This is a long and important post about the future structure of the telephony, messaging and communications services business. It argues that "ubiquitous services" are a thing of the past, and that the telecoms industry needs to embrace fragmentation, not fight it, in the future.
It leans on the work I've been doing on my Telco-OTT report, my recent advisory work with operators and others, as well as the findings that I've learnt during my Future of Comms events & operator engagements with Martin Geddes. (Sign up for our newsletter here)
- The history of telecoms ubiquity
- Why fragmentation is happening now
- The user appeal of "silo" services
- Why service standardisation can do more harm than good
- Why RCS & VoLTE may hasten the demise of the traditional operator model
- What can be done by operators about communications services
But first, let me set something straight.
I make a sizeable % of my revenues from mobile operators and "traditional" telecom vendors. I do not dislike them, nor do I want to see them go out of business. I'd like them to continue having the need and means to pay for my services, as well as the funds to build out faster and better networks.
But that doesn't mean I'm a cheerleader, especially when the telcos sometimes act collectively like a herd of ostriches. It is in my own long-term interests to see the telecoms industry maintain its health, and so I'm unafraid to make some hard calls in the short term when I see imminent strategic mistakes.
I also certainly recognise that large organisations harbour many different opinions and initiatives, and I meet as many clued-up and insightful operator employees who I respect and learn from, as I do relics of the telecom Jurassic period who make me shake my head in bewilderment.
The history of telecoms ubiquity
Perhaps the largest failing of many I encounter in the communications services industry is their belief that "ubiquity", "standards" and "interoperability" are always more desirable (and profitable) than "niche", "proprietary" and "silo" as service attrributes.
If anyone has any doubt about this, consider how a "niche" provider of "proprietary" mobile phones & computers now makes enough money to significantly alter the overall profitability of corporate US as a whole, while another "silo" provider has just filed for an IPO that would comfortably put it in the list of the Top 10 telcos by market cap.
Clearly for some organisations, ubiquity is far from essential. So why is it still taken for granted in telecoms?
This "standards first" mentality is most damaging in the area of personal communications - specifically around voice and messaging services. It is often taken for granted that "ubiquitous" services or applications are inherently better than more tailored or niche alternatives.
Now it is certainly true that we have a number of successful, useful and ubiquitous services today - telephony (fixed and mobile), SMS, and (for all intents and purposes) MMS and email. I am excluding "Internet access" as that is a connectivity service rather than an application.
Clearly, taken together (especially telephony and SMS) we have a huge precedent for ubiquity being "good". They generate getting on for a $trillion per year worldwide. We also get a warm and egalitarian feeling from their democratic nature - everyone can have them, whether they have a $20 phone or a $1000 super-device. Against that backdrop, it's easy to see why people hope to find the "next ubiquitous service".
It is also easy to see why organisations such as the GSMA are so emphatic about re-inventing and promoting "ubiquity" as a principle. It has had a hugely successful history, and they would like to perpetuate that trend - both in terms of democratisation of communications for users, and in terms of control and profitability for their members.
But the problem is that from now on, Ubiquity is Dead in telecoms. Unfortunately, the telecoms past is not a reliable indication of how the future will evolve. New services can never become ubiquitous again, while the existing ones will slowly, but steadily, die. (See Martin Geddes' fantastic article on "peak telecoms" for more).
That's not all. Not only is the age of ubiquity over, but attempts to "force" ubiquity may well backfire and make the fragmentation and cannibalisation worse and faster. Such attempts are inherently "brittle" and will lead to unsustainable tensions that will at some point break down catastrophically, rather than gracefully.
Why fragmentation is happening now
In a nutshell, fragmentation of communications services is happening "because it can".
The history of the telecoms industry has been about standardisation and ubiquity (at a service level) because of the costs and complexities involved in bringing new applications to market. They have needed to be hosted in the network, "hard-wired" into dedicated boxes, linked to complicated billing and OSS systems. The phones were essentially "dumb endpoints" with no ability to host much application logic themselves.
Also, historic regulatory structures have tended to drive national regimes for licencing, numbering and interconnection, fragmenting the providers of a few well-defined services to provide competition, rather than encourage competition between services. That is unsurprising, given the technological limitations of the past. In Disruptive Analysis' view, there needs to be some significant regulatory evolution to cope with the new landscape.
