VoIP Blog - A moving experience. By Stewart Twynham
In this "fly on the wall" blog for AccountingWEB, Stewart Twynham documents the selection, implementation and real-life benefits of a replacement business telephone system based on voice over internet protocol (VoIP) technology. In traditional blog fashion, the first instalment is at the bottom of this page. If you haven't encountered the blog before, scroll down and read the entries in reverse order - or choose a topic from the contents box below and read upwards from that entry.
Setting the scene
Discovering the client's needs
Some myths about VoIP
A word about open source software
Don't forget the extras
Preparations - time to get paranoid
BT strikes back!
The users are revolting
What have we learned so far?
Hey, this is Unix!
A moving experience
13 June - One of the benefits discussed with the client before switching to VoIP was the ability to cope much more easily with office layout changes. This week, they decided to put that to the test.
Like many SMEs with space constraints, new staff or changes of roles can necessitate an office shuffle to some degree or another.
The previous phone system was wired using fairly standard structured cabling (CAT 5e for those that are interested). This meant that each desk required two cables - one for the PC, and one for the phone. With around 20 desks in one of the offices alone, that’s a minimum of 40 cables to be trodden on, tripped over, trapped under desks, or tied around uplighters.
In reality, it means a lot more wires than that. Every time a desk gets moved or a new PC is added, cables are rarely long enough, so you end up with plenty of BT-style extensions on phones and back-to-back adaptors for network leads.
Previous moves, therefore, were a bit of a nightmare. made more complicated by wall outlets that could be for either voice or data (phone or computer) - and you never knew what you were plugging into. Confusion means disruption, especially when PCs or phones don’t work - and disruption costs businesses money. The alternative is to bring staff in on a Saturday - neither option is ideal.
VoIP brings several benefits to any office that is constantly in a state of flux:
- Less wiring - each desk now only requires a single Ethernet (network) cable. For those PCs running soft phones - the phone IS the PC. And handsets from Cisco or most other suppliers come with two network ports. So you plug the phone into the wall, and the PC simply plugs into the back of the phone. This means fewer cables, fewer wall sockets, and less “patching” of cables in the server/comms room.
- Easy relocation. Move the PC and the soft phone moves with you. To move a handset, simple move the phone with the PC.
- Less troubleshooting. With the old phone system, each extension number was effectively hard-wired to a particular wall socket. This had to be re-patched, inevitably leding to problems when anyone inexperienced swapped over multiple extensions: Gary ends up with Jane’s phone, and Gary’s phone is now ringing somewhere in accounts. With VoIP, all of that’s handled by the network - as long as you can see the network, the phone (and PC) will work.
Of course, things weren’t all plain sailing. There are a couple gotchas worth mentioning:
- Some handsets can take their power from the network wiring using a technology called Power over Ethernet (PoE). This means you only have to plug them into the wall to work. However, not all handsets can do this (even the expensive Cisco 7912s for instance) - meaning you’ll need to fork out another £20-50 for an outboard power supply and will then need to find a suitable 13A socket to plug it into.
- Although we had gone to the expense of installing two 24-port network switches which supported PoE, the model we installed only provided power to the first 12 ports. They weren’t cheap units either (hundreds of pounds) - so it was a little disappointing to find out the hard way that not every outlet did what it said on the tin.
There are plenty of examples like that with VoIP - and I suppose that’s my biggest gripe so far about the technology vs more traditional phone systems - it can all seem a bit “mix and match”. Gone are the days when you would purchase everything from the same manufacturer and it would just work. Unless you choose a company like Cisco you’ll be buying the various parts from several different manufacturers, with the potential of mixed results - or the feeling that things just don’t quite bolt together so well.
If you’re the sort of person that builds their own PC, and knows their H.323 from their uLaw, then that compromise may be fine. But for the majority of businesses looking to adopt VoIP, then it’s a useful reminder of the importance of finding the right supplier - someone who can make sense of all the “gear” and recommend tried and tested combinations.
* * *
24 May - My admin training day has arrived, and it’s finally time to look at integration. If you recall, the client has an industry-specific application running on an Oracle database. Of the more traditional phone systems we looked at, you are restricted to buying their proprietary (read: expensive) equipment, or are limited to interfacing via age-old industry-standard, but not necessarily industry-supported protocols such as TAPI.
One supplier did offer something more flexible - software which allowed integration with other systems - but that came at a price (a cool £14,000 just for the license!)
It’s a little bit like buying a Ford Mondeo, then being told that to upgrade the CD player will cost £5,000 - or that you can buy a lead to connect your own player for just £14,000, but then it still won’t work with an iPod. In today’s world, that’s madness!
My client’s phone system, however, is Asterisk - which is an open source system. You have (in theory) the freedom to bolt on anything you like. But you do need to know what you are doing – and that’s the reason for my training.
The first thing I’m reminded of is that Asterisk runs on Unix. I haven’t touched a Unix system properly for the best part of 10 years, and probably not in anger for the least 15, but somehow sat at a blank terminal emulator screen I’m suddenly home. I suppose using a computer is a bit like riding a bike - you never forget. I wonder how many accountants still feel at home using the Lotus 1-2-3 “slash” menu in Excel?
CTI – the acid test
Computer Telephony Integration, or CTI , means you connecting your phone system to, say, your CRM system for all kinds of benefits. Full-blown CTI can be very sophisticated, but for my first application, I chose something simple, in this case taking the calling telephone number of every incoming call, and replacing it with the name of the person on the other end if they are already known to the CRM system.
