« GSM in a Suitcase = Double Disruption | Main | TomTom shows Google the way »

Next Generation Mobile Networks Alliance - CeBIT

NGMN.org ran a half-day conference at CeBIT on mobile broadband. The presentations - from Vodafone, LG, Nokia Siemens, and Texas Instruments - can be downloaded here…but Hamid Akhavan’s, CEO at T-Mobile International, seems to have been withdrawn from the site now. We managed to grab it before it was. The key image is below. It shows the economic unsustainability of mobile broadband, especially on flat-rate tariffs. If you understand that low quality YouTube videos now account for 10% of all global web traffic, then imagine what will happen when the quality improves. In fact you don’t need to imagine: see the real stats of the impact of the BBC’s iPlayer (high quality streaming video) on UK ISP’s in the last 8 weeks since launch (a doubling of streaming traffic and a trebling of costs - analysis here). Then you have to ask: “Don’t we need a new business model here, in parallel with the 4G technical developments?” The answer is of course, yes, and we explained why to NGMN.org in detail over a year ago on this blog (here). But, of course, we’ve had a deafening silence from the tecchies about this (“Not our problem!”). Hence the 4th Telco 2.0 Executive Brainstorm in April to bring 200+ tecchies and commercial people together to look at this in a structured way.

To share this article easily, please click:


It is unfortunate that the guys in Telco 2.0 do not have a good memory. The point made above, decoupling of value and revenue, was one of the key reasons NGMN alliance was founded. This realisation is based on some work that was done with MIT long before Telco 2.0 was even founded. The other reason for establishing NGMN was realisation of personal broadband, another project with MIT that was well before Telco 2.0 was established. Also, I gave a talk in Telco 2.0 last year on this topic, so it is surprising that the author calims that there was silence from NGMN :-)

Now if you want to be picky, the real problem is how to decouple cost from traffic. The alternative suggested here, i.e., looking for new business models, does not address the fundamental issue of how to ensure margins when revenue and traffic (and therefore cost which is proportional to traffic) are decoupled while cost and traffic remain coupled.
New business models will not solve this problem, they may, however, temporairly mask it.


Hossein Moiin
Group VP Technical Strategy
T-Mobile International
Chief Editor, NGMN White Paper

I think we're probably in violent agreement, but the problem may be in the term 'business model', which can mean different things to different people. For us, it's simply "how money flows between different parties in the value chain".

So almost by definition, the answer to the problem in the post above has to be a 'new business model', since no single or combination of technical measures can put the vertically integrated model (that couples cost, revenue and traffic) back together.

Such new business models will need to re-combine some of the traffic with the sale of the associated content, service or device. We will also need to find new technology models that allow "scavenger" traffic like P2P to co-exist with higher priority traffic, but to only occupy the slack, and not to contend with that priority traffic or to raise peak demand (and thus capex).

The airline industry has managed this with its yield management systems, but the telco industry has barely begun the process. (Most planned NGN networks try to work like fleets of private jets -- costly and not a scaleable solution). The 'deafening silence' we refer to comes from the vendor and technology community in general, not NGMN.org specifically (apologies for unintended slur, Hossein!)

Here is a high level description of the 'new business model' we propose: http://www.telco2.net/event/april2008/slides_two-sided_business-models.php.

All the detailed analysis and modelling that's gone on behind it builds on the good work of the MIT programme (which was itself the starting point for the Telco 2.0 Initiative back in May 06). All our research and consulting since then (inc. our event series where we've tested these ideas) has focused on looking at how money could flow differently and scenarios that bring this to life.

We think the 'two-sided platform-based telecoms business model' now provides a clearer direction for the industry. And, the senior Exec groups we've presented it to so far seem to agree.

It's also a major reason for running the April Telco 2.0 event - to try to build a common framework for a.) understanding the opportunity as well as the problem and b.) working out what - holistically - to do about it (in terms that different people with different perspectives and interests can latch on to). Details here: www.telco2.net/event/april2008.


I think both you and Hossein are making valuable statements.

The traffic and revenue challange is of actually a double-headed problem. There is the issue of re-aligning revenues with traffic. The other issues is the proportional relation between traffic and costs.
Margins are derived from revenues subtracted by costs. Therefore, Hossein's point that to ensure margins new business models in itself do not completely solve the issue.

The 2-sided platform-based business model creates opportunities to derive revenues from other sources than from subscriptions (or from pre-pay). Recombining some of the traffic with the sale conatent, services and devices may allow for additional revenues from end-users of other parties in the vealue chain.

