Putting the "service" into Service Delivery Platforms

One of the things keeping us busy right now in preparing the stimulus for the Telco 2.0 brainstorm in October is the practicalities of how operators can turn themselves into platform businesses. A platform business is one that:

  • Makes money without having to define the whole end-user value proposition: that is left to third parties.
  • Is integral and value-adding to those third party products and services, beyond just product distribution or retail.

Operators already have several successful platforms, such as i-mode and premium SMS, where third parties create the content and applications. They need to develop more. The challenge going forwards is threefold:

  • What to open up to 3rd parties, beyond billing functions?
  • How to technically integrate the core real-time functions of telephony and messaging?
  • How to make money?

A key enabling technology of a platform is the Service Delivery Platform (SDP). This is a piece of software that telephony and other real-time applications can run inside. An SDP deals with the nuts and bolts stuff of failover, scalability, and monitoring. It also provides a standardised interface to that application software to the heterogeneous network and IT world beyond. They are complex and expensive to buy and deploy, so it matters if you buy one and how you use it.

However, SDPs are positioned by vendors as a technology sale, and the business model often remains fuzzy. The sales message tends to breathlessly begin "Now you can deploy hundreds of new services in weeks, not years!" Just nobody knows what they are, or how they'll be sold or supported. (Just go tell customer care you're launching thirty new services next week, and see what happens. I advise wearing body armour.)

So in this post we review the core 'Telco 2.0' business strategies and explore some of the context for SDPs and a platform business strategy.

Examples of platform businesses

Microsoft Windows is one archetypal platform business. You buy Windows, but only because you want to run the applications it supports. Apple's iTunes is another. They don't have to do A&R work, or understand anything about the vagaries of musical tastes. The VISA network is a platform for merchant commerce.

I was told the story this week of another example that really drives home the difference between a platform and a product business: Home Depot in the USA. They retail home, car and garden equipment. Yet they don't carry the inventory risk of the goods on the shelves, or even have to think about how much weedkiller is needed in Minnesota in May. They rent shelf space to the product businesses that make those products, and charge a transaction fee on both sales and returns.

This also happens to align their interests with those of the end user. Don't think that lawn trimmer does the job? Return it! Home Depot will be delighted to capture the re-stocking fee off the mower maker. They won't resist making the customer happy. This is a fundamentally different way of making money.

For most operators, it's time to trim down the product development group (if there's anyone left), and build a new business division around platform and partner services.

Why share the customer relationship with a third party?

Traditionally operators are expected and encouraged to create end-user services like MMS or video telephony. They recognise, however, that for many reasons this vertically integrated model is dissolving away. So-called "over the top" Internet applications reach the price and performance levels needed without special help from the network.

If telcos really understood user needs, this wouldn't be a problem. They would be product innovation businesses, like Apple or Microsoft, and would have highly differentiated products. Clearly, this is not the case. It took Apple to come along and remedy the glaring usability failures of voicemail. You get your telephony service from the folk who operate the cell towers and radio network because you have to, not because you want to.

The business model therefore needs to evolve, but most network operators are clinging onto the triple/quadplay offering (fixed voice, video, Internet, mobile). This leaves them with a lot of capex and debt (to build out fibre to support me-too IPTV and high-speed Internet). You then hope your way of offering video and voice service (in a one-size-fits-all package) is exactly the right way. (Are you really, really sure that the users want the TV to pause when the phone rings?) And that same Internet connection ensures they eternally suffer low margins in their voice and video services. No wonder confidence is low in the industry's future.

Get out of jail cheap?

The answer is to horizontalise. Do fewer things, but better. That leaves five (and a half) core Telco 2.0 business strategies.

  • Product innovation. Where possible and advantageous, operators should indeed opportunistically invest in end-users products and services. This can be via internal or insourced means (e.g. Cyworld at SK Telecom). However, much of the current product development creates considerable cost without compensating revenue: the top footballers are the biggest winners from IPTV.
  • Protection of existing voice and messaging products, aligned to some product innovation to extend the franchise's lifespan. Business as usual.
  • Platform business, based on partnerships and supply chain.
  • Pipe business, based on scale and price. These may start to merge into other utility and infrastructure businesses over time.
  • Packaging business, based on customer intimacy, that assembles offerings from a variety of sources, is agnostic to technology and suppliers, and works without the network or technology capital asset base of a platform business.

The extra "half" are the various other forms of diversification into adjacent industries, such as into payments and banking, healthcare, home security, etc.

If you can't stand the heat, then sweat

When you focus, you also sweat your assets far harder. The companies to look to for inspiration are folk like Costco and Ryanair. Their revenue per square foot or per aircraft leave competitors in the dust. Ryanair in particular recognised early on that the plane seat was a loss leader for rebooking fees, baggage overage charges, food, hotels, rental cars, ground transfers, and lotto tickets. IBM mastered the "smorgasbord" approach to business in the 1990s, blending services, hardware and software.

Operators have a wide variety of under-exploited assets, for which the network is just the bait:

  • A mass of user identity, user context and usage data.
  • IT assets like expensive data-mining suites.
  • Tangibles like retail stores, which are woefully short of customer-centric innovation. How come you don't go to a network operator to buy your home network gear? How come operators are the only retailers who can track goods at the individual items level, with both forwards and reverse logistics, but can't exploit this?
  • Intangibles, like brands, and established partners and supply chains.

The network can also be worked harder: it often sits idle waiting for an unfortunate roaming mobile customer to start a data session at some outrageous price.

Every one of these is an opportunity for a partner or platform interface. That may be an API, but it could equally be other formats, such as a web site for partners to do business intelligence functions based on service usage, or a weekly spreadsheet emailed out, or a co-marketing arrangement.