There are actually three parts to the problem:
1. Build a web page or web service that runs on the CRM system and accepts a telephone number, attempting to match this against the phone numbers stored on the database, returning the corresponding name if found. This is probably around 10 lines of code, and I chose the C#.NET programming language for this particular example.
2. Write a script (I used a programming language called Python) on the telephone system to pass a number to that web page, and then read back the result. Around five lines of code.
3. Modify one of the configuration files for the telephone system to run this lookup on each incoming call. A single line of code.
I didn’t time myself, but I reckon from start to finish this took less than an hour. Yes, I’ve been writing software for a while, and I’m quite at home with Unix and the good-old “vi” editor. However the point is that this wouldn’t have even been possible with any of the conventional/proprietary phone systems we saw. I didn’t need to install any special software, I didn’t have to touch or reconfigure any of the phones.
It is worth noting that I have experienced genuine surprise from vendors when explaining that my client had chosen the open source route. I don’t know why that should be. After all, I’m not the one that’s charging £14,000 for an hour’s work!
* * *
16 May - Anyone thinking about VoIP technology may look at this blog and shudder in horror. Actually, things haven’t been that bad, but the niggles we have experienced are a useful reminder that implementing any major change in either IT or telecoms infrastructure can cause “issues”.
So let’s recap on some of the highlights:
- The single biggest delay to the project was caused by BT having inaccurate records relating to the phone lines, the business and even the address, along with a limited product knowledge.
- Several BT sales staff did not appear to understand the older phone technology (DASS) that the client was using, and more specifically the process to upgrade it. I was quoted several different prices for conversion (from free to a whopping £79 per channel), told several different stories about how the conversion would take place, and given several different lead times (from a minimum of six weeks to “unfortunately, you can’t have it tomorrow, will the day after do?”). To make things worse, they then lost the order, twice.
- Once the order reached BT’s provisioning team it was like a parallel universe. Staff were friendly, helpful, and knowledgeable. They called back unprompted, and this level of customer care extended right down to the guys that came on site to do the work. An excellent job.
- One of the biggest challenges on the changeover day was simply unravelling the old phone system! Thirty-odd handsets, in many cases with multiple extension cords, all needed to be untangled from desks. Many of the existing network cables (which would now carry phone as well as computer data) were then found to be damaged through careless desk placement and needed replacing. This all took several laborious hours, but did give the opportunity to tidy up the office space.
- Training, in hindsight, needed much more thought and planning. The logistics of one person troubleshooting and providing ad-hoc training for a team numbering almost 30 in one morning just does not compute. As well as the phone system, there were several different types of handset and headset to contend with.
- The PC-based soft-phones seem to introduce their own unique set of problems. People used to handsets can find it takes several days before they are used to operating their phone on-screen as efficiently as they would a handset – for example when transferring, parking and picking up calls. The occasional Skype call will not prepare you!
With the dust finally settling, it’s now time to put the system through its paces. I’m off for a one-day training course to start integrating the phone system with the client’s other computers. We know Asterisk makes a great phone system – now let’s see if it lives up to the rest of the hype!
* * *
2 May - The users are revolting! Okay, it’s not quite that bad, but now the system has bedded in, we’ve had a few gremlins emerge. This week’s problem: call quality.
Calls from the Cisco phones are crystal clear – no apparent delays, echo or loss of quality other than the normal compression that all phone calls suffer.
For the PC-based users, things are not so clear cut – and it can vary from one workstation to the next, or one user to the next. Sometimes clients complain that the phone sounds “cheap”. Some users are very aware of degraded sound quality in their headsets, others can’t understand what all the fuss is about. It would appear to affect around 10-15% of users in some way, shape or form.
In some cases, swapping one headset for another cures the problem – in other cases it doesn’t. Two of the sales team have grown unhappy with the quality of their wireless (DECT) headsets whilst two of the accounts team on the other hand have grown unhappy with their USB wired headsets. Swapping the headsets around means everybody’s happy. Go figure!
Part of the problem is psychological – now one person is aware of a problem, at the very least others are now listening hard to try and spot the same problem (which may have always been there but gone unnoticed). This explains why many of the sound quality issues have only suddenly come to light.
In almost every case, the issue is with a PC-based soft phone. PC phones in general do seem more susceptible to quality issues – I guess due to a number of factors:
- PCs cannot take full advantage of quality of service (Q0S) technologies which ensure that voice information takes precedence over normal data across a network.
- PCs tend to do other things as well – showing spreadsheets or opening word documents - not just acting as a phone.
- The “top of the range” wired headsets offer far better sound quality than normal monaural headsets – meaning that any sound quality problems are rendered in high fidelity./li>
- The DECT headsets, although expensive, do not appear to offer the same sound quality (especially earpiece volume) when compared to a standard DECT phone (albeit about 20x larger in size!)
Add in user perception and the user’s hearing ability, and it’s suddenly easy to see how a particular mix of circumstances might suit one user but confound another.
For now, we’ve got probably 70% of the sound quality problems ironed out – either replacing or swapping headsets, and in one case resorting to giving a user a Cisco phone in replacement for their soft phone. The remaining issues are still under investigation and I’ll report back on these issues next time.
* * *
25 April - When my client called me to tell me that their phone system had suddenly stopped working, I naturally suspected the worst. Their phone lines are brought in over a mix of cables that were laid in the 1970s, never designed to last more than 15 years, and certainly never designed to handle anything more sophisticated than a “trimphone”. A BT engineer once explained to me that the particular alloy used in the cables tends to oxidise into a non-conductive powder over time and that my client was situated at the “dodgy end of the High Street”.
