This website cost over $12 Million

Does it convert great?

Let’s make it clear right from the start. They bought the icon_dot_com domain name for $12 Million. The website likely cost a fraction of that. 

But Icon is causing controversy not only because of reckless overspending, but also some of their design choices. Does this even convert? 

They don’t share a lot of real data on conversion, but the website apparently did change a couple of times already. 

There is more controversy with that site however. It’s not just the design choices, as there’s some deceptive tricks that are supposed to manipulate your psychology. 

Are they ethical? I leave it for you to decide. 

Designers are also (rightfully) asking the question now: does this condensed (asian style) design even work in the western world? 

Let’s break it down

I have my own process of analysing design, based on my past knowledge and experience of the good practices (heuristics). 

I did notice some interesting and some scary patterns here that I think are worth talking about. Most designers seem to have missed them.

How we see it?

I simulated a heatmap of where the eye would focus and in what order. These things aren’t actually that difficult to guess after seeing hundreds of real ones for years. It’s all repeating patterns. 

Here we first look at the large and very short headline (1). Then instantly towards the face (2). Then to the ad that’s the most colorful and visually heavy (3), followed by one that matches the website stylistically (4) and only then our gaze goes to the CTA (5).

The face however catches our attention the most, and the gaze seems to go back to it between all other steps. 

Segmentation

Let’s start by dividing the hero into logical segments. Since it’s a pretty condensed hero section, there’s more of those than usual. And they are: 

Of course we have a traditional “what” section, explaining what the product is and clearing some early doubts, followed by CTA buttons. 

We do have a key visual showing examples on the right side, which makes sense since it’s something clients definitely want to see right away.

Then we have two types of social proof, and an interesting section between them. I called it “overwhelm with data” because it is also supposed to convince you to pay, but through information overload. We’ll go in-depth on those sections a little later.

What’s interesting is that the $12M domain name is highlighted in two separate spots on the landing page, supposedly to give it extra credibility. 

The WHAT section

The what section itself is pretty standard. The headline features a typical kicker, headline, two subheadlines and two CTA buttons. 

It starts with a kicker (1) showing early adopter urgency. That kicker doesn’t specify “what” exactly is it for, but uses the standard “only 4 spots left”. However, a few days ago it said the exact same thing, but the end date was the second of May. 

Now on May 5th it still says 4 spots are left, but now the data is the 9th of May. No idea if this moves automatically, through a script, or someone changed it manually. Apparently however those 4 spots weren’t filled this week and they still remain. 

This is a classic example of the “fear of missing out” dark pattern that booking sites use a lot as well. You’ve probably seen this around a lot.

It’s supposed to convince you to click through a sense of urgency. I believe it’s likely, that there’s way more spots left and the date will be changed indefinitely to give you that constant fear of missing out. 

Like a special promo that’s always ending soon. 

Key visual

The key visual here is actually pretty genius. I really like it! 

The main hero of the visual is an attractive woman (1), but the photo is captured mid-frame. Like it’s a part of an actual reel/video. The subtitles make it feel even more real. 

If this was a posed, smiling photo, that superficial feeling would make it unrealistic and way less convincing. It also says the person in the photo has 2.4 Million (2) followers. Chances of this being made up are very high, as it’s just an example. 

Subconscious impact

But things like that make a subconscious impact on us. We naturally think that successful social influencers are using the tool and we want to be just like them. I want 2.4 million followers as well! 

The simplified stories UI (3) is there to connect better to what it is, but it’s shorter, bigger and easier to understand than the real (smaller UI) would be. All for you to process the information faster. 

Having the video ad out of the way, we have two regular, graphic designs next to it. The first one (4) is a big splash of color that instantly catches attention right after the woman’s face. 

I call it a “regular people” ad. 

This is what most non-designers expect an ad to look like. It’s generic but in that good, expected way. It’s something you’d probably see when you typed “ad” into a search engine. 

To balance it out we also have an ad (5) that’s more minimal, “designer” in nature. It uses a similar color palette to the website itself, so it blends in a bit better and doesn’t make the top section too heavy. 

