Retaining good employees is an ongoing challenge in any start-up. How do you keep them at your company when they have their own entrepreneurial streak? Have perks like ping-pong and foosball tables? Offer lavish stock options? Try making it easy for them to walk out the door. No, really.

That's what Daniel Saks, CEO of cloud-based b2b app vendor AppDirect, is doing. The company has gone from three employees a few years ago to more than 90 today. Given the cost of recruiting talent, you'd think he would want to implement incentives that would keep employees around. Quite the opposite is true.

Reality Check

Instead of locking people in, he wants to find engineers who have great backgrounds, come from successful companies, and want to try to go off on their own, but don't have the business background to make it work.

"What we determined out of the people that fit [our culture] is that one of the goals they often have is to create their own business," Saks says.

His thought is a dose of realism. It is difficult to ensure that people last a long period of time. In high tech, keeping engineers around for even a couple of years can be difficult. So AppDirect is a couple of months into an experiment. The goal? To see if his company can gain more than it loses by deliberately hiring people who are likely to leave and then make them business partners when ways part.

Core b2b engineers--the very people who understand the needs of business and can develop and maintain apps for corporations--have "less appetite to take risk" than many in b2c, he says. As a result, there are fewer start-ups created by people who leave established corporations that sell business technology.

But AppDirect is a marketplace that needs new and innovative cloud apps if it is to grow and thrive. So Saks makes a deal. Engineers work for the company for two years, rather than taking off after one for some other opportunity. They devote their energies to corporate goals while also learning the leadership, marketing, sales, and legal aspects of running a business.

After the two years, the engineers can go off without owing anything to AppDirect. They own their ideas. But Saks knows that many will develop apps and sell them through his company, giving it more products and services to offer its customers.

Will It Work?

As part of the plan, Saks will run an in-house incubator for some of the developers, providing six months of office space. Finally, there is a demo day when AppDirect invites venture capital backers--its own as well as others--to see presentations by the new companies.

Although it's too early to tell how the program will work, it's an innovative strategy experiment that can hardly backfire. What's the worst that happens? Engineers leave after two years because they would have anyway and AppDirect then hires additional ones? Maybe some will even decide to stay for longer terms. In any case, this is an example of how an entrepreneur can try something new that is low risk with a strong potential upside.