Forecast number of calls for software installations

[anndallas]

I'm working with a software startup to plan and implement a contact center. We're starting from complete scratch and I have no history to fall back on as we will be releasing our first product. I need to know if there is a general formula/method for estimating number of calls based on number of software installations. This is a bundled software product that ships with another product so I also first have to figure out how many people who get the bundle will actually install the software (need a best guess estimate) and then of those installations, how many will call support. Any software tech support call center people out there who can help me with this? Without an estimate of number of calls it is very hard to estimate anything else!

[chubbccm]

I used to manage hardware & software tech support for a major PC manufacturer (5 years). We would receive approximately 20 calls for evey 1000 systems shipped. This was pretty solid data so when we budgeted for the coming year, we needed the sales forecasts to project anticipated call volume. Also, for every system that we received a call on, it took an average of 5 calls to resolve the problem (real areas of opportunity for FCR).

You might want to think about the following:

Who your install base is?
How are they using your product?
What OS will it use (Linux, NT, XP etc.)?
Is there tech support on site or are you the primary point of support?
Who is handling escalated (tier II) support?
What is the difficulty level of your product?
Do you have any beta site feedback/data?
Get feedback from your install team(s) as to how well your user base is using the software.

Hopefully this helps a little bit.

[JoeB]

This is a common question yet is very hard to answer because, as chubbccm indicates, there are many variables that influence not only the number and frequency of the calls, but also their duration, the rate of repeat calls, etc. , which is also the reason why numbers from other call centers may not be useful. Nevertheless, it is possible to structure the estimation process by using engineering/QA methods that are used traditionally in hardware failure analysis and establishing a baseline. Next, use however little data you have from testing, beta sites, etc. to identify the characteristics of the problems and how they will manifest themselves in actual use.

This is only an initial estimate not only because of the lack of history, but also because it is likely that your software will evolve very rapidly in its early phases, changing many of your assumptions about failures, usage patterns, etc. Therefore, your call center design should support rapid learning and dissemination of new support knowledge.

Contact me directly if you want to discuss a systematic way to estimate the call rate and other factors that influence your call center resource planning

Joe Barkai
DIAGNOSTIC STRATEGIES
www.DiagnosticStrategies.com
jbarkai@DiagnosticStrategies.com

[gkern]

I have some general calls/user numbers that I think could help you, but I would need to know more about your enviromemnt, i.e. consumer or corporate user, toll free or local, pay or free, etc.

Give me a call if you want to talk.

Greg Kern
CallTech
512-261-0409