Product vs. platform

The Home Depot example we gave earlier shows how product and platform businesses can look confusingly similar. That means operators need to develop a finer sense of when and why to offer product vs. partnering (via a platform approach) vs. pipe. Just because you can develop the product doesn't mean you should.

Those products that are developed should clearly

  1. extend existing successful product lines (e.g. SMS auto forwarding or reply, premium SMS), or
  2. utilise an asset or capability that only operators can technically or economically offer, when compared to over-the-top players, or
  3. diversify the business outside ( e.g. adverts, mobile payments, etc.).

Aligning investment to assets means reworking the tools and processes that gatekeep the investment and product pipeline. Projected return-on-investment alone is not a strategic criterion, as you end up with dozens of "chase the buck" activities with no synergy. In particular, operators need to understand and model the weaknesses in the Internet and "over the top" business model, and thus the opportunities to partner. Sometimes these are quite subtle.

Building a services platform, not just a portal

Most of the innovation in applications and content will come from 3rd parties. This is well known. Therefore operators need to organise around these platforms that enable outsiders to reach operators' customers. To date these initiatives (e.g. Vodafone Live!) have tended to fail to meet expectations because of being simultaneously too narrow, too ambitious, and too closed:

  • too narrow in that just a portal and a billable event don't go nearly far enough to solving the user's problems, and a far richer platform is needed;
  • too ambitious in trying to fill every possible entertainment and information need rather than focusing on a few ( e.g. navigation, messaging, music) and perfecting them;
  • and too closed in that operators, with few exceptions like DoCoMo's i-mode, have put up large bizdev and technical barriers that only the biggest partners can overcome.

The glaring omission remains the integration of core voice and messaging into the platform suite. So an SDP is a piece of technology in search of a business model; and the business model can be one of internally driven product innovation, or as an external platform approach to support third party services as well as own-brand ones.

What is a service?

To the end-user, a service is a combination of ALL of the application, content, connectivity, and device; as well as support, sales, payment, etc. Therefore a service delivery platform has to be far more than just a bunch of APIs into network elements and IT systems. It has to support the full lifecycle of the application (retail to replacement), as well as payments and remittances. It's just as important to offer a usage data mining service that matches the users to market segements as it is to offer purely technical capabilities like location dips.

In short, a business platform, not just a technology platform.

Evolving the service platform offering

The most important part of the platform is delivering the right connectivity and application together. Only a telco can do this. But the ISP access model of selling expensive unlimited Internet access plans isn't working long-term. This is because:

  • the broadband incentive problem (it's more profitable to throttle the heavy users than build new capacity),
  • a proliferation of devices means it's too expensive to users to buy the same Internet access plan over and over, and
  • user confusion at being sold an intermediate factor of production, not something of direct tangible value. (You don't need to guess the voltage, phase and usage profile of your electricity supply in advance, so why ask the same of users of broadband?)

Thus we need new business models. Indeed, taking a quick peek at our online survey results (you have done it, haven't you?), I see that over 50% of commercial respondents from fixed networks believe a new business model is needed (i.e. not an ISP model or n-play).

There will be a fragmentation of the broadband product ("slice and dice") into tiers of quality and various degrees of quantity, as we've previously documented here. We already see precedents for this (Sprint Vision, X-Series, Blackberry, Sony Mylo, etc.) The ability to route money will remain critical; putting smart boxes into networks to control sessions and applications, less so.

From gatekeeper to supplier

The operator is in the best place to do this job of managing this services layer. This is because of technology (you get better performance processing the data near the user in the network) and payments (you already have billed them for access, so why not generalise that to any billable event). Furthermore, the average service developer has zero interest in learning about complex technologies like IMS. Operators will have to expose simple, easy to use web services that hide most of the complexity. Application developers can then focus on developing applications, not nuts and bolts enabling technology. Operators need to engage partners like Microsoft, with their Connected Services Framework -- and several million developers -- to become suppliers to those developers.

The future will be one of thousands of mashup applications, with far fewer mass market horizontal applications. These are going to have to be easy to create. Even simple scenarios like the TV pausing when the phone rings are fraught with complexity: what happens if the TV is turned off after the ringing starts, or what if the phone is participating in a find-me/follow-me service? An application develop will be content to let someone else deal with that part.

Hence there is an opportunity for the "phone company" to act as a supplier to application developers and hide this complexity. The developer may use far simpler web services protocols to configure the desired behaviors, and never touches any of the session control directly. Indeed, we're already seeing this being taken to its logical conclusion, where business people can create applications, just as data mining and analytics went from requiring specialist programming and database skills to graphically-driven user interfaces.

Part of the business transformation is to create all the support infrastructure to enable this: partner education, recruitment, 24×7 support for your APIs, etc.

What's the revenue model?

The operator will make the money from charging for access to these APIs, or via the APIs driving the core business metrics of the existing product set (via lower churn, lower cost of acquisition because of a more differentiated product that sells itself, etc.) This is very operator-specific, and varies with local business conditions.

In summary

SDPs can be used to support two different business models: a product business (legacy voice and messaging aligned to new internally-driven product innovation), or a platform approach. A few operators like BT and TeliaSonera are taking the platform approach very seriously and are re-structuring to support it as a business. The critical word is "service". A true service delivery platform differentiates itself from a mere application server running a bunch of telco APIs. It does this by supporting the full life cycle of the user service -- most notably the provisioning and payment for connectivity or content delivery as part of the user proposition.

Brian Levy, the CTO of HP's CME business, gave it a better name during a panel we moderated this week - "Service Integration Platform".