Wednesday, September 5, 2007

Subscribers are Over-rated

All these years, I have heard a lot about how India is an awesome destination place for mobile application/product vendors because of the huge subscription base. The 50 million, 80 million subscriber base per service provider makes all vendors lick their lips thinking that if they can even take a 10-20% of that pie, they would make it big.In fact I have had the chance to review quite a few business plans which goes on the above surmise.

I have always been convinced, subscriber base means nothing in India. And the reason for that is because even though there are so many existing consumers and likelihood for even more in the future years, the problem is that the revenue per consumer is so less that even huge volumes do not mean a whole lot of profits. And I have been harping about this for a long time. Service providers DON'T pay upfront and the model widely used is managed services/revenue sharing, which means the cut that goes to each of the vendors involved in the overall solution is so less that sometimes it is even negligible.

This is the reason why you don't see a whole lot of Telecom product companies in India enroute to success. Yes there are the Jatayu Networks/Saskens of the world, but they are more the exception than the norm and I am making an educated guess that they survived because of their international orders and not because of local sales.

Today, Sunil Mittal, the chairman of Bharti Airtel has issued an interview where he says
the Indian Telecom sector has reached a stage where prices could not decline further and would only firm up in the future. I am one of the better users of Airtel ( I pay a huge monthly fee, make international calls, roam around a bit), yet my bill is still close to 40-50$. So think of the millions that have a cell-phone only for incoming calls (incoming calls are free in India) and pay a very nominal monthly fee. Bottom-line is when analysts take about the subscriber base in India, they would need to consider two kinds of bases. The one which gives a decent ARPU (Average revenue per user) and one which is dismal. That would give a more truthful picture.

At the same time,the only people who will benefit because of the huge volumes are those companies that are looking to load-test and benchmark their products.Nothing is as good as testing your product in live scenarios and if yours has withstood 80 million subscribers, it HAS to be good. India, sadly, as of today is more a testing ground than a viable market.

I believe the subscription rates are going to rise and that is good for the Telecom vendors/service providers. Not so good if you are merely a consumer.

Tuesday, September 4, 2007

Nortel's take on Cisco-Microsoft Partnership

Tehrani has posted Nortel’s Tony Rybczynski reaction on Microsoft-Cisco partnership. What it tells me is that Nortel is not concerned too much about the new found love between Cisco and Microsoft.

It stems from the fact that Steve Ballmer made it very clear that they were on the same page with Nortel and were natural partners. Must have something to do with talks about Microsoft making a bid at Nortel. I predict it's not going to be too long till we see a formal bid by Microsoft to buy Nortel!!!

Wednesday, August 29, 2007

Third Party Voice Control

There are enough mechanisms and more ( like JTAPI, TAPI) etc which allow you to write applications that can do third party call control. My question is, are there mechanisms to do third party voice control?

Let's say that there are two ip phones which are controlled by your centralized IP PBX. Now I can write applications that can do a makecall(), answercall() etc on these two phones. But media will be to/from these phones. What I want to do is makecall() from one phone, but have the media terminated somewhere else.

I don't want to connect and then do a transfer, that would be a round-about way. But having said that, if there are no other ways to do it, then this would be the way to take. Let's come to the scenario where this would be useful ( would this be the next killer app?)

Let's say I have a dual-mode mobile phone with wifi and let's say that termination to a TDM line ( mobile/landline) is cheaper or has better quality through the GSM network. Or it could be as simple as I have way too much free minutes available during the weekend. Now, it is obvious I would be carrying the phone all the time. What if my wife is sitting somewhere else and wants to call somebody through my mobile GSM connection? Assuming she is connected on the Net, she should be able to reach the wifi enabled side of my phone and should be able to make a call out through the GSM world. But the media coming in should go to the softphone she has in her desktop/laptop rather than to the mobile I have.

The media will take the following route. Laptop->wifi side of mobile->GSM network->end user and vice-versa. This is more like a three party conferencing, only thing me sitting in the middle am not doing it, rather the third party is doing it. This would be the reason why I am looking for ways to do third party Voice control.

You don't need things like Skypeout minutes and you don't have to be with the phone to make/receive calls. Say, I forgot to carry my phone. No worries, all I need is an internet connection and I can still use my phone. Something like a pc-anywhere on my mobile phone, only simpler and easier.

