SDP Global Summit Summary: A Significant Change in the Industry. Rays of Light Emerge from a Once Cloudy Sky.

Overall Summary

The fog of ‘hype / BS’ and ‘strategic consulting organization misdirection’ lifts, revealing a focus on real business coupled with an eye to long term survival.  The conference was packed with quantified API case study successes.  The end of this weblog has some slides with a few highlights from the conference.

The biggest changes from last year’s conference are:

  • Real numbers are presented and real business is being generated through Telecom APIs, we’ll review those numbers through this weblog article;
  • A pragmatic view on the role of Telecom APIs, no longer chasing the long tail in the hope of acquiring ‘telco developers.’  Though the long tail is not given up on (remember no black and white thinking), rather the focus shifts to internal and partner developer use;
  • A recognition on the need for API flexibility, API standards are not enough, standards bodies should perhaps act more as a repository of best practices across public and private APIs, than setting public APIs in stone before the market has decided;
  • A focus on the operational benefits, business models, propositions, and innovations supported and enabled by APIs, not the APIs themselves;
  • WAC is dead (joined forces with the GSMA if you believe what the GSMA says, such BS only degrades their credibility), but the application of APIs to new and old business models continues, with operators discussing the federation with other operators’ APIs;
  • A focus on the processes around APIs, a great sign of maturity when the discussion moves away from technology to best practices in using technology;
  • Lots of APIs, far more than OneAPI and those specified by OMA, are in use and making money for operators today as discussed in the conference summary slides at the end of this weblog;
  • The recognition that there is no one SDP (it’s a nebulous concept at best) rather we’re focusing on the services domain (across people, process and technology) with the creation of new divisions in Telefonica, Etisalat and SingTel.  This change is discussed in the Services Domain Report.  The technology in the services domain is generally off-the-shelf IT components based around a SOA (Service Oriented Architecture) which is evolving towards a light-weight SOA / web services model as SOA is too chatty (using resources unnecessarily); and
  • The API gateway has become a recognized category.

The conference was packed with operator case studies giving very frank reviews of what is working and what is not.  Best practices were shared, the conversations were open and collaborative, making it a great event.  Last year Jose Valles was a lone operator voice on the challenges and opportunities of Telecom APIs, today most of the operators were similarly pragmatic.  Coming away from the conference it felt like we can finally focus on growing the business.  At the end of this weblog are some slides showing just a few of the many highlights from the excellent presentations.

I do have a few pet-peeves that I need it get off my chest (marketeers please take note, everyone else please ignore):

  • It felt like I had been time-warped back to the 1970s with the use of term “digital” everywhere, remember those digital watches, that’s when digital was cool.  So please no digital lifestyles, its weak.
  • The word ‘smart’ as a preposition to anything, for example cars don’t need to be “smart” as that’s where we use the latest devices (perhaps 1 or 2 year old at most) we carry around with us all the time not some ruggedized 5 year old technology that is built-in.   I know there are lots of well-healed technology laggards willing to pay $1500 for the “convenience” of a built-in poorer navigation service compared to the free one on the phone, it isn’t a status symbol, it’s just plain dumb, in time we’ll look on at a built-in navigation unit the same way we look at Burberry thanks to chavs.
  • The over use of the word connected, e.g. “billions of connected devices to your network”, with the implicit meaning being billions of dollars earned from your network.  This will not be the situation as the business case for most of the M2M devices is not going to be on the mobile network.
  • I’ve said this one before, but please no more use of the word monetize, the guilty phrase is, “We have technology X that enables you to monetize asset Y’, when what is meant is “We have technology X that we hope could generate some cash for you but we’re not too sure, and whenever we get specific the ideas turn out to be a bit crap.”  As an industry we make money by delivering services, and can save money in operating more efficiently.  So let’s focus on what specific services or specific savings can be achieved rather than saying monetize and then not following through with any specifics.
  • And finally on the vendor presentation slots during the conference, please do not repeat the marketing nonsense of your websites such as “smart connected digital lifestyles of billions of devices,” the slot is an opportunity for you to shine, OpenCloud let Bouygues shine for them, Oracle let Orange shine for them, or you can do your own thing but make it original, insightful and ideally entertaining.  The Orange presentation had us talking about cannibalistic penguins which though surreal did generate a laugh.  And that’s what a conference is all about coming together, learning through sharing, and enjoying the time we have together in a nice city.

