Why I am not sold on VOIP PBX

A customer, a local landscaper, called me up a couple of weeks ago looking for advice on a pending telephone system acquisition.  He had a limited budget to work with and needed less than 9 extensions (which included 2 fax machines.)  While he’d sought a number of bids, only one had come back.  He thought the vendor’s proposal was over priced and that the sales agent had not paid attention to his needs as a customer.

In all fairness to the proposed vendor, he selected Iwatsu which is a 70 year old vendor with many firsts with regard to telecommunications.  Their recent claim to fame is their VOIP capabilities for business PBX.  The issue is that it is a lot of phone system (in terms of cost) to address the needs of a small landscaping business.

In the past, I have bought, deployed, installed and maintained for the other major vendors: Panasonic, Nortel and Avaya.  A while back I was ‘certified’ to work on the Nortel systems.  And personally, I happen to think that Avaya makes the best quality VOIP system.

But I am not sold on VOIP PBX. 

There is “something” about the quality of VOIP when it comes to PBX replacement which I don’t believe any company had adequately addressed; nor with existing technical limitations do I believe that it will be addressed soon. 

Let me clarify — I am sold on VOIP for basic telephony services such as Time Warner Cable Business Class and Vonage.  These local/long distance services use simple application specific devices over the traditional Internet connectivity and work excellent.  Their cost is low — so low that as a shareholder of both AT&T (T) and Verizon (VZ) the current dominant market leaders I am almost embarrassed.

What gets me riled up is when I talk with the lawyer, accountant and business partners on their VOIP PBX, connected to their outbound VOIP carrier, connected to my inbound VOIP carrier, and finally to me.  “Something” happens…we all talk just a little slower and sometimes entire syllables of words go missing.  This isn’t acceptable quality to me on a cellular phone and it certainly is an unacceptable quality in a ‘landline’ conversation.  My bottom-line is that I haven’t seen a VOIP PBX work well with either a VOIP carrier or another VOIP PBX. 

Ultimately, the customer needed to make a PBX decision and it came down to price.  The Iwatsu system came in at over $5,500 for parts & labor.  We were able to find a Panasonic 824 Hybrid small office PBX with all of the components for $2,200 dollars plus my estimated labor of 8 hours (at the industry published rate of $45 for the 1st hour and $60 for the 2nd hour — and of course no charge for the recommendation) came to a grand total or $2,665. 

The Panasonic was the best choice because of the availability of CO ports and extension ports.  The base system was configured 3 CO x 8 EXT and the 2 x 8 expansion card costs $273 (compare this to $173 for just 2 CO’s on the Iwatsu card EXT ports cost extra.)  This PBX has a built in voicemail option, but we wanted auto attendant; so we went with the Panasonic TVA50 that supports custom menus and voicemail.  The TVA50 $498 integrates automatically with the PBX using APT integration — it took all of 10 minutes to install and about that long to program the auto attendant.  Another nice feature about the Panasonic system is that the proprietary display phones are cheap at $143 — about half price compared to the Iwatsu.

Panasonic now bundles the software for PBX and voicemail integration with their systems.  And the systems can be administered via a locally connect USB cable or over a LAN — all very cool features for a small business switch.

By the way, included in the $2,665 price tag are a couple of extra features.  We saved so much money by going ‘traditional’ that my customer was able to add a couple of features he hadn’t planned on – music on hold with a custom voiceover $168 and a loudspeaker for his shop $125.

So if you are considering a VOIP PBX, I hope this posting helps you to consider a traditional PBX.  And as always, give us a holler — the first 15 minutes are free.


The Value of Monte Carlo Simulations in Financial Analysis

Back in my days at The George Washington University in Washington, DC I had a number of professors who espoused the values of statistical analysis.  In those days which were at the cusp of transitioning big iron computational capabilities to the desktop PC it meant spending a ton of time programming data into the mainframe computer in the computer department’s lab.  It also meant that you had a limited amount of available time to process your data, so you couldn’t examine the data or tweak the numbers to wring out the fine details.