Cisco Microsoft Collaboration

I was talking with one of the thought leaders/industry observer in the Unified Communication space and he came up with an interesting theory as to what areas Cisco and Microsoft can collaborate.

As you might already know, last week Steve Balmer and John Chambers decided to set aside differences, as they call it, "for the better good" ( Read "destroy Asterisk") and work more closely so that they can offer meaningful solutions to the end customer. Based on the discussion I had with the gentleman mentioned above, it became pretty evident that Speech recognition is one such area where they most probably would work together.

Microsoft had a pretty vibrant Speech Recognition development team until one of the senior guys Kai-Fu Lee defected to the other side (Google). People who had been following that story would know of the legal battle that took place between Google and Microsoft when Lee left Microsoft. What is not commonly known though is that Lee was the chief evangelist of the Speech Recognition practice in Microsoft and most of the innovations in MS-Speech was his brain-children.So Microsoft kind of lost steam when he quit. They were already way behind market leader Nuance and this did not help their case. But it looks like they have started to become an active force in this domain. From what I hear, some releases are on the way.

Now about Cisco. Cisco has been way behind in the contact center business and even way behind when it comes to self service technologies ( IVR). Their product suite that includes IP IVR and CVP largely depend on Nuance for the speech recognition platform and licenses. So let's say a customer wants Cisco CVP with speech recognition. He ends up paying a lot for Nuance licenses and since there is nobody really to challenge Nuance's position, they have been having a free ride so far.

This is where Cisco and Microsoft could get together. Cisco, like a lot of other Microsoft products, could OEM MS speech and include it as part of its platform. What MS speech lacks today is credibility/customer base and if Cisco can OEM it, it would be a good start for Microsoft. It would help Cisco because the cost to the customer is less and a better value proposition. If Steve and John are going to put their money where their money is, Speech technology would be a good place to start!!!

Wednesday, August 22, 2007

Most Companies Don't Understand Call Center Business.

It has become the norm these days that if you are a player in the IP communication space, you think it is just a natural progression to cater to the Call Center needs. And looks like Skype is no exception. Let me get this out on the onset. IP Telephony is NOT the same as IP Contact Center. If you have agents on IP phones (be it Skype/Cisco phones) it does NOT mean you have a contact center solution on IP.

I don't blame companies that think "Hey, I have a softphone that can do voice/chat. I think I have a call center solution". Why, because I used to think that too in my previous assignments. Now that I am with a predominantly focussed Call Center solutions provider I understand that there is a vast difference between me making a VoIP call and an agent doing the same.

This is how the progression should be. Companies that have vast expertise in Call Center technolgies ( be it the traditional TDM ACDs/IVRs) must lead the way and adopt IP as their underlying platform not the other way around, where IP companies start providing contact center solutions. The problem with the latter is that, they are just "emulating" the older solutions and as far as I can see there is no real innovation.

Yes, Presence based ACD ( i.e an agent transfers the call to the subject matter expert based on the expert's availability) is something only an IP solution can offer and I am sure this was based on the presence model that IMs used to offer, but still these applications are the exceptions rather than the norms. For example lets take Cisco's ICM ( Cisco's ACD). What does it do that an Avaya AES or AIC does not do from a feature set point of view? Answer is nothing. So where is the innovation?

Now coming to what triggered me to write this post. In the news-item mentioned above, Skype is offering live-chat to customers. OK, so what is so great about that? Who does not give that? What is so "intelligent' about that? The customer gets to speak to the subject matter expert. Again, so what? It also says that Skype will provide configurations to switch between Audio/Chat windows. That shows how primitive the thinking is. Have Skype's people ever gone into a Call Center and had a chat with the floor managers? The first thing they want is an unified screen. Any novice will tell you that. What are the integration touch points that Skype provides? Do they have connectors to MS CRM, Siebel CRM? Or are they going to provide even more "configurations" to switch between three windows!!!!

Enterprise Telephony will soon merge into Contact Center Telephony. I envision that happening in the future. But not today and the people who would need to drive it are the experts and not the novices.