Still, for once it wasn’t our cables at fault – in fact for once it wasn’t even BT to blame, but the gas board who had managed to cut through several major trunk cables close to the exchange whilst looking for a gas leak. Presumably they were lucky that they didn’t trigger a full scale gas explosion.
With most of the town out and my client’s SDSL circuit still working perfectly. It was time to test the VoIP system out over the internet. Or at least it would have been had we been a little more prepared.
In our case, the company firewall prevented the phone system talking to the outside world – which is exactly what a firewall should do.
Unfortunately, the IT company had no experience of working with a securely firewalled network – their advice being “open up all the ports to the phone system and it should be fine”. Yes, it would work, but then you’ve also now opened your network to every hacker in the land as well. Not something I can recommend.
A quick Google around revealed that VoIP protocols are not exactly “firewall friendly” – meaning that you can spend a great deal of time getting everything working right – and it does go some way to explaining the “just open up everything” approach.
Firewall problems aside, let’s look at what we could have done. Statistically, phone line failures are rare but they do happen. There may also be other times you don’t have access to a conventional phone line – you may have just relocated but the new lines might be delayed, or a fire/flood/theft may have required you to dust off your business continuity plan and relocate elsewhere.
VoIP means you can move your phone to almost anywhere – you just need a good internet connection and some form of gateway (which connects your VoIP telephone system to the public telephone network) and you’re away.
There are plenty of VoIP gateways out there – we use sipgate, but there are literally thousands of others. All you do is set up an account, configure your phone system, and hey presto – an instant phone service, no BT line required. It makes a great addition to your business continuity plan, whether you relocate or simply have an extended service outage as my client did.
This time, we weren’t prepared – the firewall settings will still need time to configure and test. But we will be prepared next time…
* * *
19 April - Just over a week had gone by since the upgrade, and it was time to visit the client again. This was a good opportunity for me to see how people were using the system “in anger”.
Opinion was mixed – some genuinely great comments, but a few genuinely poor ones too. The system itself was behaving well, with none of the technical glitches that seem to dog even the most simple IT upgrade. Even the phone systems I’ve previously been involved with have had more gremlins. Incoming calls got answered and outgoing calls were made, but that didn’t stop a few people being cut off in the process. “Three days of hell, but now it’s much better” as their accountant explained.
The most consistent complaint related to the user training (or the brevity thereof). Sitting with the team, I heard several clients being told “I’ll just transfer you, oh hang on a minute, we’ve just had a new phone system installed and I can’t remember how to do that.”
Of course, this isn’t just a feature of VoIP. I’ve seen similar instances with many a traditional phone system in offices which, let's face it, can all be very different to use. All of this highlights the need for training – not just when you have a new phone system, but any new computer system, new software, or indeed anything that could materially impact upon a key business process.
Another interesting point was that the soft-phones (the ones that work on the screen in front of you rather than a physical telephone) seem to instil additional confusion – purely related to their virtual nature. They look like normal phones, they behave like normal phones, and have many of the same controls. However many users struggled to make the connection; for them it was almost like learning how to use a phone again from scratch. Because most people have used soft-phones before - for example with Skype – it is assumed that moving to a soft-phone in the business environment won’t be an issue.
Most Skype users, however, aren’t fielding multiple incoming and outgoing calls, transferring calls, picking up calls from other stations, diverting to voicemail and so on – all the things that you’d typically do in a normal business environment.
Airline pilots have similar “same but different” problems when changing from one type of aircraft to another. Same information, same controls – just presented differently – meaning they have to both unlearn and relearn certain habits. Pilots get many hours of type conversion training and testing to resolve these issues but even then will predictably revert to previous thinking if things get too stressful.
For the average soft-phone user, receiving perhaps 20 minutes of training, it’s easy to see how an important phone call which needs to be transferred, parked, held, picked-up, passed to voicemail or turned into a conference call can soon instil “flummox”. With a shakedown planned for next week, the agenda has therefore been changed to include an additional half day of extra user training to try and iron out some of these issues.
Editor's Note: There was a short delay over Easter before the latest installment of the VoIP blog arrived. We think Stewart may have needed some extra time to recover from the training traumas.
10 April - In my experience, despite all the testing, there is absolutely nothing you can do to fully test a system upgrade (computers, phones, whatever) until everyone arrives back on the Monday morning. This upgrade was surprisingly short of problems, but in hindsight the training could have been better planned.
Because of the hassles with BT, the upgrade took place over a shorter period of time than originally planned. This and other circumstances prevented any “parallel running” of the two phone systems – although in reality the need to rewire most of the desks meant that such an exercise could have been quite disruptive. No parallel running also meant no training prior to go-live. The users were not simply moving from one handset type to another - many had to cope with PC-based soft phones, wireless headsets, and the intricacies of a new phone system, so it wasn’t a straightforward upgrade.
Making and receiving calls was not an issue – and most had figured this out within a few seconds of sitting down at their desk. Business telephone systems have to do much more, however – so the team was broken down into small groups to run through the facilities that most used every day, including:
- Making and receiving calls
- Parking calls
- Transferring calls
- Call pickup
- Diverting calls
- Other features (handsets, headsets, software)
Training people “on the fly” like this might work in a smaller office, but it proved impractical for the 25+ users in this one. Even split into groups of five, five training courses were shoehorned into little more than an hour.
