysis of the IPA Effectiveness Databank - the largest study of advertising effectiveness ever conducted, covering 996+ campaigns across decades - found something that should stop every "let's just publish more" conversation dead:

Not 12% more. 12 times more.

One brilliant ad is worth twelve mediocre ones. One genuinely useful article is worth twelve that tick the keyword box. One piece of content that makes somebody feel something is worth twelve that got published because the AI tool made it easy.

The maths of AI content production completely inverts when you apply this finding:

This is the same principle behind why your cheapest leads often cost you the most. Raw volume metrics obscure what actually drives growth. Twelve mediocre blog posts generating zero leads each aren't cheaper than two exceptional ones generating six leads each. They're infinitely more expensive, because they consumed your time and attention for nothing.

Three Moves That Break the Paradox

The Jevons Paradox isn't destiny. It's a default behaviour. Here's how to override it.

If AI cuts your content production time by 75%, don't produce 4x more content. Produce the same amount and spend the other 75% getting it in front of new people. Run it as paid social. Pitch it to industry publications. Repurpose it for channels where your business actually gets found. Share it with referral partners. Sharp's reach doctrine says the biggest growth lever is reaching category buyers who've never heard of you. That's a distribution problem, not a production problem.

The Berkeley Haas researchers noted that the most effective AI users didn't just produce faster - they used the tools to think differently. Use AI to analyse your customer reviews for language patterns. Use it to research competitor positioning. Use it to draft strategy before tactics. Use it to model customer scenarios. The writing is the easy part. The thinking is where the leverage lives.

If the answer is no, don't publish it. Publishing mediocre content isn't free. It trains Google that your domain produces filler. It trains your email list to ignore you. It trains your social followers to scroll past. Every piece of forgettable content has a hidden cost: it makes the next piece slightly less likely to be seen. This connects to what we explored in why Google wants content that doesn't exist yet. The bar isn't "is this good enough?" It's "does this add something new?"

What This Means for Your Business

The Jevons Paradox tells us that when a resource becomes cheaper, humans instinctively consume more of it. AI made marketing production cheap. Businesses responded by producing more marketing. The market - your actual customers and prospects - didn't expand to match.

Byron Sharp's research says growth comes from reaching new buyers, not from talking more often to existing ones. Binet and Field's data says one exceptional piece of creative outperforms 12 mediocre ones by an order of magnitude. And Rory Sutherland's entire career is built on proving that the most valuable marketing ideas come from constraint, not abundance.

The businesses that will win with AI marketing in 2027 aren't the ones producing the most. They're the ones who used AI's efficiency gains to free up time for the things AI can't do on its own: original thinking, genuine customer insight, and creative work that people actually remember.

Your AI can produce a month of content in an hour. The question isn't whether it should. It's what you should be doing with the other 159 hours.

Further Reading

• AI Doesn't Reduce Work - It Intensifies It - UC Berkeley Haas School of Business research on how AI tools increase workload rather than reducing it
• The SEO Productivity Trap: How AI Recreates Jevons Paradox - AirOps analysis of the paradox in content production
• Rory Sutherland on Fat-Tailed Marketing - WFA interview on why creativity outperforms efficiency
• State of AI in Marketing 2026 - Averi AI's survey data on content volume, quality, and results
• Evidence-Based Marketing with Byron Sharp - Ad Age on why reach matters more than loyalty or frequency