LinkedIn launched its website in May 2003, after about six months after the company was founded, with only the most basic functionality. You could search users, connect with them if you knew their email, and, of course, user profiles with information about their network.

In 2016, 13 years later, LinkedIn was acquired by Microsoft for $26 billion. What started as a relatively simple website with just about 1,000 users now is the largest platform for professional networking and recruiting.

“If you’re not embarrassed by the first version of your product, you’ve probably launched late.” - Reid Hoffman

When LinkedIn launched, there was one feature missing that founder Reid Hoffman and his team were very concerned about. They wanted to build a “Contact Finder” so you could send a message to your contact list if you wanted to hire, say, a developer. Your contacts were supposed to see the message and forward it to their contacts and so on. But the website was launched without it, because they prioritized an early launch more than they did adding all the bells and whistles to the site’s feature set. Fourteen years later, when Microsoft acquired LinkedIn, the “Contact Finder” feature still wasn’t there.

Reid wrote about being embarrassed by your early products in an essay of his. Being OK with a product that you’re embarrassed about doesn’t mean that you have to launch something subpar, something that will piss off your users and result in a wave of lawsuits. But it’s OK if users see a product that isn’t perfect, or is even a little lame. That’s much better than perfecting your “baby” for months — months you could’ve spent on customer development, which we’ll discuss in a little bit — just because you think that you’ll be judged by your product. You won’t be judged. People are generally good; if you create something that helps them with their daily problems, they’ll appreciate it even if the product isn’t perfect at launch.

I think that it’s important for founders to think about launch as a process. It’s a marathon, not a sprint. When you run a marathon, you don’t care much about the first 100 feet because your goal is to run 26 miles. Sure, the first day of launch is important, and the time leading up to it is quite stressful. But you should focus on the first week and then the first month. Because after you launch, you’ll evaluate the results, iterate and do a soft launch again.

Imply a Glass Ceiling by Focusing on a Small, Limited Market

What if you want to launch a product that doesn’t scale very much? With an Uber or a LinkedIn — a product with a huge audience — a first launch that fails doesn’t feel like a total tragedy. You can launch again and again, often reaching different sets of people each time. A first launch that fails feels a lot more devastating when your whole audience is just a few thousand people, which is very common for B2B products.

The reality is that something is always going to go wrong in a first launch, turning some people away. The best that you can do is launch early with a minimal version of your feature set, focusing on a small, selected set of people when you do.

Even founders of products that scale make this mistake. They don’t focus on a small market; instead, they opt to build a landing page and begin pouring traffic to it immediately. Sometimes this approach works, but without a personal connection with a few early users, you get very limited and subjective feedback — all you can do is use app analytics to see how people use your product.

It’s a false assumption to assume that it’s OK to go this route. In the beginning, it’s important to do things that don’t scale, no matter the size of your audience. It’s important to focus on a small, limited market first. Because when you focus on a small market, you inevitably have to do customer development, which will provide you with useful feedback.

When I was in graphic design school, I had a great illustration tutor. He taught us that the only way to create an outstanding illustration is to imply constraints. For example, force yourself to use only black and white colors. Or create an illustration with only sharp edges. Or draw with your left foot. By implying constraints, you make your work stand out.

Same thing with marketing — even if you have no glass ceiling, you have to imply one. Otherwise, you’ll end up with a generic product.

Fail Fast

“Fail fast” is promoted so much in startup culture that it often scares founders. But it doesn’t mean what you think it does. No one is suggesting you actually have to fail with your startup. In software development, failing fast refers to a method of building a fault-tolerant system. Yes, failing fast is all about preventing failure — not the opposite.

The idea behind the fail fast approach is to build a system or a process that indicates and reports problems immediately, indicating failure and preventing bigger problems. For example, in a lot of electronics, fuses are used to prevent the whole system from burning down. If the fuse dies, your microwave will stop working, but you won’t have a fire in your house, and you can get your microwave fixed by replacing the fuse, allowing you to continue using it.

If you launch your product the right way, customer development is the fuse that allows you to fail fast. If you show your product to people in a controlled experiment, one by one, you can quickly see when something isn’t right and fix that problem before you show your product to another customer. This approach is slow and tedious, but after doing several runs of these demonstrations, you’ll be sure that the product is usable and can be shown to a wider audience.

Have you thought about narrowing down the audience of your product to have a glass ceiling? Will you begin the process of customer development, or are you doing it already?

Originally published on Medium