Startups are at their sexiest when hundreds of millions of people around the world use something that a couple of guys and girls built in their garage.

But I've noticed how that perception lays a trap for many first-time entrepreneurs. With their sights set on serving the masses, first-time founders often conclude that they must build a product that will work for millions of customers -- before they even have one.

This is such a major problem that Startup Genome identified "premature scaling" as the number one cause of startup failure. Surveying 3,200 startups in 2011, the startup-community hub Startup Genome found that a whopping 70 percent failed because they tried to scale too early -- expending resources on add-ons like expensive marketing and hiring salespeople before they truly had a product to satisfy a sufficiently large market.

Here are three ways to build a business organically:

1. Do things that don't scale.
Paul Graham of Y Combinator fame explicitly implores his companies to do things that don't scale, like finding ways to pay attention to and delight individual customers early on instead of dismissing such moves as unscalable or trifling.

In Graham's experience, first-time founders mistakenly believe that startups either take off or don't, and that's why they often impetuously prepare for startup glory. By relying on the patterns and models of well-established startups, founders miss a pivotal stage in their company's own life and therefore fumble.

Ryan Smith, founder and CEO of Qualtrics, a research software maker in Provo, Utah, describes the correct order of operations: "Nail it, then scale it," he says. Founders and their companies must go through an "adolescence phase" to nail who they are, as starting a business requires a period of growing up, self-learning and tuning of their product, identity and approach. After all, you have to learn to walk before you can run, right?

Related: The Race to Scale: How Fast Should You Accelerate?

2. Make something people want.
Part of the adolescent stage involves slamming doors on ideas and features that distract from your product and path. When it comes to making something that people want, the hard part most often is figuring out the "what people want" part. You can build something, but is it the right something?

To discover what people will pay for, find the right market, study and experiment with your product and talk with customers. The great advantage of making what people want is having a product that markets itself.

Related: Bootstrapping Your Startup to Scalability and Profitability

3. Sell it before you build it.
I've seen first-hand how successful startup founders initially gained real, paying customers before writing a single line of code -- and then scaled up their technology and personnel second.

Zirtual, a virtual-assistant service, grew its business for a year and a half without any real technology. The new customer-signup form on its website just sent an email to founder Maren Kate Donovan to alert her that someone was trying to sign up. She had to add that customer to the new customer spreadsheet and email the new customer manually, and then scan through her pool of virtual assistants to find the best fit. Signing up a new customer took an amazing 45 minutes.

Donovan and her co-founders went through this pain to make Zirtual happen in spite of their technical limitations to be absolutely sure they were building something that people wanted before they scaled up and raised a $2 million seed round.

Related: Biggest Mistakes: How JackThreads Overcame Its Growing Pains

The San Francisco-based startup ZeroCater also spent time working off a spreadsheet in its early days. New customers came from in-person sales or cold calls over the phone and were managed in a spreadsheet, all by solo founder Arram Sabeti. ZeroCater grew to tens of thousands of dollars' worth of paying customers before Sabeti decided to scale the company bringing on a developer and join Graham's accelerator, Y Combinator.

No doubt, it's scary to sell before you build and to step outside of your comfort zone. But as Graham put it in a recent blog post: "It would be a little frightening to be solving users' problems in a way that wasn't yet automatic, but less frightening than the far more common case of having something automatic that doesn't yet solve anyone's problems."

How did you know when to scale your company? Let us know with a comment.