The individuals that were yet to be trained were struggling to handle calls intended for those that were being trained (this was a Monday morning – always a busy time). Those that had been trained were then trying things out for real for the first time – often making mistakes – meaning that they had to interrupt the next group to query something.
By the time the last group had completed their training, their instruction had taken over twice as long as the first group as it now encompassed things that the first few groups had subsequently raised. And only when the system was fully demonstrated were people finally able to test and find the kind of snags you’d expect to find in any new system (“Jo can’t pick up Pete’s calls” or “My phone doesn’t ring for long enough”).
In reality, compared to most IT system installs, the training wasn’t that bad and we couldn’t blame the lack of prior training on the supplier. And, of course, the final snaglist was actually pretty short considering we’d literally unplugged one system and bolted in a new system completely from scratch.
With hindsight, however, some training prior to the system going live, better planning for the training after go-live and another person on-site to act as troubleshooter to field all the queries would have made the day run much smoother.
Thank goodness it’s over!
* * *
28 March - D-Day arrived and so did the guys from the IT company, pursued by a BT van. BT had already visited the site a few days before, so everything was ready to go, and we weren’t expecting too many surprises.
The BT guy completed the hardware changes in just over an hour: taking out the old CPE (Customer Premises Equipment), and fitting our new line. During this time, the IT guys hooked up the new VoIP switch (which is basically a rack mountable PC) in the server room, with a couple of phones dotted around for testing.
The first snag hit when BT was ready to switch over to the new equipment. The equipment itself simply refused to see the phone line. Much swapping of cables and scratching of heads later, the IT guys were forced to admit defeat and returned to base to purloin a replacement part. Et voila! It turns out that the ISDN/30e card (a £500 piece of electronics that allows a PC to connect to an ISDN line) was DOA.
This was a timely reminder that holding a spares inventory is a must for any IT company seriously thinking of offering VoIP solutions such as Asterisk. It being a Saturday, and the fact that you cannot simply buy these cards over the counter at your local PC World, I was glad that our supplier kept a small cache of “at least three of everything”.
We were rapidly approaching lunchtime, and it was clear that we were now running somewhat behind schedule. Time to divide the work and conquer.
The client had chosen a mix of hardware phones (mainly the expensive Cisco brand) along with “soft” phones which run on the user’s PCs. Ten of the soft-phone users would also get a cordless DECT headset which comes with additional software. This meant a mixture of phones and/or PCs to configure, with different pieces of software to install and configure. Plus there were two new network switches to install in the server room.
An unexpected challenge was the time taken to remove all of the old phones. These ran from conventional BT style cords, and in most cases each desk had at least one (often three or four!) extension cords in order to reach from desk to socket. It was a wonder any of these worked, given the damage some of these cords had received over the years. Worse still, many of the network cables to the PCs were just as bad (and remember that it’s this cable that our new telephone system will use). In the end, most desks were completely rewired, and we were left with a pile of old or damaged cable in the middle of the floor some three feet high!
Switching to VoIP does have one big advantage here – a big reduction in the number of cables. Even when you choose a conventional desk phone, it can share the same data connection as your PC, thereby reducing your wiring needs by half. Even where desks and sockets were not ideally placed, the finished product was much neater.
By around 7pm the finishing touches were being put in place. Incoming call handling, outgoing call handling, voicemail, call recording, DDI numbers, everything was ready and working.
That’s how it looked on Saturday night. The real test would come on Monday morning, when everybody came back to work.
* * *
23 March - As the day of the install looms, it’s time to get a little bit paranoid.
Having worked enough nights, weekends, and early mornings over the years installing, upgrading, and modifying computer systems, I know to expect at least one major unforeseen problem to occur on the day. In my book, that means between now and then, I need to focus on ironing out all of the foreseeable problems – so that I’m not already in the middle of a fire-fight when things suddenly take a turn for the worse.
Some like to use the term “risk assessment” – but I find that particular term tends to scare quite a few people away. So, in plain English, let’s just list down a few of the things that could be expected to go wrong and how we’re going to tackle them.
1. BT don’t turn up.
I cannot tell you how many times this has actually happened to me before – usually it’s a billing/ordering query which means the order goes into the system and doesn’t come out the other side.
Getting in touch with BT prior to the event is usually the best option – although to their credit, in this instance, they called me a couple of weeks ago and then this week to keep in touch.
2. Something doesn’t work
Any time a computer gets involved, be it in a washing machine, on a car or inside your phone system, you have to expect things not to always work.
In this particular case, the supplier has a “standard build” of the software, so we can expect that to work OK. They also hold multiple spares of all the hardware used in case something is dead on arrival. The person carrying out the install happens to be the resident expert on VoIP and Asterisk – whilst their office (and any spares) is just a couple of miles away.
3. We forgot something
It happens to the best of us. Don’t be afraid to go round and count numbers of staff, and then re-count the pallet load of equipment that has just been delivered, then check both sets of numbers against your spreadsheet.
In this instance, the client has made several changes to the order including numbers of headsets, spare phones, types of headset, etc. At the same time additional staff have joined the company. Better to find out that you’re three phones short before the day of the install than on Monday morning when everyone comes piling into work.
Okay, so we remembered the extras and checked out the obvious “gotchas”. I suppose we’d just better get on with it!
* * *
19 March - With BT in hand, we now have an install date to work towards. Not that we can sit down and relax – there’s an awful lot to do, including checking out all the little extras – and some of them are not so little.