In these early statistical analysis days, if you wanted to run basic statistics on your PC it could be done with Lotus.  This was a basic spreadsheet where you’d punch in a data table, then generate sufficient random numbers (by recalculating the random numbers on the spreadsheet using a macro) to examine potential outcomes of business decisions influenced by expected market outcomes.  After the data was generated, you’d enter the lists into formulas to compute basic statistical information such as average, standard deviation, min, max, etc.

A few years later SAS (the mainframe based statistics program) ported over to PC’s; and several commercial competitor for statistics software hopped into the market.  Also, our good buddy Lotus now faced a formidable competitor — Microsoft and their spreadsheet package, Excel.  (As a sidenote, am I the only one who recognized that the Excel logo is a cross across and “L”?)

As the PC’s continued to evolve and when I ditched my Mac for a real PC, I started using Decisioneering’s Crystal Ball (I think they are now an Oracle company.)  The software package worked great and integrated with Excel.  I could perform all of the statistical function I could imagine on the spreadsheet data.  But for me the problem remained that like SAS, Crystal ball came with a high price tag relative to the number of times I’d use it over the course of the year.  Also, I found that the ever evolving Window’s software often outpaced the ability of the Crystal Ball programmers…this resulted in all too frequent bugs stopping up the software.

At American Special Projects, we recently took on a pro bono job for a startup client in the petrochemical industry.  As you might imagine given the state of the US and world economy between finances and fuel, now more than ever there remains a need to be able to accurately build and model financial outcomes.  We built a complex spreadsheet for this client which considered a significant number of independent and dependent variables that drive the financial outcomes.  By changing some of the independent variables we could begin to identify basic min, mid and max numbers.  However, the model consists of 20 independent variables so the task of changing the variables to reflect different market circumstances was burdensome.

We performed a quick Google search and found a number of Monte Carlo simulation packages.  By the way, Microsoft has an excellent write up on their website (http://office.microsoft.com/en-us/excel/HA011118931033.aspx) that introduces how to use the RAND function in Excel.  And of course our old friends SAS and Decisioneering show up at the top of the searches and in the Adwords bars thanks to their big budgets for search engine optimization (SEO.)  However, the best tool turned out to be the free tool from two professors at Wabash.EDU.  The URL is  (http://www3.wabash.edu/econometrics/EconometricsBook/Basic%20Tools/ExcelAddIns/MCSim.htm) which is a website written to supplement and support “Introductory Econometrics: Using Monte Carlo Simulation with Microsoft Excel” written by Humberto Barreto and Frank M. Howland.  The MCSim Add-In for Excel is free to use and is compatible with Excel 2007.  I haven’t read the book although, if it is as well documented as the website tools, it is probably excellent.

Using the Microsoft samples as a guide to building the independent variable tables, industry knowledge and the MCSim Add-In was very easy.  Within an hour we had a model up and running which generated data with a high degree of confidence. 

The MCSim Add-In has some limitations compared to the commercially available packages — namely it only computes basic statistical information and the built-in report generator only reports on two dependent variables.  On the plus side, if you are willing to compute your own stats, then the functionality includes the ability to track up to 256 variables and iterate the outcomes up to 65,000 times.  The built in statistical functions in Excel allow you to easily work with the data generated by MCSim.  I also noted that sometimes the MCSim had a problem handling names, so I tended to use the cell references instead.

Overall, I’d rate the MCSim Add-In as a 9 out of 10 for functionality, ease of use and most certainly because it’s free .  A big thank you to the creative minds and generosity of Dr. Humberto Barreto and Dr. Frank M. Howland.


This blog is not a first for the industry.  Although it is a first for American Special Projects (AmSpec to our friends) and a personal first for me.

I’ve been in technology a while now.  And while I am trained in finance, technology is my real passion. 

Keep checking back on this blog.  No doubt I’ll rant and rave about finance, technology, business and product ideas that should be used by business.