Should we build PSTN 2.0?
The Telco 2.0 team have been enjoying some much-deserved vacation time. One of the side-effects is it gives you intellectual space to ponder some of the big issues.
Our industry survey told us that defining the services that IMS is due to support is a big issue. We also know that traditional voice service dominates industry revenue. Unlike Internet-based competitors, this voice service offers ubiquitous availability (every device supports it), universal interoperability ("direct dial" isn't even in the vocuabulary any more, it's assumed), and sophisticated pricing and payment options.
What happens when you put IMS and voice services to together? What's the evolution path? Is IMS only good for cost elimination of legacy circuit-switched gear? Or should IMS really be the bedrock of PSTN 2.0?
After all, the GSM Association has plans to keep the messaging money tap open by evolving SMS into mobile IM. Could the telecom industry out-distribute the Internet players in new voice services? Can "phone companies" extract enough value from their billing, identity, distribution, support and retail assets to overcome the inevitible functionality gap with nimbler Internet players?
Undoubtedly the challenge is steep, sufficiently so that many would hesitate to even ask the important questions:
- Do the features of advanced telephony require common standards, or are most compelling features ones that each operator could deploy independently?
- What would PSTN 2.0 look like if it came to pass?
- How in practise could it be brought to life?
- What are the competitive challenges and technical problems with it?
- What alternatives are there to maintain revenue from voice telephony?
So, let's lie back in the sand, plug in our headphones, move our sunhat over our face, and contemplate each of these.
As if by magic, my voice whispers in your ear from across the world
If I'm going to talk to you, we need to have a common means of doing so. The constraints of SS7 gave us a pretty uniform experience in POTS. Internet protocol creates a new neutral interface for underlying connectivity. SIP (despite some standards fragmentation and design problems) provides a common signalling mechanism, and likewise there's no shortage of standard media codecs.
Yet, to create an application we need a common meaning overlayed on top of that SIP infrastructure. This is both a technical and social issue. To have the option of depositing a message directly in your voicemail inbox without your phone ringing, we need to agree the API.
I also need to know what your social expectation is. In today's telephony, if I call your mobile number, it's not my embarassment it wakes you at 3am in your hotel room whilst roaming on the other across the world. Add in explicit time zone support as an 'availability' feature, and it does become my concern. If you assume I ignored that information, when in fact I didn't have access to it, we've got a problem. Without a common social framework, the application may break even if technically flawless. This implies a common underlying meaning ("semantics") to the application, identities, service capabilities etc. with relatively small operator-specific enhancements that don't disturb that.
So that answer to the question of "Do we need a new common application standard?" is "Yes!". Features will cross operator/application network boundaries, necessitating common technical standards; and a common set of features is required to make the system socially scale. (I've also not mentioned resisting voice spam, which also demands no leaks and common defenses.)
The question is who gets to do it, and what's the business model given the ultimately inevitible demise of the metered minute?
Features, features, everywhere
Before we get to the business model, let's just look at what people actually do with existing IM/voice clients such as Yahoo! Messenger and Skype that they can't typically do with a cell phone or landline:
- Ability to send multiple media types (audio, text, emoticons, URLs, pictures, video, files) as part of a single conversation.
- Presence, availability and "meta-presence" (e.g. additional mood text as well as "Available").
- High levels of privacy and access control via buddy lists, multiple identifiers and opaque directory data only visible to approved receivers.
- Integrated directory, history and search of users and past communications.
- Extension across multiple device types.
- Always free on-net usage (that's "free" without an asterisk).
They sometimes offer integration with voicemail or message storage systems, although the functionality and integration tends to be weak and may come at a price. Communications may also be encrypted, although this is not universal.
This list gives us the baseline functionality: this is table stakes. Yes, even free on-net calling between consumers once they've paid for the underlying connectivity -- remember, we've already said last rites for the metered minute.
Yet this list tends to address the strengths of the IM players (i.e. PC software development) whilst ignoring those of telcos: payment, universal access (including businesses), advanced management, distribution and support, high quality, and strong identity (SIM cards, credit checks, or simply knowing where you live and who you are). Most notably, current IM systems fail to cross boundaries between consumers and businesses, as well as performing poorly at scaling to mass, open public use in the way the PSTN has (this requiring pre-approval as a "buddy" of potentially hundreds of people).
Thus we should expect additional features from a PSTN 2.0:
- Explicit negotiation of any charges between sender and recipient of any media type, including voice calls.
- Ability to bundle connectivity for free ("next-gen 800 numbers") for all types of media.
- Acceptance of multiple types of identity, not just phone numbers, and to behave appropriately in each case.
- Strong forms of privacy to protect users from intrusion by businesses with whom they interact.
- Conversely, the ability to reveal a great deal more identity information via digital means to businesses or other users. No more dictating names and addresses to bored Indian call-centre operators.
- Integration of payment systems for completion of transactions with businesses.
- Integration of all forms of device and personal identity into the communication. This offers stronger privacy protection and blurs the difference between a buddy and a general contact or address book entry.
- Higher and more predictable service quality.
- Stronger audit and encryption for commercial use.
- Better integration with other identity sources (e.g. enterprise directories) as well as fine-grained role-based management of all these relationships.
- Radically better voicemail and messaging, given the comparative advantage here to IM systems.