Take networking. Simply slotting a VoIP phone system for 25+ users onto an already over-burdened computer network is a recipe for disaster. Dropped calls, poor call quality and a host of other problems could all be on the horizon, simply because the phone calls now have to share a computer network already congested with Excel spreadsheets and Word documents.
Not that you’d probably notice that right now. The fact that Excel takes half a second longer to load is no great shakes – but apply that half-second gap to all of your phone conversations and you’ll soon be breaking down the door of your telephone company.
A network upgrade is clearly called for, and in this case the client requires a couple of 24 port network switches. These new devices feature QoS (Quality of Service) and PoE (Power over Ethernet) – two essential components of a larger VoIP solution.
Quality of Service ensures that whenever the network gets congested, priority is automatically given to voice traffic passing through. Your massive Excel spreadsheet will unfortunately just have to wait another half a second.
Power over Ethernet does exactly what it says on the tin – it provides a small amount of electrical power to devices that are connected to it via the network cable. So you can plug your brand new telephone straight into the network point coming out of the wall, you won’t need a separate power supply or a nearby 13A socket. PoE can also power other small devices, such as networked security cameras – but not laptops or PCs.
The downside – cost. Two branded “conventional” 24 port switches might set you back around £350, but add QoS and PoE and you’ll be looking at over £1,000 from the same manufacturer.
You also need to remember the UPS (Uninterruptible Power Supply) – the emergency back-up should the mains fail. Whilst it’s often recommended, the old phone system never had one. But then again, as soon as the power came back it was up and running again after about five seconds, every time.
The same is not normally true of most computers – there is a small but significant chance that a sudden power failure might cause long term damage to the hardware or software – and even if it doesn’t, it may be 10 minutes before the system fully reboots and is 100% operational again. Moving to a computer-based telephone system means that a UPS is now a necessity, so bang goes another £300.
* * *
14 March - Most of the past week has gone by in frustration trying to organise and confirm the installation of the new phone system with BT.
The setup is slightly unusual in that the client is using an early form of ISDN/30 phone line, a product called DASS. This causes several problems:
- Not many people in BT seem to have actually heard of it
- Even when they have heard of it, trying to get a straight answer on how much it costs to convert to ISDN/30e is proving even trickier. I now have 4 different sets of figures.
- On several occasions, there was no-one available to talk about ISDN/30 and I was asked to send an email for the “offline” team to reply some 24-48 hours later. No-one available in the whole of BT to talk about ISDN/30??
- There was much confusion about arranging a weekend install, essential for the project to go ahead. Some BT staff even claimed it wouldn’t be possible or would take much longer to arrange (although I know from experience that this isn’t true – BT Engineers queue up for simple weekend work like this, often resulting in a shorter lead time, not a longer one).
Add to that the fact that the client had changed their company name about five years ago, but BT were never informed, and this week has been completely wiped out with little to show in the way of progress, but plenty of confusion.
I wanted to compare notes, so I rang up a colleague who had completed a similar exercise about six months ago – but this time with a traditional and established telecoms company. I was concerned that I was rather being left to “get on and deal with BT” by my supplier.
In actual fact, she had experienced exactly the same problem. It seems that she did get some prompting from her supplier with regards what to ask BT for – but the actual communication with BT was strictly through her.
She also relayed identical problems to mine: a lack of product knowledge (especially DASS to ISDN/30e conversion) and the fact that sometimes there was simply no-one there to talk to. She also confirmed my suspicion – BT doesn’t like giving out overtime if they can help it, but if you talk to the right people (those actually scheduling the work) then weekend/evening work is really no problem – and she’d booked with less than 48 hours notice.
There was one final thing she pointed out. Conversion from DASS is essentially a “one way” process. When it happens, you cannot go back – so if your new phone system doesn’t work (for whatever reason) then you’ll be stuck without a phone system.
It’s time to start planning for D-Day. And there’s no turning back…
* * *
9 March - In the good old days, you used to buy your telephones from the same people that provided the phone line (eg British Telecom). As the industry became deregulated, and with renewed demand, independent companies shot up to provide equipment and support to businesses. And so it has been for many years. Private companies specialising in business telephone systems. It’s all they do - so they tend to do it rather well.
Then along came VoIP. Suddenly, telephones and IT started to converge and many IT companies in the late 1990s jumped on to the telecoms bandwagon thinking it was a bit of a gravy train. After all, the net profit on supplying a single business phone was probably five times that of the PC sat on the same desk.
In reality this turned out to be a complete disaster for many IT companies - and especially for their customers. A combination of a complete lack of product knowledge (especially technologies such as ISDN/30), ropey equipment/software, and the kind of vague support that generally accompanies anything to do with IT and it is clear to why such a plan was destined to fail.
These days, things are generally much better. Equipment has improved, and certainly become cheaper – although the ever increasing complexity has grown enough to keep the majority of IT companies away. If you are considering buying a phone system from an IT company, here are some tough questions for you to ask:
- How experienced are they within telecoms? (Have they been installing/maintaining these systems for years, or are they new to the game?)
- How many systems like this have they actually installed before? (Don’t count the system they’ve installed at their office – only paying customers count!)
- When can you meet with reference sites? (For larger systems, this is essential, even for smaller systems you may want at least a quick email or phone conversation with a reference site to confirm that all is well).
- Who will be doing the work? Will they contract any parts out? If so, why? It’s okay to contract out the structured cabling on cost grounds, but getting another company to carry out the actual installation should be a red flag.)
- What is the underlying business relationship with the product(s) they are recommending? Do they hold relevant product certifications, are they development partners? Are they simply one of 10,000 “resellers” offering little in the way of added value?
