The release of InformationWeek's Startup 50 raises the old question: Should IT departments do business with startups at all? The unequivocal answer is yes, and here's my advice on how to pick the right ones.The issue, of course, is that doing business with startups entails risk, even more so than working with established vendors. Startups have a shorter track record, fewer customers, and they're generally not as financially stable. They may be acquired, have a change of product strategy, or go out of business.

The flip side of the argument, which argues in favor of working with emerging companies, is that they're more innovative and responsive and their technologies tend to be cheaper or otherwise help lower costs. How do we know this? In an InformationWeek Analytics survey last year, IT pros told us so.

With this week's publication of the InformationWeek Startup 50, we're re-releasing a report I wrote about six months ago titled "Vetting Startups." It's a guide to the world of up-and-coming tech companies-what to look for in evaluating them and what kinds of questions to ask.

Here are seven steps to evaluating tech startups, excerpted from the report:

1. Hands-on product testing. This is an obvious starting point and, appropriately, the most frequently mentioned screening method in our survey, with 88% of respondents indicating that they review or test a startup's products or services.

2. Customer references. There are few better ways of learning about a company than talking to its customers; 72% of business technology pros do that as part of the background check.

3. Check its financial status. Startups can tell you if they've received venture funding, how much, and from whom. VC funding alone isn't a great litmus test--good startups can be self-funded. Fifty seven percent of IT pros say they investigate a startup's funding and financial well being.

4. Meet the company. One of the nice things about working with startups is that company founders, senior executives, and top engineers are more readily available than in larger companies. Fifty-six percent of respondents to our survey meet with the CEO, CTO, or other top management.

5. Inquire about exit strategy. IT managers tend to overlook how a company will advance from being a startup into the big leagues of the tech industry--only 30% check into it--but it's a question you should ask.

6. Check on management background. Contrary to the stereotype of two guys in a garage, many Silicon Valley entrepreneurs targeting the enterprise market have been through the lifecycle of a startup before, sometimes two or three times. It helps to know that an experienced hand is at the wheel of the new company.

7. Stage of development. Startups can range from stealth mode, where plans are under wraps, to six or seven years old. PricewaterhouseCoopers and the National Venture Capital Association divide startups into a few buckets: seed/early stage, expansion, and later stage. Some IT departments shy away from the newbies and toward those with a few years of experience under their belts.

The above steps mitigate the risks associated with tech startups, but don't eliminate them. The bottom line is that IT pros need to do their homework on startups, but the benefits are often worth it.

You can download the full "Vetting Startups" report at no cost, along with InformationWeek's Startup 50 list and profiles of many of the startups on our list, here (registration required).