It’s a perfect blend of photo, heavy visual and minimal that also shows the diversity of outcomes the platform can generate. 

On top of that, there’s a stronger shadow (6) under the whole key visual section. It brings the whole segment more towards the perceived “front” of our eyes. And things closer to us (even when simulated) catch our attention. 

Social proof

The social proof secton has all the bells and whistles we expect. Of course there are logos of brands using it. This is typical.

It also has a star rating (1), and because it’s condensed we quickly jump to the next item without giving it much thought.

But we should.

Who voted for it to be 5 stars? Where did they vote? Can we see how many people voted? No. It’s all an arbitrary metric, that normally would raise suspicion but we have so much stuff here we quickly process it and move on. 

Internal validation (where the users vote for something on the platform itself) is the least trustworthy. Of course most businesses won’t post negative opinions on their own site. It will be curated. 

External validation (like app store rating) are a lot more realistic, so when making your own website focus on those instead. 

Then it uses numbers (2) to show they have users and that those users care (3) through social media engagement. This is a good idea for any business that actually have users — show the numbers. The problem with the user number is the same though. It’s unverifiable. You don’t know if it’s real. 

With social media engagement it is easier to check, but we’ll get back to this in a minute, as it’s quite interesting. 

Then it says they’re funded in 2025 (4) which makes it a “hot startup”. Normally companies wouldn’t brag about not having market experience, but in the AI world this is a part of the fomo tactic. 

This is so new, you probably haven’t heard about it yet! 

And then it says how much the domain cost right in the social proof (5) and then adds a section below. This is to show the company is serious. Well funded. Not messing around! Commited to a vision and believing in it. 

Was the marketing power of the domain worth $12M? We’ll have to wait and see. 

Overwhelm with data

This website would be pretty standard if not for this section. Without it, and with a little more whitespace it would look like any other website. 

And it’s likely I wouldn’t be reviewing it at all. The images on top of each section are there to create chaos. Visual clutter. To make you not want to process ALL of it, but instead look for some visual hooks and move on.

The section is kind of like an “about us”, but it’s using a lot of external “validation by proxy” to achieve that. 

Validation by proxy means you’re using other people’s success to boost your brand

It talks about the founder trough monetary success. That’s good, as it’s at least direct and not proxy. 

But then talks about big investors (Peter Thiel, OpenAI) and previous job experience of the team. 

Now, it doesn’t say what roles at Nvidia, Tesla they had. But we subconsciously associate big brands with quality and assume it’s the same here. 

Then it uses recall from the key visual, by showing the same ad creatives again. Our brain really likes recall. 

The next section is all about showing social media activity to make it even more realistic. 

But is it?

When you look at some of the videos, you notice 528K+ views next to the button. But then you go and see the video itself and it only has 53K views.

That’s because they’re showing a combined number of views of all videos, but this is not what the user is thinking seeing the number.

It’s a small oversimplification that is there to trick you into thinking all their videos are pushing big numbers. 

Most people won’t check.

Overwhelm with data again?

The no one else comes close claim is there to prime you into reading the table below in a certain way. 

It uses popular brand names, as it’s likely you’re already using one of them and seeing that name there will have you pause and try to compare it. 

But the table itself is super long and complex on purpose. You’re not really supposed to read it. You’re supposed to scan it and just see all the checks in the first column. 

But to me it falls apart in the last four items. 

Tier 1 investors and tier 1 engineering team may be true, but showing that only OpenAI checks those boxes is another comparison by proxy. We assume they’re all as great as openAI, which may or may not be the case. 

However bragging about a 7-day work week may please investors and show commitment, but it’s definitely not a good workplace. Overworking and burnout are not things to brag about and people do need that free time.

Also saying founded in 2025 and then showing all other brands as NOT founded in 2025 is odd, isn’t it? Is it a bad thing they were founded before and already have market experience? 

Tricks

I have mixed feelings about this website. On one hand, it goes bold and tries something different. 

On the other, it manipulates the user in slightly too many places. We can say that all landing pages do some sort of harmless deception, and that’s also true. 

But I’ll let you judge for yourself if this is still fine, or is it crossing the line.

What do you think?