Cloud Series 5:Hosted Voice – Just say no

by Dave Michels

Welcome to Part 5 of my cloud series. So far, most of the posts have been about cloud computing trends in general. This one dives (and jumps back out) into hosted voice.

Basically, if you don’t want to read this whole post, let me just summarize it here with this: Cloud computing is going to to be big, resistance is futile. With that in mind, you should take a good strong look at hosted voice, the very epidomy of a cloud service (cloud2), and run like hell.

This post is much better if you read the entire Cloud Series:

Part 1 The Cloud
Part 2 Cloud Computing
Part 3 Why the Cloud Now
Part 4 Cloud PUNishment

I am becoming a real fan of both SaaS and cloud computing (yes, I believe there is a distinction, see parts 1 and part 2). I miss Outlook a bit, that rascally hour glass, the inability to find anything, and I just don’t know what to do with all this memory. The migration has been a real downer for all the remaining servers which are acting very busy lately.

The reality is they are safe as I can’t find the SaaS I need. In fact, all these articles that are pointing out cloud migration is slow, evidently have not noticed the limited selection of services. Both personally and professionally I can’t find the apps I need. For example, I would like to implement a SaaS personal finance solution (ala Quicken) but can’t find a suitable offering (recommendations?). Co-location is an option, but there is no incentive, seems like a distraction right now.

One area where there is plenty of choice is hosted voice. No, but thanks for offering.

Vonage is the best known hosted voice service, but mostly as an alternative of consumer land lines. Most consumers are happy with their cell phones only – but those that do want a family line should look at the various IP services. MagicJack, Vonage, Packet8, knock yourself out (seriously). But this post is more about PBX replacements or Virtual PBX options. These services vary, but have the following in common:

  • Typically charged per user per month
  • Typically SIP based, often Polycom phones.
  • Advanced web consoles for self service administration.
  • Unified messaging and some amount of calling minutes are generally included.
  • Usually there is no term commitment – though as with cell phones this disappears with subsidized hardware.
  • Each phone comes with a direct dial number, as well as “internal” extension dialing.

I was a believer in hosted voice. I got into reselling the service about 5 years ago. Back then, the service was a bit rougher around the edges. My assumption was the service was going to get better and cheaper and just might be the next ground floor opportunity. Well, I got one right.

The services have definitely improved. The web consoles are more intuitive and the features are more accessible. But the prices haven’t really dropped – they continue to hover around the $30-$40/user/month. Exactly what is included in the monthly price varies slightly more than the price itself. Just about every brand of VoIP switch has been tried in hosted offerings, but two solutions stand out. The larger systems tend to be based on Broadsoft’s platforms (or a platform that Broadsoft acquired). Smaller carriers, Interconnects, and especially ISPs often implement it on Asterisk platforms. There are hundreds of players in this game despite continuous consolidation.

Hosted voice fits the SaaS profile nicely. Modern phone systems are effectively all software – use of SIP for both trunking and telephones eliminates specialized hardware. The hosted providers take care of operations, software assurance, upgrades, backups and all those other pesky responsibilities associated with on-site phone systems. The flat monthly rates are positioned as a hassle free solution that easily expands or shrinks with business changes. There is no upfront costs and for users of traditional TDM switches, the feature upgrades are significant.

Despite my interest and excitement about SaaS, most organizations are better off owning or leasing a phone system. There are lots of ways to implement phone systems now – traditional implementations of course fit the majority, but implementation on virtual servers on or off premise could make sense. Or another option I’ve discussed before is implementing simplified solutions and using hosted services. I’ll discuss more about cloud options this in future Cloud Series posts, but suffice to say the cloud will have significant impacts to voice. I consider hosted voice another (limited) tool in the toolbox. I am not opposed to hosted voice, I just think it is often misunderstood (explaining its growth).

