SEO is yesterday's game, Boomer
The new game is context.
In Little Shop of Horrors, the plant gets bigger every time someone feeds it. Audrey II has one demand, repeated with mounting urgency: “Feed me, Seymour!”
She also, at one memorable moment, says “No shit, Sherlock.”
We’ll get back to that.
When did you last use (or trust) Google’s first page?
Be honest. For most of us (and I’d wager you in particular, if you’re the sort of person who reads pieces like this), the default has quietly shifted. The first move is now ChatGPT, or Claude, or Perplexity. And it’s entirely possible there’s a new model du jour between me writing and publishing this. Google still gets a look-in for “restaurants near me” and “padel court hours.” For everything else, the questions with nuance, the ones that matter, you’re talking to a model.
So is everyone else. Including the people deciding whether to do business with you.
The conversation you’re not in
Try this…
Open your preferred model and ask it about your company…
Read the answer carefully…
Notice what’s there…
Notice, more importantly, what isn’t there, what’s slightly wrong, and what tone the model has settled on.
Someone wrote that. It just wasn’t you.
The model assembled its answer from whatever was lying around: a competitor’s better-positioned blog, a Hellopeter complaint from a customer who didn’t get their refund in 2021, a Reddit thread where your company name was misspelled, the press release your PR agency wrote when you launched a product you’ve since discontinued. All of it goes in. All of it! And it shapes what gets said about you when a prospect, a journalist, an investor, or a future hire types your name into the box.
You’re being described by strangers, in rooms you’ll never enter, to audiences you’ll never meet. And the description sticks.
SEO won’t save you
For twenty-odd years, the game was density. Pack your page with the right words, in the right places (and even the right font sizes), often enough, and Google would float your blue link above the other blue links. That was the craft.
It was a craft built for a specific reader: a human typing keywords, scanning a list, picking one. That reader is leaving. The new one asks a question in plain English and reads a single synthesised answer. There are no blue links to climb above. The only question is whether your business gets mentioned at all, and whether what gets said is what you’d choose to say.
Density doesn’t help here!
The model isn’t counting words. It’s working out who has the context.
That’s the shift worth pausing on. The old game was content: produce more of it, stuff the right keywords in, hope the algorithm rewards the volume.
The new game is context: who actually understands the territory, who can speak to the nuance, who’s been in the room.
Content can be manufactured. Context has to be lived. And the difference between the two is exactly what the models are now sophisticated enough to detect.
What it looks like when you win
Picture the version of this where it’s gone right.
A prospect asks ChatGPT about your industry. Your business comes up by name, with a credible summary of what you do and why you’re worth a look, drawn in part from things your own people have actually said. A journalist researching a story gets a quote that sounds like you because, well, it is you. A potential hire forms an impression of your culture that matches the lived experience. An analyst building a market report finds your point of view well-represented and well-articulated, because (again) you put it there.
The models become a distribution channel that runs while you sleep. Your expertise compounds. Your reputation gets built by the same machinery that’s currently eroding it.
That’s the prize. Now the question is how you get there.
You already have what the models want
Somewhere in your business, probably in more than one head, there’s an enormous amount of expertise. Years of it. Pattern recognition, war stories, frameworks, the unwritten rules of how your industry actually works. The kind of knowledge that takes decades to acquire and minutes of conversation to demonstrate.
The models are ravenous for exactly this. Not marketing copy. Not press releases. The real stuff: substantive, plain-spoken, opinionated, voiced by people who know what they’re talking about. They want it consistently. They want it in depth. They want to recognise the source.
Most businesses haven’t worked out how to get it out of the heads and onto the internet at any meaningful scale. Writing is slow. Experts are busy. A blog post takes a week and produces 800 words.
The unfair advantage
Now do the maths.
A blog post: a week of effort, 800 words.
A podcast: an hour of conversation, 8,000 words.
Same expert. Same expertise. Ten times the output for a tenth of the friction.
If you’re keeping tally, that’s not a mere marginal improvement, it’s two orders of magnitude of leverage. That’s a different category of leverage on the knowledge already sitting in your people’s heads. You’re not asking them to write more. You’re asking them to talk, which is what they were going to do anyway, in meetings and on calls and over coffee, except this time someone hits record.
Do it quarterly and you’ve laid down a foundation. Do it monthly and the models start to recognise you. Do it weekly and you’ve out-positioned every competitor in your sector. Not by working harder. By capturing the expertise that’s already happening in conversations, meetings, and corridor chats, and pointing it at a microphone instead of letting it evaporate.
Get those conversations transcribed, sitting on a domain you control, marked up so the models can find them, and you’ve quietly become the authority in the room. The next time someone asks an LLM about your industry, the model has more of your context to draw on than anyone else’s. It cites you. It quotes you. It frames the answer around your point of view.
8,000 easy words of context beats 800 hard words of content.
No shit, Sherlock.
You’re already doing the written half
You are, at this moment, reading this on Substack. Which is also being indexed, scraped, and fed to the same hungry models. That’s not an accident. It’s the written-word version of the same play. Long-form, in a recognisable voice, on a platform crawlers like, published consistently enough (well, I try) that the models start to know who’s talking.
One detail worth borrowing: this Substack lives at substack.solidgold.co.za, not on the default solidgold.substack.com URL. Our own subdomain CNAME, ten minutes of DNS work. Same words, same effort, same Substack, but every piece builds our domain reputation instead of theirs.
If you’ve made it this far down, you already understand the written half of the argument. The audio half is the same argument with a microphone, and significantly more throughput per hour of effort.
The choice
The beast is being fed either way. Right now, today, while you’re reading this, someone is asking a model about your industry, your competitors, possibly even you. The model is answering. The answer is being assembled from whatever was lying around when the crawlers came through.
The only question is whose words it reaches for next.
Feed it, Seymour. Or don’t. The plant gets bigger either way.
Solid Gold makes corporate podcasts for businesses that have something worth saying and would prefer the LLMs heard it from them first.