- Google their recommended product(s). Are customers happy? Does it look the right fit for your business?
- What support SLAs (Service Level Agreements) do they offer? (Consider response times, out of hours cover, staff availability, etc)
- What spares inventory do they maintain? (If they don’t hold an inventory, how soon can they guarantee replacement parts on-site?)
- Who will interface with other parties/suppliers, for example BT? Will they expect you to?
* * *
1 March - Responding to our decision last week, several people have questioned my client’s choice of an “open source” package. Was it price? Isn’t it somehow “bad news” to choose something which isn’t “professional” or “commercial”?
For clarity, I have tackled the general question of open source software in another article, but feel the need to clarify a few important points here about the specific benefits of Asterisk over commercial VoIP systems.
1. Your choice of hardware
The biggest single benefit of Asterisk for any user is that it comes as a piece of software, written and designed to be hardware independent.
At no time are you forced to buy a particular type of server to run it on (any Linux compatible server will do).
Furthermore, at no time are you forced to buy a particular type of telephone – any VoIP phone will do as long as it supports industry standard protocols such as SIP or Skinny. This factor alone makes a huge difference. You don’t have to spend £200+ on a manufacturer’s own brand telephone when you can pick up a SIP phone with display for around £40 in most high street computer shops.
2. Nothing is extra
Commercial products did pretty much everything, but at a price. I’ve already talked about the fact that CTI (Computer Telephony Integration) was built in, but so is everything else – Voicemail, Conferencing, Call recording, Failover handling, and the ability to connect the system to other offices all come as standard. No add-ons, no option packs, no system extensions, no additional licences. It’s all there.
For me, the biggest potential concern of choosing open source software wasn’t the software at all, it was the fact that we were moving from the world of telecoms companies who install telephone systems every day and putting our trust in an IT company.
Of course, we all know about the reputation of IT systems. So the question is whether an IT company can really provide the skills necessary to deploy and support a 25-30 user phone system without suffering the major hiccup along the way that IT systems generally get?
It looks like we’re all about to find out!
* * *
20 February - Having eliminated Siemens from the contest, we were left with two remaining systems. For the price comparison, I specified identical systems with 20 digital handsets, 10 “fancy” handsets (larger displays for reception and certain “privileged” staff) and all the standard features such as voicemail and music on hold. We also threw in a curve ball: later in the year the client wishes to have additional features including connecting the phone system to his CRM system and having a wall board to keep track of call volumes and so on. Here are the two contenders:
This is a major international phone brand with a good reputation, and was presented to us by Admiral Voice and Data. What’s particularly impressive is that Inter-Tel offers a very consistent VoIP offering. For example, its VoIP handsets look and function identically to its “normal” handsets.
Costs: Basic system around £14,000 but with almost £2,000 as a part exchange for the client’s old system. Additional cost for database integration/reporting: a staggering £14,000 - "just to buy the additional software licence", they said.
Asterisk is well known to the open-source community – and is to telephone systems what Linux is to Windows. In fact, Asterisk runs on just about any Linux machine really rather well. Asterisk was proposed by Omniis, which is actually an IT company. As Asterisk is a purely software solution (although there is a company behind the software that makes some of the components needed to make a PBX such as ISDN/30e and ISDN/2e cards), you have to look elsewhere to find compatible phones. On the plus side, Asterisk does support most of the standard phone system protocols including SIP and also Skinny (used by Cisco phones).
Costs: Basic system around £12,000 almost identical to Inter-Tel, but without the trade-in option. However, this includes ALL of the features required of the system, including the ability to integrate to databases and web services.
- Seamless support for VoIP – right down to identical looking handsets
- Good brand name
- Cost competitive if you just want a “plain old phone system”
- Cost goes out the window if you want to integrate with other systems (eg CRM).
- In-built support for music on hold (plays MP3s), voicemail (will even email you the message), auto-attendant and even call recording out of the box.
- An open platform means you can get the system to do almost anything!
- Nothing is an option, everything is built in. No additional licences, modules, add-ins, etc
- Supports a wide range of phones and devices
- Although it’s essentially “free”, there are still implementation costs, plus the cost of the hardware which will bring it to within 20% of a conventional phone system IF you don’t want it to do anything clever. It wins hands down the moment you want to expand or scale the system.
Both systems were impressive, as were the suppliers themselves, however the decision came down to price. The costs to expand the Inter-Tel system were too great to be selected – although users with less demanding needs would probably do well to consider it. The client is going for Asterisk.
* * *
13 February - Unlucky for some...
Now, I’ve nothing against sales people... but I have spent much of the past week in the company of several of them. My experiences are as follows:
- The price quoted seems to be pretty elastic. Aside from any margins the sales person may have to play with, there appear to be quite generous offers from many of the manufacturers for users looking to upgrade, downgrade, or otherwise trade in. I would suggest that there is always a discount to be had – so anyone looking to replace their phone system should be prepared to bargain, and bargain hard!
- Every conversation almost without fail seems to lead on to the subject of the cost of phone calls (aka Least Cost Routing). I want to buy a phone system, not change call provider!
- Salespeople might know their products, but rarely know the products they are replacing. Being told that System N can do X and Y – when your current system has been doing X, Y and Z for the last 10 years is a pain.
- Salespeople seem to sell VoIP as an after thought. Partly, this is down to many telephone system manufacturers not having a particularly credible VoIP roadmap. VoIP for many systems is an “add on” which allows you to hook a conventional digital telephone system (PBX) to a VoIP one, possibly via the internet. It is not a methodology.
