Matthew Friedman offers a fairly long but useful article on five things to consider when trying to get the maximum value out of a VoIP implementation for your enterprise. The whole articles worth a read, the main points are summarized here.

When it doesn't work, enterprise VoIP can be a disaster, Friedman notes correctly. He quotes George Goodall, who says that while Voice over IP can be very successful in the enterprise, the key is to step very carefully and make sure that everything is in order before trashing that old private branch exchange: "Most of the problems with VoIP implementation can be caught before you implement it.

First, make the business case. Do you really need VoIP? Put down on a sheet of paper exactly why VoIP would work better for you than your current PBX system.

Those benefits, and not the amorphous promises of technological novelty, are what make the business case, says Avaya director of communications applications Lawrence Byrd: "It's always easier to start with the business case of what you're trying to accomplish. Do you want the kind of features you have now with a PBX, but with VoIP, or do you want something different?"

Secondly, have reasonable expectations. One part of making the business case for VoIP is to have reasonable expectations for the new system. Success and failure, after all, depend to a great extent on how you define success.

"One company I know about wanted to use VoIP to do toll bypass using its home-office workers' DSL lines," Goodall says. "What they didn't factor in was the data demand spike at 3:30 pm every day when the kids came home and started using their PlayStations. So people switched over to their traditional telephones, charging the company for their long distance at full price."

Thirdy, test your network. All the benefits of VoIP will mean nothing, of course, if your data network collapses under the weight of the voice traffic. Just about any IP network can, in theory, run VoIP. However, almost none are properly configured for voice traffic right out of the box, Byrd says.

Fourthly, have a plan. One approach, which is hotly debated in the VoIP market, is to rip out your PBX, phones and wiring and replace everything with IP phones. That might work in same cases, particularly where the old PBX is long past its expiration date, but it isn't always the best approach.

"It's often a good idea to pilot or start small," Goodall says. "You should look at where you really need to implement VoIP, whether it's a PBX overlay or at a branch facility. As with any major project, it's often a good idea to start small and plan from there. If you have to pull the plug for any reason, you still have that option."

Finally, choose the right vendor(s) on the site Friedman gives instructions for how to do this. "The problem is that there is such a cornucopia of options." Goodall says. "So you need to know what you need a vendor for. You have to be sure that the VoIP system talks to your data infrastructure and that whatever vendor you go with can support your handsets."

Above all, you have to choose a vendor in terms of how its technology and features work for you. "The big question is 'will it change the way I interact with customers or the way I do business?'" Byrd says. "If it will, are these changes for the better?"

David Sims is contributing editor and CRM/contact center columnist for TMCnet.