Friday, August 17, 2007

Buzzworks & SearchtoPhone : Same Apps in two different parts of the world

I know what Buzzworks are trying to do, since they demoed their product in I have been intending to write more on their product but never got around to it. Today GigaOm has a piece on a company called SearchtoPhone and once I read what they are upto, I thought it was time I wrote about Buzzworks.

Basically, this is Buzzwork's premise. You want to reach one of your neighbourhood services but you don't know whom to call. So you call up a number and do a voice search ( through open source speech recognition) and the centralized software searches for the most likely candidate and calls them up for you and you get to talk to the guy whose service you want to use. But from what I see, SearchtoPhone is more or less similar except for one bit that the author neatly summarizes

Using voice recognition and knowledge of your location, Search-To-Phone determines merchants or service providers who might be able to help. These businesses receive calls from Search-To-Phone and listen to your recorded job request, decide if they want to talk to you further, and place a call to you via Search-To-Phone. Your phone number remains private.

Interesting to see two companies in entirely different parts of the world coming up with similar ideas at the same time. Had they been big companies, this is when the litigation would start :) Of the two, SearchtoPhone has a slighter edge because the caller maintains his privacy. Having said that, it shouldn't be too long for Buzzworks to incorporate this feature as well.

I don't know about what powers SearchToPhone's software. But in case of Buzzworks it is open source all the way ( Asterisk, openSER and oopen source Voice recognition software). So I am going to take an educated guess that Buzzworks would be the cheaper solution.

The problems I see either of them facing is that they have to work with the service providers and that requires lot of time and patience. On the technology front, I am not a big fan of Voice recognition software ( yes, that includes leader Nuance as well) and the accuracy is not something I would bet my life on. Buzzworks would have more problems because they are powered by open source ( not sure how stable that would be) and secondly, the dialects/languages in India are so much and I would be surprised if accuracy is more than 20%.

Two companies that I surely will be following for the next few months.

Monday, August 13, 2007

Irony : India fights China using the Chinese Model

Rumours have been flying around on this for sometime now but has just been made official. DoT ( The Department of Telecom) is planning to setup a Testing and Certification wing which will monitor security compliances to which all telecom operators in India have to comply by. The first rollout of these compliances will be setup as early as early 2009. Whether you are a private or a public operator you would need to pass these security tests to operate in India. This recommendation closely resembles the Chinese model where operators/equipment vendors have to adhere to strict norms. China's Information Technology Certification Centre enforces these norms.

The government can say this is to prevent "all" hardware/network vendors from accessing their systems through back door channels. But the bottom-line is, it is to prevent the Chinese vendors from doing this. They are not concerned about Alcatels or Cisco's of the world. They are more worried about the Huaweis of the world which has the Chinese government's backing. China, a rising power, has a lot of countries nervous ( including India) and they believe China might use the back-door channels to cripple Communications in case of conflict. The release says

DoT is concerned about the advent of 'new names' in the equipment vendor space which are of 'suspect origin', but extremely competitive in cost with regard to Telecom tenders.

This is specifically targetted at Huawei. What Huawei typically does is, off loads the network/communication gears to the Telecom provider for literally no cost and has an agreement that the Telecom provider can pay in say 10/20 years. This obviously has caught the attention of the Indian government and the security compliance is just a beginning. I foresee more stringent laws coming.

The compliance would include removing any inbuilt remote diagnostic facility in the equipment. This would be a major blow for Product companies that are planning to build NOCs ( Network Operating Centres) to monitor their products across locations. This would increase the cost of support, since resources have to be 'on-site' throughout.

These extra tests and support would mean more money spent and it remains to be seen who is going to take the brunt. It is very unlikely that the equipment vendors are going to take the hit, and I say this because the equipments are already given to these Telecom vendors at throw away prices. I know of a case where one major networking giant had to give an eighty percent discount on its products to convince one major Telecom provider in India to buy it. So they will not bear the brunt. And we know the Telcos never take the hit ( Is it just me, or are they turning to be the next-gen Oil Companies?). So it might be the end consumer who might be forced to shell out more.

It needs to be seen how badly the consumer will be hit. On the brighter side though, if you are in the Telecom testing space, this is bonanza time :) Interesting times ahead!!!