On that basis we have already said “adios” to one system.
Having run the sales gauntlet, we narrowed the selection down to three systems – two “conventional” telephone systems from Siemens and Inter-Tel, and a third outsider – an open source system called Asterisk.
The first to be ruled out was the Siemens. Nothing wrong with the system – the client had used the same one at a previous site 10 years ago. The trouble was the system being proposed was almost identical to the one they’d been using 10 years ago! Furthermore, it was being sold on one of these “free phone system” type of deals where you sign up to a five year lease, and the lease is paid for by the savings made on call charges. The theory goes that the customer gets a small reduction (but not a great reduction) in call costs, AND gets a free phone system.
The trouble is – in reality – the savings that can be made by switching supplier may more than cover the phone system. Last year, another (much bigger) client looked at the same deal. In that example – capital cost for the phone system was £24,000 we estimated (and it was even better in reality) that the saving on call charges would be in the region of £1,800 per month through careful supplier selection.
Even if they had to borrow money from the bank, payback was under 15 months – rather than the five years they were being offered. Siemens, you are the weakest link. Goodbye.
And then there were two...
* * *
9 February - Now that we’ve exploded a few myths about VoIP, let’s look at some of the benefits and how they relate to the client.
My client asked for several things. One of them related to the creation of a more focussed sales environment, including:
- Better (online) access to call statistics
- Recording of calls for quality and training purposes
- Better integration with IT systems (CTI – linking the telephone system with their in-house database)
In a true VoIP-based PBX setup, all calls – incoming, outgoing and internal - pass through a central computer not unlike the one sat on or under your desk right now.
The “handsets” used by staff might look very similar to what you used to have on your desk, or they could be in an entirely “soft” format – simply running as an application on your computer screen and feeding off your computer’s soundcard (or preferably a dedicated handset or headset).
Either way, the computer network is being used to transmit the voice information. Because all of the voice information is passing through a central computer, rather that simply being routed or switched, the information is already in a digital form that can be processed by the computer. It follows that many of the applications that used to cost serious money in more traditional setups can now be less expensive to implement:
- Call logging – where details of calls are recorded for billing or statistics purposes
- Call recording – where calls are recorded for later playback
- Screen popping – where incoming calls are checked against the CRM system, and the name/number appears on screen
- Automated dialling – dial direct from an application such as Outlook
- Voicemail – recording messages in your absence. You might even be able to retrieve them via the web or email.
One of the biggest single drivers for many VoIP installations is the potential cost saving on calls. These generally come from several methods:
- Businesses which have multiple sites can use VoIP to channel inter-site calls over their existing broadband Internet links.
- Remote offices may also “break out” – for example a user in Ireland can dial a UK number and be routed via VoIP to the UK office exchange before being passed over to the UK telephone network. They pass to the UK for free, so they pay UK local and national rates for that call, not the international rates they would normally pay from Ireland.
- VoIP gateways allow you to break out into other countries as well. For example, you can connect to a US gateway via the internet and effectively have a US phone number. This allows you to call the States as if you were there, and your customers in the US to call you the same way.
Now I’ve run through some of the benefits, it’s time to engage with some potential suppliers.
* * *
5 February - Last time, I looked at some of the needs and wants of my client. But as we go into the second month of the project - and before we approach any suppliers, it's important to look a little closer at VoIP technology, and perhaps explode a few myths:
- VoIP will save you money
Well, it might, and that was certainly one of the intentions of VoIP, where existing data networks could carry voice instead of having to rely on expensive public switched telephone networks (PSTN). These days, however, voice telephony costs have continued to tumble, and the market is pretty cut-throat. The real-life payback of a VoIP setup will depend on what you buy and how you use it.
- VoIP systems are all pretty much the same
Wrong. VoIP can be implemented in multiple ways. At its simplest end, you can buy a low cost telephony service which uses VoIP and get cost savings without spending a penny on changing the equipment in your hand. Alternatively, you can opt to buy some equipment - perhaps a setup using Skype or an internet phone service such as Vonage.
The systems to which these articles refer relate to telephone systems (switches) which are used by businesses to process their incoming and outgoing calls. They typically come in two flavours: those that are basically a traditional computerised private branch exchange (PBX) with “added VoIP”; and those that are designed from the ground up to be VoIP only. Both types of system will normally support all of the same types of handsets and functions – the main difference is how they handle the calls internally.
- VoIP makes it easy to connect different systems together
Just as there are many ways to provide a VoIP service, there are also many standards used, some of which are incompatible. I won’t dwell on the detail of the protocols and codecs – just point out that some standards are “open”, for example SIP, whilst other providers’ standards are “closed”, a good example being that of Skype. I’ll discuss the differences between open and closed systems and what that can mean for you in a later blog.
- VoIP is cheap
Well, certainly the cost of the equipment is cheaper than it was, but many traditional PBX suppliers still charge a premium for VoIP options.
- You need VoIP to connect to your IT system
CTI or Computer Telephony Integration has been around for years before the advent of VoIP. Certainly, connecting your IT system to your telephone system may be easier with some VoIP systems, but it will depend on exactly which system you have.
- VoIP can be used anywhere - just “plug and play”
Try to force VoIP across any poor quality network and the results will be disappointing. Typical examples of poor quality networks include office networks that don’t use the right networking equipment, or substandard internet connections (for example dial-up or services provided over satellite - which also includes whole countries such as South Africa).
