IT departments have gotten more conservative about using new technology from startups, says Marc Andreessen, the founder of three startups himself. Andreessen advises emerging tech companies to steer clear of corporate IT. Let's hope they don't listen.Andreessen has been there/done that, having founded Netscape, Opsware, and most recently social networking site Ning.com. He's sharing his experience on the dos and don'ts in a Guide To Startups series on his Pmarca blog.

Andreessen agreed to answer a few questions via e-mail for my story, "Evaluating Tech Startups: The Risks And Rewards." Here's the interview:

Question: What are some of the common concerns you hear from IT departments when it comes to considering a new product or service from a startup? And how do you respond?

Andreessen: Well, to start, just to be clear, my current startup does not sell to IT departments, so I don't respond :-).

In general, there are two kinds of IT departments -- call them "early adopters" and "stuck in the mud". "Early adopters" want to harness technology as a competitive edge and look for reasons to act. "Stuck in the mud" departments actively look for excuses to not act.

Since the great crash of 2000-2003, the percent of "early adopters" has dropped dramatically, shifting much of the innovation in the industry towards the consumer side. Corporate IT as a whole is going to get what it deserves as a result, which is far less innovation in enterprise products and services.

The concerns all flow from which kind of IT department you're dealing with.

Question: Your Pmarca blog has a Guide To Startups series. What advice would you give tech startups to give them a better chance of success with CIOs and IT professionals?

Andreessen: Don't build products that are dependent on CIO's and IT professionals to succeed. Top-down selling in IT has become prohibitively difficult for most startups. You're much better off starting a company aimed at a market that WANTS to adopt new technology, such as consumers, or a self-service model for workgroups or small businesses.

Question: Was Opsware an easy sell to businesses? Any anecdotes to share?

Andreessen: See above comments :-).

Question: What about Ning, your social networking site? Compared with Opsware, Ning's cheap and easy to deploy. That would seem to lower the barrier to entry for business uptake. Does it?

Andreessen: It doesn't matter :-). Ning is either a free service (if we run ads) or a very inexpensive service (if we don't run ads) that is entirely self-service. If businesses want to use it, that's great. If not, there are 1.3+ billion consumers on the Internet who will :-).

Question: In general, how would you describe business awareness/receptiveness to tech startups these days? Is the Welcome mat out? Or are deadbolts and chains more likely?

Andreessen: Deadbolts, chains, shackles, boiling oil, pits full of sharpened steel spikes, iron maidens...

It's grim. So is the falloff in innovation in those markets as a result. Luckily, the new consumer markets are on fire.

There are exceptions, of course. Opsware did get to north of a $100 million annual run rate in four years selling entirely to enterprises. VMWare has obviously been extremely successful, although greatly aided by the fact that their product could be easily adopted by individuals well in advance of any corporate-level sale. And of course Salesforce.com is doing very well and increasingly among large companies -- although again, they got to critical mass based on self-service to smaller businesses and workgroups. But I think these exceptions prove the rule, which is that it's become very hard going in the enterprise market.

--End of interview.--

As my story shows, the evidence supports Marc's assessment that most IT departments aren't exactly in an experimental mode. But I have to disagree with him on the matter of whether startups should target business IT or avoid it. Businesses need innovative new products and services, we know that. And some CIOs are willing to go out on a limb when new ways of doing things (software-as-a-service, business intelligence appliances, commercial open source) are compelling enough.

The alternative is that business technology innovation dries up -- all the good ideas happen outside the firewall and on the Web. In which case, IT infrastructures become a wasteland of legacy software and utility servers, all in support of old ways of doing things. Don't let it happen, IT professionals. Open the IT gates to newcomers with something to prove.