Eyeballs Is All You Need

I used to think that if you build something useful and put it on the internet, people would come and use it. However it seems that there is a lot more happening between building the thing and someone actually finding it.

I had thought about distribution a lot while building SuperFlights in Mumbai. This time I wanted to try something small myself. Build a product, put it on Twitter, run some ads and launch it on Product Hunt. See what happens.

So I built Unleaf. An online LaTeX editor. No login, no workspace, no clutter. Just write, compile and get your PDF.

I built it in around 30 mins using GPT-6 Astra. This was just for experimentation and finding out how marketing and other stuff works. I was curious. Building had become quick enough that I could spend more time thinking about how to bring the thing to people.

The product

Overleaf felt like enterprise software to me. I wanted to write something and get a PDF. Why did I need to go through all the other stuff first?

An editor on the left, a preview on the right and an export button. Very simple.

This is the actual editor at unleaf.lol. I was writing about whether tiny transformers learn algorithms or representations. The product was useful to me, so I thought it might be useful to other people too.

Buying some eyeballs

On September 27 I put it on Twitter. I wrote that Overleaf made LaTeX feel like enterprise software, so I made Unleaf. No login. No workspace. No clutter.

I posted two tweets and put both of them on ads. I spent $2 per tweet, $4 in total, and got around 40K views across the two tweets. That was it. Four dollars to see what would happen.

The screenshot above is of one of those two tweets. It had around 21K views, 52 likes, 6 replies and 2 reposts when I took it on October 13. Across both tweets the views came to around 40K. For $4 in total I thought that was interesting. People had seen the thing. Some of them even went to the site.

The analytics for the last 30 days showed 238 visitors, 340 page views and an 84% bounce rate. There was a spike around the launch and then it fell off. You can see it here.

Twitter was the biggest named source. t.co brought 93 visitors. Google brought 8. LinkedIn brought 6 from the website and 4 from its Android app. Product Hunt brought 4.

So Twitter worked in the sense that it brought people to the site. The screenshots combine the launch and the traffic around it; they do not separate the paid visitors from the organic ones. I also did not track how many of those people compiled a document or exported a PDF. These are visitors, not a count of people who kept using Unleaf.

But I kept thinking about the $4. What if I had spent x10 or x100 more? $40 or $400. What happens when you keep pumping money into the thing, like bigger companies do with much larger budgets?

I felt that with that kind of push I could have had a hit product. More eyeballs would inevitably bring some more people to my site. That was my thinking, though I did not run the bigger experiment, and I cannot assume that spending ten times more would give exactly ten times the traffic or make people stay.

Still, it made me think about how different the game looks when you can afford to keep buying attention. I stopped at $4. Someone else can keep painting the town red until everyone has heard the name.

Product Hunt

I also put Unleaf on Product Hunt. The result was 9 points, 1 comment, #58 for the day and #317 for the week.

Four visitors from Product Hunt according to the analytics. No magical #1 day or week for me.

I thought putting the product there might bring people on its own. It did bring a few. But Twitter was doing much more of the visible distribution in this experiment. Putting something on a launch platform is easy. Getting people to care that it is there seems to be the actual work.

Going viral on command

I see people going viral nowadays and sometimes it looks purely on command. It looks all prepurposed. A tweet is made, medium size follower people respond in comments and repost, ads are run to pad the stats, and suddenly the product is everywhere.

Then the same product is on Product Hunt and somehow it tops the day and the week. I wonder how much of that audience was already waiting before the launch even happened.

If a company has ten employees, especially active ones, it already has people who can comment, share and bring their own circles into the post. The same Twitter community can show up on Product Hunt too. It is a lot easier to create momentum when you have people around you from the beginning.

I had written in my notes that maybe someone spends $200 and gets around 200K views. That was a rough thought, not something I measured here. But the thought behind it was simple. What looks like sudden organic attention might have money and planning behind it.

I also have a theory that some people are paid to post things in these tech circles and massive Discord servers. People tweet, repost and quote in multitudes and the thing starts looking mainstream. Maybe it is a side income for some of them. I do not have proof of that happening in any particular launch, or of people using multiple accounts to vote. It is my theory from watching how these things unfold.

I don't think that every successful launch is fake. I think we see the tweet and the numbers, but we don't see all the work that happened before it. The communities, the relationships, the ads, the people ready to share. We call the final thing viral.

And once a thing is everywhere, more people look at it because everyone else is looking at it. The product starts getting a story around itself. An aesthetic, a vibe. People feel that it is the thing to try.

I think this is why I keep coming back to distribution. You can build something useful in 30 mins now. But if nobody sees it, it is still just sitting somewhere on the internet. Gold in a desert.

This was a small experiment. Two tweets, $2 of ads on each, around 40K views, a Product Hunt launch and a product I wanted to use myself. I wanted to see what would happen instead of just thinking about it. These are the results.

You can try it at unleaf.lol.