VentureBeat presents: AI Unleashed - An exclusive executive event for enterprise data leaders. Network and learn with industry peers. Learn More

Moshe Greenshpan is founder of Skakash, a face recognition based app.

So you’ve decided you are going to turn your great idea into a product, but you’re still facing a major problem — your budget.

Facing the “low budget problem” myself has led me to crowdsourcing, what seemed like the one and only possible solution.

Simply put, crowdsourcing is a way of using a group of people to perform a certain task, usually online. It encompasses the bustling virtual marketplaces and web platforms of excited, talented, and inexpensive contractors from all over the world.

From engineers and developers to designers, marketers, and data-enterers, millions of young contractors from Asia, Europe, and the US are at your disposal, and they’re constantly looking for their next online project.

After interviewing hundreds of contractors, hiring a couple of dozen and running my self-financed startup, I can definitely say that building your product through crowdsourcing at a fifth of the market cost is not an easy task, but it’s clearly possible.

I’ve been told many times that if I pay peanuts I’ll get monkeys, but I believe that there are more than a few expensive monkeys among us.

Crowdsourcing is a game of people, and choosing the right people for your team is crucial for your success. The quest for ideal candidates can be long and tiring, but it will inevitably make you a crowdsourcer — a user of the collective intelligence of talented people, and that’s before you’ve even hired anyone.

When I started with the idea of Skakash, I neither knew the basic difference between face detection (face found) and face recognition (face identified), nor did I realize the obvious fact that women are trickier than men recognition-wise (makeup blurs the facial features).

And although talking with a bunch of professionals about your project doesn’t make you the field’s authority, it does make you a true expert of your own assignment.    

There are many ways to use crowdsourcing in your favor, but if you are an entrepreneur or a small business owner on a tight budget and your focus is on developing new and unique software product, using the following golden rules will help you build your product cheaper, faster, and better.

Look for little experience with high rating

Crowdsourcing platforms provide project experience history and rating for each contractor, and those two factors will determine your cost. Newcomers with no experience are very cheap, but they are usually a gamble; better to let someone else hire them first.

Contractors with a five-star rating and dozens of medium to big projects under their belt will usually come at around market price or even higher. They have a long list of “rich” clients begging for their help, and they can’t dedicate themselves to your project, not with your budget.

In order to get a talented contractor at a fraction of the market cost, you need to look for 5-star contractors who have completed no more than a few small projects. But make sure those projects are not much smaller than yours. That way, your young contractor will see your project as it is — a challenge.

I also recommend giving your contractor incentives such as shares in your company or a success bonus, just like any startup would do with its first employees.

Look them in the eyes

Communication is the key. Running a remote project offsite has its problems and limitations; however you will avoid most of them if you keep an open communication with your contractors via video calls over services such as Skype throughout your project.

You should insist on a video interview. It will tell you a lot about the candidate and will help you make the right decision.

If the candidate refuses the video interview, it usually means he has something to hide, so don’t waste your time and energy trying to convince her otherwise.

Also, don’t be tempted to overlook the language barrier. Your candidate’s technical expertise isn’t worth much if you have hard time understanding her, and she can’t discuss your project fluently.

Look for motivation, intelligence and experience in that order

You are looking for someone who is truly excited about your project and is willing to put real effort into it. Intelligence and experience will only get you so far, but motivation goes a long way. You can have the most experienced and smartest engineer on the planet, but if he is not committed to completing your project, you’ll be in big trouble.

I did put intelligence before experience because I’m assuming you have somebody to guide your contractors, which leads me to the next rule.

Get a local chief technology officer or vice president of research and development

Your CTO/VP R&D is the guy who’s responsible for managing your developers and delivering your product. You need a true partner on your side who shares your language and vision — someone you can meet with face to face frequently and take with you to meetings. And for those reasons, I advise you to get a local one.

Use open source and revenue share

Before you start reinventing the wheel, make sure it is not already out there on the web, for free. Even if you do come up with an idea for an original product, chances are you could find and use open source projects for some parts of it.

Researching for open source projects before you start your development can save you months of hard labor and a lot of cash.

If you need third-party software for your product, you can save a lot of money if you’ll be willing to give up some of your future revenue by getting into a revenue share program, provided the third party offers or is willing to accept such a program.

When you start using crowdsourcing, use only one or two big platforms for your project. I recommend using Odesk for hourly work and for fixed price projects.

VentureBeat's mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.