The old backdrop has now changed because of multiple trends and drivers:
- The growing base of open smartphones and appstores allows easy distribution and installation of 3rd-party mobile communications applications
- Home PC & broadband penetration catalysed the initial growth of popular new communications services, that have then reached across to the mobile domain (webmail, IM, social networks, PC VoIP etc).
- The general trend towards “cloud” is catalysing the separation of application hosting and network access.
- Operators are starting to de-layer into NetCo and ServiceCo divisions anyway, either because of regulatory intervention, or by the direct entry of new wholesale-centric "UTF" (under-the-floor) operators that are forcing separation through competition.
- All-IP networks – and especially LTE – are forcing a range of decisions to be made now about future operator-hosted applications and supporting infrastructures.
- There is growing awareness that “telco grade” QoS is not always the pre-eminent driver of customer adoption, or arbiter of communications excellence. Some customers will accept lower network quality in exchange for higher experience quality, for example through clever web-based interactions. Facebook and Skype exemplify this.
- Software and development tools and techniques have improved massively, along with supporting capabilities such as Amazon cloud platforms to host scalable applications.
- CFOs are aware that much of the value creation in the communications and Internet industry has been driven by asset-light, rapid-growth services such as Google or GroupOn. They want to capture some of this capital growth for shareholders, even if it does not fit traditional revenue/margin business models.
- WiFi allows a variety of communications services to circumvent operator visibility and control, exploiting fast and low-latency IP connections.This then seeds the market with applications capable of using HSPA+ and LTE networks when they reach the necessary quality.
- A few "non-Internet" OTT services have also emerged - notably BlackBerry Messenger (BBM) and various corporate VoIP/UC products.
The user appeal of "silo" services
The generally-accepted assumption in the telecoms industry is that end users always want ubiquity. Services need to work everywhere, on every platform, for every user. Isolated or proprietary “islands” are seen as impediments. Ideally, services are standardised and adopted by all operators. Stories abound of how SMS only “took off” after it become possible to send messages across networks, or how "everyone just wants to pick up the phone and call someone".
But this assumption needs closer scrutiny. Users do indeed want some services that work over all access networks but they also seem to be happy to tolerate access-neutral service islands – even where there are other constraints such as lock-in to specific device platforms.
Experience seems to prove that users are willing to tolerate (or embrace) fragmentation in messaging in particular – many see it as a benefit, in compartmentalising their groups and slightly-different modes of engagement. Users seem entirely at ease having multiple IM and email accounts, SMS/IMS, Twitter, Facebook, BBM and so forth. Disruptive Analysis believes that the same will happen with voice capabilities – they will fragment and multiply, absorbed into multiple applications and web services, rather than just being “phone calls”. Apple’s Siri is a good example of this.
Fragmentation often adds value and consumer utility, rather than reduces it.
Customers seem highly able to find the application(s) which precisely fit their needs and price/quality points, even if they don’t have the same level of ubiquity and interoperability seen in the traditional telco world. In some cases (eg BlackBerry Messenger), a lack of ubiquity has counter-intuitively been re-framed as desirable exclusivity for example.
In the past, 99% of human "distant" communications has used a few standardised models - telephone calls, SMS, email and so on.
Now, the picture is shifting. For maybe 80% of communications, there are now better or cheaper tools available - especially among closed groups such as friends or employees. "Ubiquitous" services are being downgraded to being just a fallback. If I want to invite 50 friends to a birthday party, I'll send 48 invitations via Facebook, and the other two by SMS to the recalcitrant non-users. Similarly, once in a while I'll make an old-school phone call for a pizza, as the restaurant isn't a "friend" on Facebook or a contact on Skype.
The problem is that for the telcos, the standardised services (old and new) are still expected to be primary channels, not just the secondary option, or lowest-common denominator.
In other words, new services need to be either
- The best primary option for a given use-case (& to stay the best through regular updates)
- The cheapest fallback as a compromise, when none of the preferred silos works
Why standardisation can do more harm than good
But even if we accept that services like iMessage and QQ will take a good share of the "optimised communications" market, surely it makes sense for the industry to at least try to create new ubiquitous services? Like, to name a couple of examples, RCSe or VoLTE?
On the face of it, that sounds like a smart solution. If it doesn't cost that much to try for a collaborative, standardised solution, then what's the harm in trying?
And that's where the problem lies.