The roadmap for the telco telephony system thus diverges significantly from most of the Internet players. Adverts, auctions, entertainment content and operating systems take a second place to simple communication that scales beyond circles of friends to all public spheres. Some revenue may come from select premium features, such as centralised storage, or aggregation of capacity to enable conference calling. The big bucks come from from bridging the islands of consumers and businesses, as well as lowering the barriers to safe conversation between strangers.
A hard journey with many dangers where most get lost and perish
The soothing waves sloshing up the beach have lulled us into a dream state that allows us to imagine what the destination might look like. Is there any viable route from "here" to "there"?
This is where things start to unravel. Your correspondent naturally took a plane to the seaside, and discovered a few things along the way. Sending text messages whilst roaming sometimes simply doesn't work: Message send failed. (You still get charged.) The "+" international prefix for making voice calls doesn't work everywhere as it should; dial "00" as the prefix from your mobile phone, and the call goes through.
If the industry can't get vanilla voice and messaging to work properly (let alone reliable interoperable MMS), what chance is there of executing on such a grand plan? And how likely is the target functionality going to fit the real needs of users (again, shades of WAP, MMS, and unused portals)? Services designed by committee don't have a good track record of market traction.
The rapid iteration and permanent "beta" status of Web 2.0 applications is an outcome of Darwinian commercial evolution: anyone with a heavyweight approach to discovering real user needs went out of business in Web 1.0. Few telcos are likely to acquire such competence; and a mass outbreak of excellence at application design and implementation is even less likely.
Thus the Telco 2.0 suspicion is that operators have only weak competence at the content and applications space. Timescales for deployment are glacial -- and like the glaciers, such efforts are prone to melt and disappear before their time. It is necessary to specialise in the horizontal functions you do well. That means realising some of the same commercial goals of liberating value from payment, billing, identity, etc., -- but without the baggage of application and hardware design that defining a full service entails.
So, how do we get the feature list and revenue, but without betting everything on one particular application architecture?
Collaboration, collusion or cartel?
We already have the capability to replicate POTS using SIP and ENUM. However, the definition of advanced features is almost totally absent. Rather than invent a new application called "PSTN 2.0", it may be necessary to take a more nuanced approach. To see why, you need to understand the competition.
There are thus two ways in which common technical and social understanding of a communications system can come about. A group of telcos and vendors get together to create a new public standard, or a private entity grows to become such a de-facto standard. An analogy might be with payments: "Cash 1.0" was based on analogue paper and only required you, a government (the public standard issuer) and the merchant with no interlocutor. "Cash 2.0" is digital and tends to be mediated by a private payment network such as VISA.
Ideally, telcos would like that private entity to be themselves, collectively -- just like VISA and the banks.
Ultimately, the telecoms operators are competing against a number of other players: the Internet giants, media companies moving into social media and communications, and transactional players such as those same banks as well as Ebay/Paypal/Skype. We suspect the industry is going to polarise as a number of associations based around different scarce resources: telcos and their licensed spectrum; commerce hubs (Google, Ebay) which introduce buyers and sellers; Microsoft, as a 3-way network effect between users, software/interactive content developers, and OEMs; and media players like Yahoo! and BSkyB/News International. Whilst the telcos think they're competing against each other, others plot to snatch away the pie before the winner is chosen.
So if the network operators might struggle to define new application standards, they instead need to ensure they win regardless of the power struggles between the above axes of power. That means adopting a platform play around the assets the others lack. This is a much richer idea than just technical APIs, and broader than IMS. It means defining means of access to billing, distribution, logistics and support systems -- in other words, the whole business, not just the network.
Perhaps the best current example is SMS short codes. These are a very simple instance, but you can easily imagine a common platform for creating and managing these; standard means of short code business customers interfacing with the telco CRM and support systems; common interfaces to manage payments and refunds, and so on. A business platform, not (as with IMS) just a technology platform. The seed is there; it just needs watering and fertilser to grow.
The network operators will have to work together to protect their turf. This, luckily, is something they are more than accustomed to doing. It will require some careful positioning with the regulators, however, for them to see that genuine inter-system competition is emerging, and that real public benefit is being generated. Moving from technical standards to common business standards carries political risk.
Fight for your right to profit
The "phone companies", for want of a better term, are at risk of carrying all the costs of the physical infrastructure (network and distribution) and messy credit and fraud management, whilst receiving few of the spoils. The traditional telephony system will slowly erode, as new forms of communication that are more convenient to their context encroach, and people with powerful hubs external to the industry use their leverage to the maximum.
At the moment, the telephony industry isn't even in the ring to put up a fight, having lost its confidence following a half-decade of doldrums and decline. The worst-case outcome a decade hence is that Skype takes the 800 number business through integration with Ebay and Paypal, Yahoo the premium call revenue by extending other premium content businesses, some unheard-of upstart the consumer-to-consumer calls ("MySpace/MyTalk"?), and Microsoft/Nortel federate every business communications server. Many other combinations are plausible -- pick your own bogeymen. In each case, the telco outcome is bleak.
The escape route is a "Telco Industry Inside" platform approach that ensures consistency and interoperability, low cost of development to business users, and an unaccustomed openness. The outcome may not be called PSTN 2.0, but a Telco 2.0 platform and connectivity business will most certainly lurk underneath.
Sunny days beckon if the right moves are made, even if immediate retirement to a beachside property remains a distant prospect.