Highlights from the API day included:

AT&T is generating close to 1B API transactions per month through their APIs, an API transaction does not necessarily correspond to revenue, particularly for cloud based services, where the APIs are simply using the service.  The numbers we’ll need to focus upon as the business matures are: revenues, margin, and operational cost reduction per bit or per transaction, etc.

On the payment API there was a recognition of the problem of the mobile payment fee, people were not using the term revenue share anymore.  This demonstrates a much better understanding that it’s about a payment service to merchants enabled through an API.  It was heartening to see the language, thinking, and realism moving forward so much in just one year.  There was a frank and open discussion on the opportunities to copy the leading experience of ‘one click purchase’ under customer control with trusted partners, which is likely over 80% of the transactions.

There was a recognition that Twilio has got it right, and there is no reason why operators should not copy, only the flexibility of their platforms is limiting them.  An important point is it is not just about copying the Twilio API as developers want cross-carrier APIs, extensive how-tos, great dev community engagement, and all the “for free” functionality like conferencing Twilio provides. And once developers have made a decision to build on Twilio, they will be unlikely to change unless Twilio has performance problems. So it’s not just about copying APIs and offering them cheaper, it’s a broader business problem.

There was no discussion on WAC (Wholesale Application Community), last year’s WAC day made its death a foregone conclusion.  But there are important lessons to learn and operators now understand their strength is in working together and sharing APIs, so they can offer cross-carrier capabilities.  Operators can federate APIs, their platforms have this capability.  We’re learning faster, now we need to execute faster.

Surprisingly the role of WebRTC (Web Real Time Communications) remains unclear.  Simplistically within 18 months any browser becomes a communications end-point.  WebRTC doesn’t magically enable browsers to communication without a server in the middle to enable the connection, it’s a communications triangle / trapezoid, like every communication service on the planet.  So operators have a great opportunity to extend their services across many more devices and use cases, but they need to pull their finger out!

mBaaS (Mobile Back End as a Service) appears to be the current hype concept.  That is a PaaS (Platform as a Service) that does everything from enabling cross platform development for client software to having lots of prebuilt capabilities such as Facebook sign-on, web and telecom APIs, user profile repository, etc.  It is an important emerging category, especially for helping channels and enterprises build apps for operators’ cloud platforms.  But its perhaps a hype cycle too far given where the market is today.

There were some discussions on analytics, but little specifics.  The devil really is in the detail on this one.  I point to O2 More as a great example of the baby steps operators can adopt in eating their own dog food on analytics.

The generally accepted business model for an API is its free at low volume, chargeable at moderate volume, and let’s cut a deal at high volume.

There is no super platform SDP, there is a services domain (across people, process and technology), it’s a service oriented architecture, within it there are a number of capabilities, most of which are shared, its federated with the IT domain, and each operator has a different implementation because of the legacy infrastructure, market situation (mobile-only, converged, fixed-only, consumer focused, enterprise-focused, mostly prepay, etc.), business objectives, etc.


Highlights from Day 1: Lots and lots and lots of operator Case Studies

There were lots of operator case studies, I cannot pick out one as the highlight, they were all important.

We kicked off with an excellent presentation by David Moro on the Telefonica UNICA SDP, it’s a global SOA using APIs internally, with partners and a subset through BlueVia to the long tail.  They have a mature API development process, they use standards were appropriate but generally define their own APIs as they can be easily changed.  They have over 50 APIs defined.  The API gateway is a critical part of their architecture in enabling flexibility.  For 2012 they focus in QoS (Quality of Service), the API is likely to be used internally for example with TU Me.

Luis Fernando Almansa Delgado described Telefonica’s Cloud Services.  A key point is there is no need to build stuff yourself when you can partner.  Telefonica are partnering with Joyent and resell Joyent’s cloud services and its leadership in node.js to make it easy for enterprises to build and migrate applications into the Telefonica cloud.  Telefonica are taking a multi-channel model, using their local presence and extensive SI (System Integrator) and VAR (Value Added Reseller) networks to drive cloud adopting in their countries of operation.  They are also bundles their services with other OTT services to meet specific customer needs.  Luis gave a very pragmatic strategy advising partnering with the technology leaders, and focus on using the local market leadership to embed your services and network into enterprises.  Highlights from Luis’s presentation are shown below.

