Your algorithm is (wrongly) telling you yes.
I've just come back from a whole week off work, spent relaxing in Mallorca. Wandering down cobbled, colourful streets, eating far too many tapas dishes and maximising time in the last of the summer sun.
Of course, I turned to social media for recommendations on where to visit, and with Mallorca a current TikTok favourite, my notes app soon filled with things I simply *had* to do.
One of the top recommendations was heading to a wine tasting in a vineyard - idyllic. Sign me up. But no vineyards were easily accessible from my location — I obviously wasn't planning on driving or cycling home, and taxis on the island can be a bit of a nightmare. So naturally, my mind wandered to how, in a dream world, I would buy a vineyard and set up a winery and wine tasting experience in that area. The dream job in the dream location, just a simple lottery win required.
As though it heard me, every social media app began to show me countless examples of booked-out wine tours in the Mallorcan mountains. People complaining they couldn't get a booking because everything was already booked up. Influencers promoting the days out to all their followers. This proved the business idea was already in demand; why shouldn't I be the one to supply it?
Fortunately, AI was there to tell me I absolutely should!
If I start with the experience first (before my lottery win that would unlock “a real asset that appreciates while it builds a brand”), then it's “easy to test, easy to scale, and you control the experience end-to-end.”
What a no-brainer! Thanks Claude.
Unfortunately, I know this isn't as easy as my algorithms make it out to be.
There's so much to think about: the seasonality, licensing issues, competition, and the fact that I don't actually have any experience in (besides the drinking of) wine, to name just a few.
But how easy would it be to get swept away and put a significant chunk of my life savings into a business venture, based on an algorithm showing me what I want to see, and a robot backing that up with what I want to hear?
This is the perfect example of an increasingly alarming pattern I'm noticing with business owners this year. More and more potential clients are coming to me, genuinely perplexed that a new product or direction hasn't landed.
The general pattern is:
The problem with the increase in personalisation in all tools we use is that we're seeing an increasingly narrow view of the world.
Social feeds show us only what we're already interested in, or what fits into our curated, small set of desires.
AI tools that want to please us are again reinforcing what we are already doing, or suggesting new things based on existing interests.
Less real-life socialisation means we're exposed to fewer people with different views, backgrounds, and interests.
We are becoming more and more disconnected from those outside our box. So it's no surprise that when we have an idea we think is great, we don't understand why everyone else doesn't agree.
You're being flattered by your idea
It's easy to mistake scrolling social media feeds or talking to AI as research. They're not.
They're a hall of mirrors, built to agree with you, and perhaps even more importantly, keep you using them. They can trap you by showing similar success stories, drip-feeding you serotonin boosts through “proof” that your ideas are good. Always surfacing an “and” of something else you could do to keep you hooked.
This is all simply evidence that social media and AI are doing exactly what they were built to do: handing you a shiny version of what you walked in wanting. Think of them like a sleazy salesman that flirts with all the women in order to get record-breaking sales and tips, but a few steps back will make you realise they're just a bit icky. Definitely not the person you're going to for actual business advice, even if you want them on your sales team.
Plus, this is all for them to answer the wrong question. They might prove that an overall category is wanted, but that's not the same as your specific take on that being wanted.
It's likely true that a lot of people in Mallorca want a wine tour. But would they want my wine tour? What do they actually want from a wine tasting experience, and is that what I am offering?
Visible demand for a broad category alone is a pretty useless signal. Yet it's the exact thing your feed is built to serve you on a loop. Real demand is much harder to find, but much more valuable when you do. And sadly, the latter isn't on your For You page.
To come back to my Liquid Death example from a few weeks ago: on their social feeds, I am sure they were confident the world was crying out for their product. They had festival sponsorships, and no doubt a feed of people who loved the brand and were begging for it to come to the UK. I'm sure they launched in the UK market confidently – no doubt an AI tool would've told them it was a no-brainer. But they lasted no more than two years before withdrawing. Leaving with roughly £2m in sales vs Evian's £207m.
The UK market might've posted photos of them with Liquid Death cans and said they loved it on their social channels, but when it came down to buying bottled water, the cheaper, plastic-bottled alternatives solved the customer problem better than the more expensive, metal-canned equivalent Liquid Death offered. The social media hype wasn't to be believed, and Liquid Death wasn't actually the solution their customers wanted. That's the question your digital tools cannot answer.
What does your customer actually want? What does your solution actually provide to meet those underlying needs?
The best solution likely isn't the most visible
If every app has begun to show you your own idea succeeding elsewhere, and if the AI tells you it's easy, scalable, and a no-brainer, then don't treat that as a green light. It's just an existing solution already being visible.
The questions worth asking are the ones your feed can't answer for you: what can you offer as a new and improved solution to their problem? Are they already paying for a worse version of something you could make? Or is there an entirely new idea that would better slot into what they're asking for?
My algorithm already knows what I like. It can show me a hundred booked-out wine tours, but it can't tell me whether a single person would want to book my take on that. The real world is the only thing that knows what I don't: whether the idea's any good, and who it's actually for.
So the next time your feed and your favourite chatbot agree you're onto something, remind yourself to treat that only as a starting point, and begin to build out your own evidence in the real world.
An algorithm will never tell you no, which is what makes its yes worth so little.
