Boring doesn't stop the scroll: how premium brands get seen when everything sounds the same

Do you ever feel as though you and your team spend hours every week, sometimes daily, creating and sharing content – on socials, on emails, in ads – only to receive crickets in return? Like you’re literally just shouting into the void? It’s kind of like the digital version of being the guy outside the empty pizza restaurant wearing the sandwich board flogging $5 pizzas, whilst the Thai joint across the road is absolutely pumping.

The lack of engagement or any kind of visible return-on-effort is pretty demoralising. Not to mention, it’s very easy to compare one’s own lack of engagement and results to what appears to be a very successful competitor’s social media highlight reel…

But the painful truth is, if your content doesn’t stop the scroll, it’s not because your packing-orders Reel lost to a competitor.

It lost to the four hundred similar reels that came before it, which had already taught her that packing videos contain nothing new.

When something you’ve made lands flat, we instinctively assume it wasn’t good enough. Wrong hook, wrong time, wrong format, algorithm’s changed again. That’s human nature, right? So you tighten it up and post more often.

Unfortunately, there was probably nothing wrong with it. Which in itself is a problem.

Now, regular listeners will know I’ve banged on about sameness before, and I have. I came at it from the customer’s side, because a brand that looks like every other brand gives her nothing to choose on except price. That still stands. 

But, today I want to come at it from the other direction entirely, because something has changed that turns sameness from a positioning problem into a discoverability one.

Something has shifted in the last eighteen months that I don’t think most brand owners have fully joined up yet. You’re now being discovered in two completely different ways. A human scrolling, and a machine retrieving. Instagram, TikTok and Pinterest on one side. Google’s AI Overviews, ChatGPT, Perplexity on the other.

They work nothing alike. One is a person deciding in a fraction of a second whether to stop. The other is a system deciding whether your page is worth pulling into an answer.

And here’s the thing that surprised me when I sat with it properly. They fail you for exactly the same reason. The human doesn’t stop because she’s seen it before. The machine doesn’t cite you because it’s already got it from somewhere else. Same disqualification. Familiarity.

Let’s do the human side first, because that’s where most of you are feeling this daily.

You post consistently. The photography is beautiful. The captions are considered. And it goes out into a sort of polite silence. A handful of likes from people who already buy from you, and nothing that looks like discovery.

The reflex is to assume you’re doing it badly. Wrong format, wrong time, wrong hook, algorithm’s changed again.

Usually you’re doing it perfectly well. You’re just doing something recognisable.

Because a person scrolling is not evaluating your content. She’s pattern-matching at enormous speed, and the only real question her brain is answering is: have I seen this already?

Not this exact post. This shape. This kind of image, this kind of claim, this kind of tone. And if the answer is yes, she’s gone before any conscious judgement takes place. Your beautiful flat lay didn’t lose to a better flat lay. It lost to the four hundred flat lays that came before it, which taught her brain that flat lays contain nothing new.

This is why “post more consistently” so often does nothing. Consistency multiplies whatever you’re already doing. If what you’re doing is recognisable, consistency just makes you recognisably invisible, more often.

And it’s got worse recently, for a reason nobody really wants to say out loud. An enormous number of brands are now producing their social content with AI. Same tools, same prompts, same hook structures, all drawing on the same pool of what already exists. So the category converges. Everyone’s captions start sounding like they came from the same person, because in a sense they did.

Which means the amount of content your customer has to scroll past to reach something genuinely unfamiliar has gone up sharply. The bar for stopping her has moved, and most brands have responded by producing more of the thing that isn’t stopping her.

Which brings me to the other half, and this is where it stops being my opinion and starts being documented.

In May this year, Google published its first proper guide to showing up in AI search. AI Overviews, AI Mode, all of it.

Most of the coverage focused on the technical section. That’s not the interesting part. The interesting part is that the strongest signal in the entire document is editorial.

Google’s term is non-commodity content. And they say plainly that creating content people find unique, compelling and useful will likely influence your presence in generative AI search more than any other suggestion in the guide. More than anything else they list.

Their example is worth sitting with, because it’s brutally clear. Commodity content is an article called “7 Tips for First-Time Homebuyers.” Common knowledge. Could have come from anyone. Adds nothing you couldn’t get in fifty other places.

Non-commodity content is an article called “Why We Waived the Inspection and Saved Money: A Look Inside the Sewer Line.”

Same category, same reader. One is a summary of what everybody already knows. The other happened to a specific person who is now telling you what they learned.

Then Google adds the line that should genuinely stop you. Don’t just recycle what others on the internet have already said, or what could easily be produced by a generative AI model.

Read that second half again if you’ve been using AI to write your blog posts.

I want to be precise here, because there’s a version of this going around that isn’t accurate. Google has not said AI-generated content is penalised. There’s no statistic showing AI content underperforms. Its position on using AI tools is that it’s fine, provided the output meets the same standards as anything else.

The penalty isn’t on AI. The penalty is on sameness.

It’s just that AI has made sameness enormously cheaper to produce, in exactly the same way it has on social. And the mechanism is simple. These systems work by retrieving pages and using them to ground an answer. If your content contains nothing that isn’t already sitting in forty other places, there’s no particular reason to pull yours. Not because it’s bad. Because it’s redundant.