In my view, failed attempts at creating ubiquity may in fact be worse-than-useless. There may be additional damage caused to operators' legacy revenue streams and customer loyalty, above and beyond that which will happen anyway.
- Some new services will be integrated with existing telephony or messaging. If the new experience is poor, then usage of the older services may drop too, and the user may investigate alternatives more thoroughly. For example, mid-2000's phones which put 3G videocalling above telephony in menus caused significant user resentment.
- Users may opt for devices that do not feature the new "forced ubiquity" services - for example unlocked $100 Android phones, or Apple devices. They may also "root" or jail-break operator-supplied devices with different software stacks
- One or more operators may pull out of the GSMA/3GPP model unilaterally (for services, not underlying network access) leading to a crumbling of traditional models in an unpredictable fashion - perhaps through removing SIMs from the value chain and moving to alternatives (eg Soft-SIM)
- Operators may be forced into infrastructure investments that are expensive or cumbersome (eg full IMS implementations). They may also delay investment in LTE or migration of users towards it, while they find alternative voice and messaging solutions.
- The time and effort wasted on failed "ubiquity" projects may have impacts beyond the upfront costs. Delay may mean relinquishing other opportunities, a reduction in work to enhance "old" services, and also allow other companies to further overtake the telcos and dictate events and pricing.
The problem is that services designed to work in a ubiquitous fashion often work poorly in a non-ubiquitous context. They're perfect at 100% compliance, useful at 90%, OK at 80%....but then below that, they rapidly become a liability, especially if they get in the user's way in the device UI. At 70% "ubiquity", the chance of any two users having the capability is 49%.
Conversely, services designed as silos typically have mechanisms to cope with that fact. They have sophisticated "invitation" mechanisms, APIs for importing and exporting communications if needed (eg SkypeOut / SkypeIn), or are intended for specific niches and communities. Their makers expect them to exist in a convoluted world - and even Facebook and Google aren't arrogant enough to prohibit their apps from being deleted or uninstalled.
Why RCS & VoLTE may hasten the demise of the traditional operator model
In two weeks' time, MWC will launch a huge upsurge of talk about the GSMA's favoured RCSe and VoLTE programmes. We might even get the long-delayed launch of the Spanish RCSe service, rumoured to be called Joyn.
For the reasons above, I think that RCSe in particular is one of the most damaging and divisive services that the telcos and their vendors have proposed. It is a huge and unnecessary distraction at precisely the wrong time. It has numerous flaws in design and conception, and in any case will have (at best) patchy acceptance by vendors. It is too little, too late, and in the wrong direction anyway. The industry can't even decide if it's supposed to compete with Internet-OTT players or complement them.
That approach can most charitably be described as "fake it 'till you make it"
Firstly, it isn't "just there" on all devices - the iPhone implementation will be app-based as Apple is unsurprisingly unwilling to implement it natively, as it would compete with its own communications products such as iMessage and FaceTime. It's unlikely to be in the native Android stack but implemented by OEMs on certain specific phones. Its relationship with RIM and Microsoft remains unclear (especially its fit with/against BBM and MSN/Skype), and it seems improbable that "vanilla" featurephones for the prepay market will integrate all the software, unless the manufacturers are convinced it will sell more devices. Even Samsung, a supporter, will have to position RCSe carefully against its own ChatON service.
In addition, the attempts to strong-arm manufacturers by getting handset certification rules changed at the GCF, or specifications adopted at OMA, look like a triumph of power over consumer demand and "ecosystem good practice". Such moves appear Machiavellian - they put the "cart before the horse". When/if RCSe has proven itself to be liked by users and successful, then it is the time to entrench and optimise it.
Worse, the effort to embed RCS capabilities low-down in the device "stack" will make it difficult to update (or bug-fix), and possibly unable to be deleted or removed by users if they decide they don't like it. Keeping a disliked and possibly invasive feature front-and-centre on a phone is a good way to breed user disloyalty.
VoLTE is a bit different and will probably emerge to the market in some usable form - but likely later than anticipated and only by some operators for some users. Also, the timelines are a bit different - it won't be deployed (in most cases) until LTE coverage is good enough, while the SR-VCC mechanism for handoff to circuit-calls has only just been lab-tested. By that time, it should be apparent to all that phones will have multiple voice services (not just telephony) and so hopefully it will be implemented to be a "good citizen" in the company of various other apps.