Michel Burger form Vodafone again gave an excellent and insightful presentation, covering many of the key points listed in this weblog.  Highlights from Michel’s presentation are shown in the slides below.

Loai Garelnabi, Vice President/Digital Technology at Etisalat Group gave an insightful presentation on their Global SDP covering the services, architecture and implementation issues.

Jalel Kamoun from Tunisiana highlighted the importance of local market conditions in defining what is required in an SDP.  They implemented self-care and business analytics on a SDP, and were able to demonstrate a 50% uptake of data packages amongst customers that had previously not accepted such offers.  Given they have 98% prepaid market, the only way to drive revenues is to encourage customers to use services. Highlights from Jalel’s presentation are shown in the slides below.

Antonio Cruz from Portugal Telecom gave an excellent presentation on their services domain.  He is from SAPO, think of it like Telefonica Digital with its own development and also running the technology.  A key piece of advice he gave was to focus on using APIs internally, then with business partners, and only once that model is mature explore the B2B2C model.  Their services cover IPTV, web-based, and smartphones.  They make pragmatic use of ITIL, WS-I, TMF, OneAPI (re-use standards where appropriate but not limited to those standards) and have built most of the platform themselves.  They have a number of mature processes around the use of APIs.  Highlights from Antonio’s presentation are shown in the slides below.

Martin Prosek from O2 Czech Republic gave a frank, open and honest review of Telco Payment services.  He explained clearly the challenges operators face with fraud, and the opportunities they can create with a payment service.  Highlights from Martin’s presentation are shown in the slides below.

Highlights from Day 2: And even more operator case studies

Vincenzo Amorino and Francesco Vadala from Telecom Italia gave a quantified review of the API program.  It was the highlight of day 2 for me, as I consider Telecom Italia to be one of the most successful API programs on the planet.  They have roughly 1B transactions a month, and that does NOT include cloud services, unlike the AT&T figure, so many more of those transactions are revenue earning events.  The API transactions are roughly split 50:50 between internal and partner use, with 3rd party making up perhaps 1%.  They described their API processes and plans for process improvement, the importance of non-functional API requirements like policies, SLAs, AAA.  Highlights from Francesco’s presentation are shown in the slides below.

Hashem Sharrab, Director of Customer Support and eBusiness at Mobily gave an impressive presentation that highlighted the many APIs they are exposing including voicemail, Caller Ring Back Tone, Missed Call Notification, and In-Call Streaming.  Highlights from Sharrab’s presentation are shown in the slides below..

Sebastian Grabowski, Orange Labs focused on the role of an open middeware 2.0 across telecom and web; and need for operators to act together with a great video.  It was an Oracle session, but Sebastian made it shine.

Patrice Crutel from Bouygues Telecom gave an excellent presentation highlighting challenges operators have in innovating on the core services across the silos of IMS and NGN.  He explained how with the OpenCloud service broker enables new services to be created for a few $k not a few $million.  Highlights from Patrice’s presentation are shown in the slides below.

Some of my take-aways from the event are:

  • Flexibility in platforms is essential to support APIs, and emergence of the API gateway as a required component.
  • There is no SDP, the acronym should be dropped.  It’s a Services Domain across people, process and technology.  Just like the IT and Network domains in an operator.  See the Services Domain Report for more info.
  • Because operators buy routers and base stations they think an SDP should similarly be productized, it is not, and it should not be productized.  The variety of market situations and accelerating rate of change means the services domain is a constantl
    y evolving component of an operators business.
  • We’re now seeing emerging best practices in the services domain and the extensive re-use of off-the-shelf IT, not custom systems.  But please do not let the TMF (Tele Management Forum) get at these processes else they turn them into an overly complex mess to keep their affiliated consultants busy on projects.
  • Most of the cost in an SDP project is in the integration into existing platforms, and the fragile state of many service silos.  Not the platform itself, here Telefonica shows the steps to enable better services and IT domain integration with ESB (Enterprise Service Bus) federation and discussed in the Services Domain Report.

Coming away from the conference it felt like we can finally focus on growing the business rather than working through all the misconceptions created by the “strategic” consulting organizations.