Hi,
over the years, I have found that the difference between hardphone and softphone has been reduced from flexibility to stability. With almost every vendor now moving from harware-based solution to a software-based, many a phone are actually an application running on a dedicated hardware, so from a feature point of view, they tend to be identical.
There was a tendency to separate Genesys CTI signalling and actual voice through the use of toolbar to control the phone operations and the nice shiny black hardphone to carry the voice. As CPUs got faster and Microsoft introducing RTC 1.2, many hardphone vendors started offering a software version of their phones albeit without the shiny black box, yet the overall user excitement so far has been somewhat muted. The reason is that your application is only as good as PC that it is running on, and given that many of PCs in a call center usually tend to be a low-end Dell and system administrators have a strange tendency to run daily virus scans during the odd hours of day and night, system sometimes can feel very sluggish. Add to it the fact that it is Windows (I am not a Linux fan myself) which manages to somehow become cluttered with additional DLLs which increase its boot time by a factor of five, factor in the possibility of system freezing for a few seconds whenever you open an excel file or a browser (I still have no clue why it happens even on the fastest of PCs we have) and that operators always find a way to install or change something in the system that would start throwing COM exceptions, and most reasonable people start feeling quizzy about having the whole thing running on their box.
Having said that, the last call center we have installed, which is running on SIP TServer, is purely software with no hardphones what-so-ever. With over 800 seats, it is the biggest one for us both in terms of using SIP T-Server as well as our SIP phone. So far, problems with it that we had were all related to PC performance, where there would be a slight degradation in voice quality or even total application crash due to a poorly schedule virus scan. The great thing about having a software-only version is that you know exactly what operator pressed. It is also much much much easier to implement new features, not to mention customizing rings, the way headset behaves or even how the inbound call is handled. There is also a major benefit to not having a clunky box around. The downside is that the headset is still there, so the problem associated with having a headsets and the ease of how easy they manage to break are still around.
On top of that, one of my biggest headaches so far was with Dell PCs. Dell PCs using SigmaTel audio boards and MaxAudio (or something like that) for some strange reason require customized mic and speaker adjustment per PC, meaning that you cannot preset the mic volume in the software and then just roll it out to hundreds of PCs: mic level has to be adjusted per PC. On top of that, there is a problem with impedence - yes, I know it is something that we all learned in school and has something to do with taking resistance and inverting it, but it is important that your headset impedence matches that of the board, otherwise, you will find your operator speech incomprehensible at best, no voice whatsoever at worst.
Hardphones are easier to rollout; however, in the long term, they usually tend to take too much space and offer too little functionality given that you are paying not only for the space that they are using, but the maintenance as well. You also do not have much say in how they look or what they do.
Software-only phones offer the flexibility and ease of use; however, require extremely careful planning in terms of both infrastructure (peripherals, PCs, network) as well as use.
I think softphone-only approach would be more than suitable for smaller call centers (those with less than 50 seats) where the calls were helpdesk-related and you can get away with having a call dropped due to a system malfunction (I would expect one in 10,000 to 50,000 range) but it would not be suitable YET for a large scale rollout unless you carefully plan every little detail of the system and its workflow.
My guess is that during the next five years we would see a lot more call centers using software-only phones; however, it will still be some time before software-only phone use would be widely accepted for a large call center installations.
I hope my rambling has helped
Vic