There is also much less rhetoric about VoLTE being "ubiquitous" - except in some vague, "middle of the 2020's utopia" sort of fashion. Apart from anything else, the telcos know that it will have to coexist with 2G and 3G circuit-only phones, as well as various flavours of VoIP. That said, if early implementations are problematic (eg poor voice quality or battery life) they will hand the battle to Microsoft/Skype or perhaps one of the new Telco-OTT alternatives. But that is a different issue to the "forced ubiquity" weaknesses of RCSe
What can be done by operators about communications services?
On a recent call with an operator strategist, a point was made to me that telcos increasingly need to be "portfolio managers" rather than being focused on traditional single-product metrics. This is absolutely the case for next-gen communications services around voice and messaging (and also video).
The winners in this space will be the ones which acknowledge that they are not the only game in town, and which aim to gain traction gradually, and alongside others, pulled along by users actually liking the service and its ongoing iterations. Designing and solidifying a platform technology, before getting real customer feedback on early service implementations is foolhardy.
Designing the mobile IMS architecture before services like RSCe and VoLTE get wide user acceptance is a 1970s-era view of telecom service development. RCSe should have been designed as an OTT-style service first and then when it met user demands (and was embraced enthusiastically), it could have been "platformised" later. Given that there are various RCSe-type OTT options (eg Metaswitch Thrutu) it is perhaps not too late.
The GSMA and the G5 operators urgently need to drop the claim to ubiquity for RCSe, ideally before Barcelona. Facebook's 400m+ mobile users make them look foolish - if reach is the critical indicator of success, they have already lost the battle. The claim in various RCS slide-decks that it's somehow difficult to to install 3rd-party services is even more risible: it's much easier than going to a shop to buy a new RCSe-capable phone, while the operators themselves will be depending on that much-derided model for iPhones and existing Androids.
Even "it's just there" is empty rhetoric, when even prepay SIM cards can work out-of-the-box with OTT services like BBM, while it clearly won't be "just there" on iPhones at least, and probably many other devices. And let's wait to see how RCSe actually operates in the user's hands before we claim "it just works....", when we know if the fourth word is "well" or "badly". If it's the latter, it will do the operators more harm than good.
My view is that proponents of RCSe should forget about "ubiquity" as a driver, and instead focus on specific niches or use-cases, and evolve the service over time based on user feedback. Because this may be country-specific, it should also move away from the multi-operator interoperability model, and towards a Telco-OTT basis. It may be that Vodafone does something clever with RCSe implementations that Orange does not. It could be that we each end up with three or four different RCSe-OTT apps on our devices if they're done well.
Operators should be thinking about specific use cases and service "look and feel". They should absolutely have a long-term aim to launch services that could reach a billion people or more eventually, but should not try and leap straight there on the first iteration. They need to adopt Google's approach and try services like Wave, Buzz and Google+ and be prepared to kill them and move on if they don't work.
Creating binary "unkillable" services that can only ever be ubiquitous or entirely absent is an existential mistake. Given what's going on in the industry, trying to create "forced ubiquity" will likely accelerate the movement towards "total absence".
Economists might want to think about an unfortunate parallel here, with attempts to force a certain "ubiquitous" currency on a diverse group of nations, leading to a possible "disorderly breakup" when one defaults on its debts, with "contagion" taking down the other dominos via "interoperable" holdings of debt by banks.
When the first major operator "defaults" on the 3GPP/GSMA world-view, the outcome will not be pretty either.
So for telecoms, ubiquity is dead from here onwards. Even if RCSe miraculously succeeds, its most optimistic destiny is to become the lowest-common denominator fallback when nothing else better is available. The sooner telcos, their vendors and the GSMA admit it, and embrace fragmentation and flexibility, the sooner they can hope to recreate their role as services providers of choice.
Whether operators partner with Internet companies or each other, acquire promising Internet startups (eg Viber, Tumblr, WhatsApp, Pinger, Path, Yammer etc), or build their own Telco-OTT services from the ground up, an acceptance of "communications diversity" is a non-negotiable starting point.
They need to find best-of-breed communications apps, because the value is going to shift permanently. "Ubiquity" = "lowest common denominator" = "minimal value".
