I made a big mistake. I was convinced that the website I was going to launch was going to go viral, that I never talked to anyone about it while I was building it.

It was going to be so big, that upon launch, tech publications are going to want to cover every aspect of it. My Facebook friends are going to share my website like crazy. I'm going to be a trending topic on Twitter and LinkedIn. The launch couldn't come fast enough. All of this time behind the scenes building the perfect product is going to pay off!

You can guess what happened. Nothing. Well, a few users signed up, but they never returned. I was "stealth mode" while building the product. I was so focused on the engineering of the product, I forgot about the people that matter the most: my customers.

I made the cardinal mistake. I didn't start marketing the company until after I launched it.

If I did start marketing before, I would have realized that what I was building wasn't actually needed. And it would have forced to me realize that I actually didn't care about what I was building. I was obsessed with the idea, not my customers.

So, here's my advice to you: The best time to start marketing is the moment you agree to pursue your idea. Build up momentum to the real launch. And by the time you launch, you'll have a much better idea if what you're building is actually going to work or not.

Here are three tips on marketing your startup.

1. Marketing doesn't mean telling the world on social media you have a new product.

All marketing means is telling the right people that you have something that they might need or want to invest in. Anyone else, in an early stage company, doesn't matter. Focus on the people that matter.

Marketing could be sending physical packages to potential customers who could use your product. It could mean going to the right meetups where your customers hang out or hanging out on Facebook groups where you're actively building a product that can solve their problems.

Basically, marketing doesn't have to be a full-on marketing blitz. You're a startup, not Google.

2. The big, huge launches are almost always not necessary.

I remember specifically an entrepreneur who was so adamant about throwing a massive launch party. He planned months and months for it. He brought an amazing celebrity speaker in a great, expensive location. The event was perfect.

The company shut down six months later.

Big launches just don't work like the way the used to or what they intend to. Instead, save that money, energy and time to plan for milestone one year after launch. Tell everyone about your success and growth. Then a big party works out great because you're celebrating with substance. And this time, you'll have actual customers and investors to invite!

3. The best marketing is a product or service that actually solves a problem.

As someone who has seen the guts of a lot of startups and through several conversations with successful entrepreneurs, success always originated from a product that has been validated by the market in terms of acquired users, paying customers, etc.

Great marketing and sales for a bad product, almost always ends badly. And honestly, it's hard to have great marketing and sales playbook for a product people just don't want. The lack of traction will deflate everyone from the engineers to the product managers all the way to the CEO and founder.

If I had $100, I would spend $100 on building a product that people want. That means spending it on networking events with your target audience, surveys, and any way possible to talk to your target customer.

In the end, marketing is a mix of art and science. There is no right way to market a product. The only thing you can do is to start now and figure out what works best for your situation.