Google Doesn't Want Better Content. It Wants Content That Doesn't Exist Yet.

Search "best Google Ads tips" and read the top ten results. Then try "how to improve my website conversion rate." Or "digital marketing for small business."

You'll notice something. They all say the same thing.

Different websites, different authors, different word counts. Same advice, same structure, same recycled talking points. One page says "optimise your landing page," the next says "improve your landing page," the third says "enhance your landing page experience." Same idea wearing three different outfits.

Google has noticed too. And in March 2026, it started doing something about it.

Google Built a Sameness Detector

Google has a patented ranking signal called Information Gain. It scores every page by how much new information it contributes beyond what the other pages ranking for the same query already say.

Not better information. Not more information. New information.

The March 2026 core update made Information Gain a dominant ranking factor for the first time. Early analysis found that sites with original research and first-hand evidence gained 15 to 25% visibility, while templated content dropped 30 to 50% in affected verticals.

Read that again. Google isn't asking "is this content good?" It's asking "does this content say something the other results don't?"

This changes the game for every business publishing content online. "Better" is no longer the bar. Different is. And not different in tone or style. Different in substance. Different in data. Different in what you actually know that nobody else does.

The Distinctiveness Principle, Applied to Content

Byron Sharp's research at the Ehrenberg-Bass Institute proved something counter-intuitive about brand growth: brands don't grow by being better than competitors. They grow by being distinctive. By being easy to notice, recognise, and remember.

Most marketers misunderstand this. They think distinctiveness means a unique selling proposition, some rational reason you're superior. Sharp's data across 130+ brands in 13+ product categories showed the opposite. Most successful brands have very low perceived differentiation. What they have is distinctive assets: recognisable colours, shapes, sounds, and patterns that make them easy to pick out of a crowd.

The same principle now applies to content.

We've written before about why Google stopped rewarding good SEO and started rewarding good businesses. Information Gain is the mechanism behind that shift. Google doesn't need another well-written article about email marketing best practices. It has thousands. What it needs, and what it now actively rewards, is content that contains something those thousands of articles don't.

92.9% of Traffic Winners Own Something You Can't Copy

Cyrus Shepard, founder of Zyppy, analysed more than 400 websites to identify what separates the sites gaining organic traffic from the ones losing it. He measured each feature's correlation with year-over-year traffic change using Spearman correlation coefficients.

One finding stood out: 92.9% of the sites gaining traffic owned proprietary assets that are difficult to replicate. Datasets, original images, studies, tools, or products that can only come from that source.

The five traits Shepard identified, ranked by correlation strength:

Here's the part most businesses miss: Shepard also tested features he expected to matter but found no correlation with traffic changes. These included first-hand experience, personal perspectives, user-generated content, and "uniqueness of information" as a subjective quality.

The distinction is critical. Having a unique opinion doesn't help. Having unique data does. The difference between "I reckon Google Ads works best for tradies" (opinion) and "across 47 trade service accounts we manage, the average cost per lead dropped from $85 to $31 when we restructured campaigns around offline conversion data" (proprietary data) is the difference between content that Google treats as noise and content it treats as signal.

AI Search Makes This Even More Urgent

If Information Gain matters for traditional Google search, it matters even more for AI.

ChatGPT, Perplexity, and Google's AI Overviews don't just rank pages. They cite them. And the research shows they are extremely selective about what earns a citation.

Content with original statistics gets cited up to 40% more than content without them. Comparison content ("X vs Y") is cited 2.8x more than generic feature pages. These are the patterns of an AI system that's actively looking for content that adds something new to its answer.

And the traffic that AI search sends your way? It converts at a fundamentally different rate.

A Seer Interactive case study found ChatGPT referral traffic converted at 15.9%, compared to 1.76% for Google organic on the same site. That's a 9x difference. The volume is still small (Google organic traffic is roughly 47x larger than ChatGPT referral traffic), but pages cited in AI Overviews see a +35% increase in organic click-through rate compared to organic results that aren't cited.

This is exactly what we explored in why your business now has two customers. Human buyers and AI systems both want the same thing: content that tells them something they can't find anywhere else. The difference is that AI systems are explicit about it. They cite what's novel and skip what's redundant.

What "Proprietary Data" Actually Looks Like for a Small Business

"Proprietary data" sounds like something that requires a research department and a six-figure budget. It doesn't.

Every business that serves customers generates data that nobody else has access to. You just aren't publishing it.

A plumber who tracks their average job completion times across different suburbs has proprietary data. An accountant who knows the most common BAS errors their clients make has proprietary data. A Google Ads agency that can show the real cost-per-lead across 30 trade service accounts has proprietary data.

Avinash Kaushik calls this "AI-resistant content" in his AEO series: content of genuine novelty and depth that AI can't generate because it requires access or expertise the model doesn't have. AI can summarise what's already published. It cannot create data that only exists inside your business.

Here's what proprietary data looks like at different levels of effort:

This connects directly to why creating content isn't the same as creating information. Content is a format. Information is what gives that format value. And Information Gain is now how Google measures the difference.

The Uncomfortable Maths of Content Sameness

Here's what most businesses don't realise about their content strategy.

If you hire a freelance writer to produce a blog post about "5 tips for better Google Ads," they will research it by reading the top ten results and synthesising them. The output is a competent summary of what already exists. Its Information Gain score is effectively zero.

Google sees this clearly. So does ChatGPT. Neither system needs another summary. They need something that changes the answer.

Rory Sutherland makes an adjacent point in Alchemy: "It doesn't pay to be logical if everyone else is being logical." Logic gets you to exactly the same place as your competitors. In content, "best practices" are the equivalent of logic. Following them perfectly means your content looks, reads, and contains the same things as everyone else's. The most valuable content is the content that no one else would publish because it requires access, effort, or a willingness to share what most businesses keep private.

This is why Shepard's study found no correlation between "uniqueness of information" as a subjective quality and traffic gains, but a strong correlation for proprietary assets. Subjective uniqueness is just a different writing style on the same facts. Proprietary assets are facts that don't exist elsewhere. The first is differentiation. The second is distinctiveness. Sharp proved that only one of them drives growth.

What This Means for Your Business

Stop trying to write better versions of content that already exists. Start publishing things only you can know.

The businesses that figure this out first will be the ones that both search engines and AI systems treat as sources, not summaries. And as we've explored before, the difference between being a source and being a summary is the difference between earning traffic and watching it disappear.

Further Reading

• 5 Data-Backed Features of Websites Winning Google in 2026 - Cyrus Shepard's full study of 400+ sites identifying the traits that predict organic traffic growth
• How LLMs Search for Citations: What They Find - Deep analysis of AI citation patterns and what content earns mentions
• Information Gain SEO: The Google 2026 Ranking Signal - How Google's Information Gain patent works in practice
• The State of AI and SEO in 2026 with Lily Ray - Lily Ray's keynote on citability as the new visibility metric
• ChatGPT Referral Traffic Converts at 15.9% - Case study data on AI referral conversion rates vs organic