Here are my concerns with hosted voice: [As usual, I refer to a the “PBX” as the CPE call manager – VoIP assumed since 2005]

  • WAN bandwidth: Most hosted services are delivered over the public Internet which creates some unique quality challenges. Many providers require dedicated circuits and equipment to improve QoS. Bandwidth contention issues are significant with hosted voice. Remember, all voice traffic must pass through the WAN, including extension to extension/intercom conversations. A PBX does not necessarily change the criticality of the WAN, but reduces WAN traffic by keeping local traffic local. Also the ability to connect traditional lines into the phone system provides both route diversity and support for all those analog devices.
  • Dedicated bandwidth: All phone systems (with the exception of true Centrex) have less trunks than lines. Meaning not everyone can make a call at once. Ratios vary, but commonly twice as many users/phones as trunks is common. Most hosted providers manage this at an aggregate level, not a customer level. I rescued a small business that was fired from a hosted voice service. This customer sold tickets twice a year which wreaked havoc on the provider’s service. The tickets were so popular, that at one point they had 80 people in queue (they had 3 agents). This impacted all of the other customers on that provider’s network – which is why the customer got fired. We put them on a low cost PBX with some nice call center features and a single PRI. They were releived to know that the 24th caller would get a busy signal. The customer reused their SIP phones and obtained total control at a lower cost.
  • Features: Hosted voice features are competitive with most VoIP phone systems – but the PBX model is changing. Phone system features are just a component of voice now. Another critical component is integration with other systems; email, presence, collaboration, conferencing, FMC, Skype, and more. This actually hurts one of the biggest benefits of SaaS – free automatic upgrades. Sometimes upgrades are not wanted. Look at Office 2007 or Vista as recent examples – both created significant learning curves and broke existing features.
  • Cost: The cost of hosted voice is failry linear; that is the per user cost doesn’t vary much regardless of size. Phones systems are historically notorious for the opposite (I need one port, why do I need to buy 12?). But software changes that. Now, the core infrastructucture components are available in SMB packages. The upfront cost of a phone system will always be higher, but the overall/total per user costs; particularly over several years is usually quite a bit lower than hosted. If you don’t have the upfront capital, consider leasing.
  • Analog: Hosted solutions can’t do analog. The providers generally suggest ordering analog lines separately if they are needed. They are generally needed for modem related things – which are decreasing but not gone yet. Modem technology is used in fax machines, alarm systems, postage machines – and various other systems. Modems are generally not compatible with SIP trunks (T.38 faxing is slowly changing this). Beyond modems, analog circuits may be needed for a variety of systems; door phones, conference saucers, etc. Running analog lines separate from the phone system increases costs, particularly dedicated analog trunks.
  • Elasticity: This is a confusing one because SaaS normally wins this category. The issue is DID numbers, and the issue is not as significant as it once was. Organizations typically want blocks of DID numbers so a reasonable dialing plan can be created and maintained. With phone systems, blocks of DIDs are typically reserved for little to no additional costs from the carriers. In the case of hosted, most providers resell individual DIDs at a much higher price. Organizations that shrink or plan to grow may be spending too much on maintaining reserved numbers.
  • General features: Hosted systems offer an attractive set of features associated with each extension, but overall system features are sold a la carte. This could include an auto attendant (each level), spell by name directory, night greetings, call center feature sets, conferencing services, and many more. All charged per month. Many of features are included in CPE PBX systems for no additional charge.

The cost of the on-site voice PBX has dramatically dropped over the past few years. The features are getting better. The APIs are getting powerful and plentiful.

Pointy Stick Questions:

So does this mean that hosted voice never makes sense?
Not really. Hosted can make sense in several situations for many organizations. My Beef is that people often select hosted for the wrong reasons. It is also important to realize there are a lot of folks out there that don’t want fancy features – dialtone and a voice mail light will do. These people don’t care about UC, APIs, or any of advanced features. LudditeComm just might be the right partner for them.

When Does Hosted Make Sense?

The best example is numerous small offices. Hosted solutions can create a virtual PBX across locations fairly seamlessly. This can also be done with an onsite PBX at one location, but requires a little more work. In tragic irony, hosted voice is best for small offices, but it lacks “key system” features often desired by small offices.

Does this mean the boom in cloud computing does not apply to voice?

Absolutely not. I think the cloud is going to change things dramatically and so far, I like what I am seeing (except all these clouds are obscuring visibility). Multiple vendors are racing toward certifying their products for virtual servers. SIP trunks know no distance and can distribute systems easily and inexpensively. Most importantly, various systems can integrate with each other creating powerful capabilities – of course this requires version control and access to the APIs.

Bottom line? Hosted voice is great – it has little risk and solves a lot of problems. Of course, the same can be said about arsenic. The right answer depends on the question – but the quetions around voice change over time and a voice solution should last a minimum of five years. Voice CPE has never been so open or versatile or inexpensive. The best way to ensure the right answer to changing requirements is to focus on a solution that offers flexibility, options, APIs, and of course strong features. The best scenario of hosted voice is continuous increasing (for growth) payments for technology you will never own or control.