(Disruptive Analysis has recently published a report on Telco-OTT Strategies. Click here for more details)
Very nicely written. Believe we end up with "younified communications". No ubiquitous PSTN-like entity but a combination of diverse tools - selected by each individual on a use case by use case basis - http://goo.gl/5Rn0h
Great post Dean! One related point is that I think NEPs are at least as responsible as carriers for choosing a trajectory towards 'forced ubiquity'. They have many representatives at the standardisation bodies who decided to specify a single service layer technology for all networks. 'Cookie Cutting' the same IMS architecture and deploying it on top of all access networks is more in the interests of NEPs than carriers. The dynamic for NEPs has also changed recently though ... not least due to the prevalence and wide adoption of open source projects. The reason why you have so many cool new Apps from small companies is that projects like Asterisk and Mobicents now let a start-up do for free what NEPs used to charge millions for. The failure of IMS to generate the levels of innovations that all the OTT start-ups offer is as much a challenge for NEPs as carriers.
Agree with the second part of the headline and finally someone writes it down. eRCS is just a scam to please investor analysts and telcom executives: "we're doing something", but doubt it will be success with the consumer.
But on the first part I don't think "ubiquity" needs to be dead. actually the opposite, the time just hasn't come yet. all these new players are using the same standard xmpp that will allow ubiquity once everybody starts to open up, like for example Google already dead. plus there are new services like yuilop in germany that combine the best of both worlds, the user experience of the IP-based OTT services and the ubiquity/reachability of the telco services (like SMS).
Wouldn't it eb easier to just write one b'day party invitation message and send it via IP to your social graph and via SMS to the few 'old-school' friends, all for free and simply in one app?!
I think ubiquity and best user experience go well together.
what doesn't work are OSS (network) layer based services like telcos still try to push.
it's about software and cloud, light core, heavy leaves. not proprietary network based services
Quote #1: Clearly, taken together (especially telephony and SMS) we have a huge precedent for ubiquity being "good". They generate getting on for a $trillion per year worldwide.
Quote #2: "Ubiquity" = "lowest common denominator" = "minimal value".
I'll take $trillion per year worldwide "minimum value" Dean. You can have what's left.
So, you do make very many good points, but there's so much other noise its hard to sit through a reading. A few good points are:
1) Previous use of standards has generated very large user bases generating enormous amounts of cash. Good.
2) Carriers are through natural evolution becoming ServiceCo and TransportCo.
The other junk about don't design IMS until you've designed the applications that require it and prove its required is silly. And saying that a hugely fragmented situation is good assumes that all users are power users, which is also silly.
The carriers are going to do IMS and RCS and VoLTE because that is what makes the most sense given the work they're responsible for (e.g., E911), and the work they want to be responsible for (e.g., sexy sticky multimedia). Multimedia applications will RULE. And, multimedia applications tied to telephone numbers and which are "smart enough" to know what complex service experience to deliver will RULE. Skype and Facetime will be marginalized if not outright exterminated. And, if the sexy stick multimedia applications are ubiquitous and standardized, so much the better. Now, all that being said. Dean may still be right and I may be wrong. ;-)
Each individual is their own garden in the commons. No walled gardens.
Go back to 1984, when WAN competition in the US was created, followed by data digitization en masse (unlimited ISP dial-up) and then competitive digital wireless competition (aka CDMA and iDen). It all happened in the US after the genie was let out of the bottle; along the way Moore and Metcalfe kicked in and assisted. My epiphany in the debate you do a good job documenting goes back to 1991 when a little band of market driven competitors were pushing frame relay vs the monopoly bell SMDS standard. Knew back then which would win; know now which will win. The Facebooks, Apples and Googles of the world are developing "exchanges" that provide for centralized procurement of services, much like 800, vpn and calling card that developed from the competitive WAN in the US. A great blogpost from Fred Wilson frames this phenomenon: http://bit.ly/yQ73xT . The bigger problem as I see it is the tidal wave of applications and resulting demand overwhelming the lower layers; much like with Web 1.0 in the late 1990s and the lack of broadband. Changing metaphors, LTE is a firehose against a forest fire and WiFi offload is a bandaid at best; although there is a good deal of opportunity in the latter. Portfolio management would work if they thought along horizontal, not vertical, constructs and were positioned to generate ROI from vertically complete solutions (there's and other's)in the competitive "exchange" model. Monopoly driven standard solutions have been failing for the past 25 years, with only Wall Street and govt regulation perpetuating the model.
Post a Comment