Have any other AccountingWEB readers been disappointed by VoIP thanks to the hype? If so, I would be interested to hear from you: use the Post Comment button below, or email me via AccountingWEB.
* * *
31 January - Armed with my set of questions (see Monday's post), I ran through them with the client before we even started to consider approaching any suppliers.
We identified five broad areas of interest:
1. A Step Forward. The new system should be at least as feature rich as the old one – so would need to include support for Music on Hold, Voicemail (with external access), direct dial-in numbers etc, etc, etc. So far so good.
2. Voice over IP (VoIP). The client wanted (felt the urge?) to move to VoIP. Putting my consultant hat on, I asked the client to explain to me what he understood by VoIP and several factors emerged:
- He had heard a great deal about VoIP recently, and felt it was “something we should be doing”. Not a particularly powerful argument for ripping out a phone system, but I’ll go along with it for now...
- Cost savings: “Everyone knows VoIP saves you money, right?” Again, not a particularly convincing argument in this case, but we’ll explore that next time.
- Future proofing. It was understood that a VoIP system would be one which is easier to connect into other systems, and easier to upgrade and expand.
3. Integration. The system should help not hinder the sales operation – and a number of desirable features were mentioned: wall boards to display call stats, call recording capabilities for quality and training purposes, etc. In short – integration with existing IT systems, or CTI (Computer Telephony Integration).
4. Scalability. The system should be able to grow. Again, out came the consultant hat, and I asked the question, “Define scalability”. The client cited that two or three of the home workers dotted around Europe might suddenly turn into proper offices – could the phone system be hooked up to them, and expand should their offices or the UK HQ expand?
5. Budget. The client had a total budget of £15k for the work (capital costs, installation, training, integration).
With the client’s points noted, I added two more points to the list:
6. Quality. Voice quality must match that of a “conventional” telephone system.
7. Training. There is little point in deploying a brand new system if no-one knows how to use it! If people don’t know how to pick up a call, pick up someone else’s call, transfer, park, divert, hold, conference, etc – then you’re heading for trouble!
Taking this list, along with the technical specifications of the system we currently had, it was now possible to approach potential suppliers. But before we do, I’ll take a more detailed look at what VoIP technology has to offer.
* * *
29 January - If I’m going to help my client select and install an internet-based telephone system, I need to look a little more closely at finding out their needs and wants. But how does one go about that?
In their rush to make the sale, salespeople often overlook the real needs of a client. Not only could this result in a system that won’t cut the mustard today, it could mean the client could be really stuck in 12 months when it’s time to upgrade or expand.
Often, it comes down to a lack of questioning. For example, the client “thinks” that they know what Voice over IP (VoIP) is all about, and hence what they might get out of such a system. The salesperson, faced with a fairly knowledgeable client, might be tempted just to sit there and nod graciously – a recipe for disaster!
A good salesperson will ask questions – here are some you might expect them to ask:
- What is the client trying to achieve out of a new system – upgrade, expand, do new things, save money, etc, etc?
- If the client drops in a buzzword such as VoIP, clarify it. What do they understand by the term VoIP? What do they think it might do for them? What are their expectations?
- What’s the budget? There is no point in recommending a Rolls Royce when the client is only prepared to pay for a Focus.
- Some technical question: most importantly how the existing system is plumbed into the phone lines – for example BRI (ISDN 2)/PRI (ISDN 30) etc.
Running through these kinds of questions prior to meeting up with any suppliers is a good first step. It helps you better understand your client, ensures that the client isn’t making any wacky assumptions (“the system will save £2,000 a month and cost no more than £5,000”), and will generally ensure that time spent with suppliers is more fruitful.
* * *
24 January - So, this is where our VoIP adventure begins. When I first talked about doing a VoIP case study with AccountingWEB, we started coming up with all sorts of side issues and diversions from the project programme. That's what made us decide to tackle it as an episodic web log. As I document the progress of our VoIP project, this format gives me the opportunity to explore the technology and terms along the way.
And for the all the accountants reading this, I will be especially keen to compare the costs and benefits very closely, to see if all that VoIP hype really does measure up in a real business setting. Expect to share in several problems and pitfalls along the way, and I hope, eventual success.
My client is a fairly typical sales organisation - a recruitment company based in the South East of England with around 25 staff in the UK, plus a couple dotted around Europe. I have worked with them for some 10 years now, and for 60% of the staff, the telephone is still the business-critical tool it was 10 years ago - essential for finding candidates, managing client accounts, and making those all important sales calls. The human voice relays so much more information than words on a page or in an email. The need for good quality human to human voice contact in business may have diminished, but it will never completely go away.
The client's original phone system came as part of the fixtures and fittings when the business relocated back in 1999. Despite being approached several times in the intervening years by companies offering great sounding deals - (including “get a free phone system when you switch providers” - an offer which never quite stacks up when you read the small print), there was never enough incentive to change.
Last year, however, the client received notification that their system was no longer being supported by the manufacturer: any future problems with the telephone system itself would be repaired, but the system could not be expanded in any way. The voicemail system would be scrap if it ever broke.
Clearly an attempt by the maintainer to drum up some business - the letter included a “great part-exchange offer” for those thinking of replacing. How many simply took up the offer and replaced like for like I do not know - have any other AccountingWEB readers have had similar experiences? In our case, the time was right not simply to replace, but to review the client’s needs and the emerging technology, and decide on the best phone system strategy that would see them through another eight years.
So next time, that’s exactly where I’ll start.
Bawden Quinn Associates Ltd