Labour Leads to Love: The Hidden Bias in Every Website You Build

There’s a particular feeling that arrives the moment you finish building something. You lean back, look at the screen, and think: that’s actually really good.

These days that feeling comes faster than ever. You describe a website to an AI design tool, watch it assemble itself in front of you, nudge a few things, and within an hour you have a living, breathing site that looks like a real company made it. The rush is real. You made that. And it’s good.

We want to talk about that feeling, because it happens to be one of the most well-documented biases in psychology, and it has real consequences for anyone designing a website or product today, whether they’re using AI, a page builder, or a team of professionals.

Labour leads to love

In 2011, three researchers, Michael Norton at Harvard, Daniel Mochon and Dan Ariely, ran a now-famous set of experiments (later published in the Journal of Consumer Psychology). They had people assemble plain IKEA storage boxes, fold origami, and build Lego, then asked how much each creation was worth.

The results were lopsided. People were willing to pay roughly 63% more for furniture they’d assembled themselves than for the identical, pre-built version. They rated their own wonky origami as about as good as an expert’s. And, this is the part that matters most, they assumed everyone else would see it the same way.

Norton and his colleagues called it the IKEA effect, and summed it up in four words: labour leads to love. The effort you put into making something doesn’t just make you proud of it. It quietly inflates how good you think it actually is.

It’s not a beginner’s mistake

Here’s the uncomfortable bit, and the reason this isn’t a post about other people. The effect didn’t only catch beginners. It hit committed do-it-yourselfers just as hard. Experience at building things granted no immunity at all.

There was, however, one condition that switched it on: you had to finish. When people built something and then took it apart, or failed to complete it, the warm glow disappeared. The love depends on a completed product.

Which is where AI design tools do something genuinely interesting. They all but guarantee completion. You will end up with a finished, working website, that’s the entire promise. So the one prerequisite the bias needs is handed to you, every single time, in minutes.

The tools are brilliant. We build with them every day. But they strip out the friction that used to act as a natural reality check, and they reliably deliver the exact conditions under which humans most overvalue what they’ve made.

You can’t see your own work

There’s a second bias stacked on top of the first, and together they’re potent.

It’s called the curse of knowledge, and the cleanest demonstration of it is a 1990 study by Elizabeth Newton, then a PhD student at Stanford. She had people tap out the rhythm of a well-known song on a tabletop and predict whether a listener would recognise it. The tappers were confident, surely it’s obvious. The listeners almost never got it. Inside the tapper’s head the song was playing in full, all the listener heard was tapping.

When you design a website, you are the tapper. You know what every button does and where every link goes, so the navigation feels obvious, the flow feels natural, the clever bit feels clever. The visitor arrives cold, hears only the tapping, and gets lost on a page you were certain explained itself. You literally cannot un-know what you know, which means you cannot experience your own site the way a stranger will.

The honest part

So can a professional designer simply see past all this? Not really, and we’d be overselling ourselves if we claimed otherwise. We feel the IKEA effect too. We fall for our own drafts. The research is blunt on this point: knowing a bias exists barely dents its grip on you.

What experience actually changes is two things. The first is calibration. When you’ve shipped a few hundred designs and watched how real people used them, your sense of “this is good” is anchored to a lot of outcomes, rather than to a single rush of pride. The second, and more important one, is that you stop trusting your own eye. You learn the hard way that the only reliable way out of a bias you can’t introspect away is to build structure around it: critique from other people, testing with real users, a day away from the screen before you decide anything is finished.

That’s the whole game. The antidote to creator bias was never talent, and it was never even experience. It’s distance. And distance is the one thing you cannot manufacture for your own work, because the bias is strongest precisely where you’ve done the most.

Where this leaves AI and your next website

None of this is an argument against building things yourself, or against using AI to do it. We’re genuinely enthusiastic about both. AI design tools have put real creative power into the hands of people who would never have touched it otherwise, and that is a good thing.

It’s a word of caution, and it applies to everyone, us included. If you’ve built something with an AI tool and you love it, that love is real, but it isn’t evidence. It might be a great site. It might also have a confusing checkout, a hero section that means nothing to a first-time visitor, or a structure that makes perfect sense only to the person who built it. You are the last person who can tell the difference. So are we, about our own work.

The most useful thing you can do, the thing the research quietly insists on, is to get a set of eyes that didn’t do the building. Someone with no labour invested, no curse of knowledge, no rush of completion clouding the view. That isn’t a knock on what you made. It’s simply the only known way to find out whether the thing you love is also the thing that works.

If you’ve built something you’re proud of and you’d like that second, unattached perspective on it, that part happens to be our favourite part of the job. And the help doesn’t stop at an opinion. We can evaluate a design and tell you honestly where it’s working and where it isn’t. We can take a design you love and build it into a real website or product properly. Or, if you’ve already built a website or product, with or without AI, we can review the code under the hood, point out what’s fragile or insecure, and improve it so it’s something you can actually grow on.