Everyone Can Make Ads Now. Almost Nobody Knows Which Ones Are Good.

Global ad spend rose 33% between 2023 and 2026. Marketing effectiveness fell nearly 20% over the same period.

More money. More tools. More content. Worse results.

If you're an SME spending $3,000 a month on Google Ads and Facebook Ads, this trend should worry you. The thing that used to separate great marketing from mediocre marketing, the ability to produce polished creative at speed, just got commoditised overnight. And the thing that actually determines whether your marketing works has nothing to do with the tools you're using.

It's a skill that can't be installed, prompted, or automated. It's called taste.

The Bottleneck Disappeared. The Results Didn't Improve.

Every day, 4.4 million new blog posts go live across the internet. 97% of content marketers plan to use AI to support their work in 2026, up from 83% in 2024. Over 80% of enterprises have deployed generative AI in production environments.

Production is no longer the constraint. Any business can generate 50 ad variations in an afternoon, write a month's worth of social posts before lunch, and build landing pages without touching a line of code.

So why isn't all this marketing working better?

Because production was never the hard part. The hard part was always knowing which of the 50 variations was actually good. And that skill, the ability to look at a piece of marketing and know whether it will work before you spend a dollar testing it, is the one thing AI hasn't replaced.

If anything, AI has made it more valuable. We've written before about why AI makes most marketing faster but not better. The data behind that argument has only gotten stronger.

The Creativity Crisis That Started Before AI

Les Binet and Peter Field have spent decades analysing the IPA Effectiveness Databank, the most comprehensive record of what makes advertising campaigns actually work. Their findings reveal something uncomfortable.

Between 1996 and 2008, creatively excellent campaigns were 12 times more efficient than average ones. Same media spend, radically different results. By 2018, that multiplier had collapsed to less than four.

Today, creatively awarded campaigns are no more effective than non-awarded ones. The link between creativity and effectiveness hasn't just weakened. It's broken.

AI didn't cause this crisis. But it poured fuel on the fire by making it trivially easy to produce more of the short-term, forgettable content that was already failing.

What's Actually Missing

Tomlinson defines taste the way a sommelier would: it's developed through 1,000+ deliberate tastings. Not by reading about wine. Not by following a formula. By repeated exposure to work across a wide range of quality, combined with the active effort to understand why something works or doesn't.

The absence of taste in marketing shows up as three symptoms: incoherence (every touchpoint feels disconnected from the next), mundanity (nothing is distinctive or memorable), and interchangeability (your ads could belong to any competitor in your category).

Sound familiar? That's what most AI-generated marketing looks like when nobody with judgment is steering it.

The editorial team at Every, one of the most AI-native publications in the world, demonstrates what taste-led AI use actually looks like. They use AI at every stage of their process, from research to drafting to final polish. But here's the critical detail: every piece of output runs through human editorial judgment before it goes live. Their social media manager described his role like being a DJ: when AI can generate 50 variations in seconds, taste is what makes the difference between a post that sounds like the brand and one that sounds like every other newsletter on the internet.

Their staff writer observed something equally telling: with the mechanical effort of putting every word after the other taken off her plate, she had more mental bandwidth to think about the craft going into the piece. Whether the introduction was compelling. Whether the thesis was solid. Whether the writing sounded like a specific human wrote it.

AI didn't replace her judgment. It freed up bandwidth for more of it.

Why Best Practices Produce Average Work

In 1963, BBDO creative director Fred Manley gave a satirical presentation called Nine Ways to Improve an Ad. He took the legendary Volkswagen "Think Small" ad, widely considered the best ad of the 20th century, and applied nine standard advertising best practices to it.

Make the product bigger. Put the product name in the headline. Add a testimonial. Make the logo larger. Show people enjoying the product.

By the ninth improvement, the ad was indistinguishable from every other car advertisement of the era. Every "improvement" made it objectively worse.

Manley's point, made over six decades ago, has never been more relevant: following every best practice produces the most average possible work. The VW ad was great precisely because it broke the rules. It used negative space. The headline was two words. The car was tiny in the frame. Everything about it felt wrong by the standards of the day, and that wrongness is exactly what made it impossible to ignore.

This is the fundamental limitation of AI-generated marketing. AI is trained on the average of everything that came before it. It produces work that is, by mathematical definition, derivative. It can generate a thousand Google Ad variations, but every variation will converge toward the median. It follows every best practice perfectly. And as Manley demonstrated, that's the fastest path to being ignored.

This connects to why following every marketing rule makes your business invisible. The rules exist to prevent bad outcomes. But preventing bad outcomes is not the same as creating great ones.

The 14% Gap That Compounds Into a Different Business

If the taste argument sounds subjective, the data isn't.

That might sound modest. It isn't.

Ipsos has separately found that ads scoring well on their Creative Effect Index produce 44% higher average sales lift. A 14-point gap in creative quality doesn't translate to 14% less revenue. It translates to a completely different growth trajectory over 12 months of campaigns.

As we've explored in why your competitors aren't beating you with better ads, the visible quality of an ad isn't always what determines its effectiveness. Two ads can look equally polished while one generates $30 leads and the other generates $130 leads. The difference is taste applied at every decision point: which hook to lead with, which benefit to emphasise, which image creates the right emotional response, which words to cut.

None of these decisions can be A/B tested in advance. They require judgment.

What This Means for Your Business

This isn't an argument against using AI. It's an argument for knowing what to do with what AI gives you.

Binet and Field's research proves that creative quality is a multiplier on media spend, not an addition to it. A $3,000 monthly ad budget with strong creative can outperform a $10,000 budget with mediocre creative. The difference isn't in the platform, the targeting, or the bid strategy. It's in the 50 small judgment calls that shaped the ad before it ever went live.

The tools have never been better. The question is whether you have the taste to use them.

Further Reading

• Taste Is a Competitive Advantage by Sam Tomlinson: why AI can't replicate the accumulated judgment that separates distinctive marketing from noise
• AI Ads Are Good Enough, And That's the Problem by Ipsos: the full research behind the 14%/17% effectiveness gap between human and AI creative
• The Crisis in Creative Effectiveness by Peter Field for the IPA: how the creativity-effectiveness link collapsed across 24 years of campaign data
• Nine Ways to Improve an Ad by Fred Manley (1963): the satirical masterclass on how best practices kill great advertising
• The Long and the Short of It by Les Binet and Peter Field: the foundational research proving creative quality is a multiplier on media spend