So look at what we’ve actually got. A woman scrolling won’t stop for something she’s seen the shape of before. A retrieval system won’t cite something it can already get elsewhere. One’s instinctive and one’s computational, and they’ve arrived at the same standard.

Here’s why I think this is genuinely good news if you’ve built something real.

There’s exactly one input that can’t be produced by a machine or replicated by a competitor, and that’s what actually happened to you. Not “we’re passionate about quality.” What happened.

Let me show you the difference, because it’s stark.

The boring version: we use organic ingredients sourced from sustainable farms to create effective, luxurious formulations for discerning customers who value quality and environmental responsibility.

Now the same brand, telling the truth: when Sarah’s daughter developed severe eczema at three, every dermatologist prescribed the same harsh chemicals that made it worse. As a trained chemist, Sarah knew there had to be a gentler option. Two years of kitchen experimentation with botanical extracts from the family farm produced the first formula that actually healed her daughter’s skin.

Same information. Organic ingredients, effective formulation, environmental care. It’s all in there.

But a machine could have written the first one. A machine could not have written the second, because the second requires a daughter, an eczema diagnosis, a chemistry degree and two years of failed batches on a kitchen bench.

And notice it passes both tests at once. It’d stop a scroll, because you haven’t read it before. And it’d get retrieved, because there’s nowhere else to get it.

That’s the whole test, really. Could anyone have written this? If yes, you’ve made wallpaper.

I call the thing Sarah has an origin catalyst, and the distinction matters, because most brands think they’ve got one when what they’ve actually got is a founder bio.

A founder bio says: I was passionate about this category and wanted to make something better. Lovely, and something every one of your competitors could say word for word.

An origin catalyst is the specific problem, frustration or realisation that gave you a different view of your category than everyone else in it.

Patagonia’s wasn’t “we love the outdoors.” It was Yvon Chouinard realising the climbing gear he was making was damaging the rock faces he loved. That one specific discomfort produced the equipment innovations and eventually the entire environmental position.

Frank Body’s wasn’t “we wanted to make skincare.” It was noticing cafés were throwing out enormous quantities of used coffee grounds that could be doing something useful.

In every case the catalyst is small, specific and slightly awkward. That’s precisely why it’s uncopyable. Your competitor can match your price, your packaging, your ad format and your ingredient list. They cannot have had your realisation.

Here’s where I’d push back on how most brands run at this, though.

When founders hear “tell your story,” they nearly always tell the wrong one. They write about themselves. The founding, the journey, the values, the mission statement.

Your customer is not, on the whole, that interested in your history. She’s interested in her own.

The story that actually converts a considered buyer is the story of what changes for her. Not softer skin, but confidence in her own skin. Not better cookware, but becoming someone who cooks properly. There’s a version of herself she’s moving toward, and your brand is either part of that or it isn’t.

This is also, incidentally, where most social content quietly goes wrong. Behind-the-scenes posts, packing orders, the new stock arriving, the founder at a market stall. All perfectly nice. All about you. And the reason it doesn’t stop the scroll isn’t that it’s badly made. It’s that she’s watching someone else’s life rather than recognising her own.

The content that stops her is content where she sees herself.

And I’ll be honest with you about this part. Working out what that transformation actually is for your specific customer, in her words rather than your assumptions about her, is the hardest work in this discipline. You can’t introspect your way to it over a coffee. It comes from asking her properly, and knowing what to ask.

That’s a bigger conversation than one episode. But I’d rather tell you it exists and that it’s hard than hand you a five-step formula that produces the same wallpaper as everyone else’s five-step formula.

While I’ve got you, Google’s guide also has a myth-busting section, and it’ll save some of you money.

You don’t need an llms.txt file. Google Search doesn’t use them. You don’t need to chop your content into little chunks for machines to digest. There’s no ideal page length. You don’t need to rewrite everything in some special AI-friendly voice. These systems understand synonyms and meaning. Structured data isn’t required for AI search. Keep it for rich results, but it isn’t the lever people are selling it as. And chasing mentions across the web for the sake of having mentions doesn’t work the way you’d hope.

Google’s position, stated about as directly as Google states anything, is that optimising for generative AI search is still just SEO. There’s no separate discipline. Anyone selling you one should be treated accordingly.

There’s also a new Generative AI performance report in Search Console. Go and open it. It’ll show you how you’re actually turning up.

So here’s what I’d do this week.

Take your three most recent pieces of content. A blog post, an email, a Reel.

For each one, ask the only question that matters now: could a competitor have published this with their logo on it and nobody would have noticed?

If the answer is yes, that’s commodity content. It isn’t wrong. It’s just not stopping anybody and it’s certainly not getting cited.

Then go looking for the material that fails that test in your favour. The awkward specific reason you started. The thing that went wrong that you fixed. The customer whose situation changed how you formulate. Not the polished version. The actual one.

That’s what gets retrieved. That’s what gets remembered. And it’s the only thing your competitors genuinely cannot have.

Because they can copy nearly everything else. They can’t have been there.