You refered to the airline industry. It provides indeed some interestiing 'lessons'.
Low fare airlines make their money in different ways than the incumbents. Low fare airlines charge for additional services that are included by the incumbents. But in addition low fare airlines run lean operations allowing for big cuts in operational costs compared to the incumbents. Yield management in fact helps to ensure margins by covering marginal costs. Hence for low fare airline it is the combination of 'product packaging' (offering low fares and charging for add-ons) with low cost operations and high asset utilization that make up its business model.

So in new business models, incl. 2-sided platform based ones, costs need to be lowered as well.

An example of how costs could be lowered for traffic is to use real estate financing models. Of course (other) operational costs need to be reduced as well as proportional to the traffic.

So solving the
'economic unsustainability of mobile broadband, especially on flat-rate tariffs' is not just 'new business models' or just 'decoupling of traffic and costs'. It nedds to be both. But it is actuaally even more challenging.

New business and new business models have a higher risk and higher uncertainities about economic returns. Some of the tools typically used for financial analysis, and the decision making about investments, distort value, importance and likelihood of success of investments in innovation. The combinations of risk, uncertainity and the tools make it very difficult to understand new business and new business models. Hnece, almost by nature, decision makers give preference to cost reductions and innovations building on existing products and markets.

Your events (and blog!) can help to raise the awareness to innovate along the lines of new business models, esp. the 2-sided platform-based business model. (the agreement of senior Exec groups you've presented it to so far prove that point)And also to eductae how to create 2-sided buisness models.

Hossein kindly responded too, as follows: "I think we are talking about two different approaches to solving the same problem. The problem, as I see it is the "network business case". This can be described as follows:

The network business case hinges on the fact that the revenues from the network must exceed its costs. The cost of a network (for a packet switched network) can be roughly expressed as a function of volume of the traffic that it carries or (# of bits carried by the network) * (cost per bit) where cost per bit includes both CAPEX (including spectrum) and OPEX.

The revenue of the network depends on the business model; two business models have been used; a metering business model where the revenue is a function of total traffic (MBs, messages or minutes) and a subscription based model, where revenue is a function of the number of subscribers and is independent of the traffic that they generate. The subscription based model is commonly referred to as flat-rate.

When voice and messaging were the dominant applications of the mobile (or fixed) networks, the principle business model was metering; which allowed the operators to tune their network and make a positive business case for the network. However, as data becomes the dominant application on the network, the value of data decouples from its volume. This means that the business model needs to migrate towards a subscription based model. This in turn implies that the cost and revenue are now decoupled. Two solutions exist; first, make revenue somehow coupled with the traffic (an example is the two-sided business model or advertising business model) or make cost independent of traffic and instead dependent on the # of subscribers. In reality, both models will be tried. Technology can help with making cost dependent on the # of subscribers and business models can help with revenue being tied again to the total volume of traffic.

NGMN approaches the network business model from both angles where initial effort was on the cost side; hence, the emphasis on technology and the later activities concentrate on the revenue side.".

"However, as data becomes the dominant application on the network, the value of data decouples from its volume. This means that the business model needs to migrate towards a subscription based model."

Is this really true?

I agree there are different types of data with different types of value

eg a favourite movie is more valuable than some advertising we are forced to watch to see the movie

or with text based stuff

a search result which answers a particular perplexing question is probably more valuable than the daily browse of someone's favourite site.

But was this not always the case with voice?

Most people view a call to a voicemail box being a particularly waste of both time and money?
A five minute chat contemplating life's rich tapestry with our kids is priceless.


A text message asking for product details from an offline advertising medium is more valuable to the recipient than the sender.

We have developed pricing scenario's to deal with these eventualities in the voice & text world - why can't we develop similar solutions in the data world?

The elegance of the 2-sided model is that we can apply value to not only volume but also CONTEXT - this is the combination that we require to extract the maximum return.

The other problem I have is the cost element of the equation - it is not simply true that cost is purely proportional to volume. This is too much of a basic equation. Although I admit I am also guilty of making this generalisation periodically.

In reality it is proportional to three main variables:
a) volume - for sure;
b) capacity - think of spectrum auction, lighting up new fibres, buying new centrals. Very lumpy cash outflows; and
c) subscribers - think of both sales and support effort.

So yes, we all like the abstract to illustrate our points, but the reality is more complex.

And as with all business models we need to add both complexity and fine tuning to make them more representative of the real world.

Agreed the pricing issue is complex, but I would challenge the view that we have solved the problem for voice.

We've known for over a decade that our aim should be for transparent, value-based pricing, especially in consumer markets.

But value is as much about perception as reality.

We killed this for voice with otherwise lauded pricing policies such as "friends and family", which perversely operate on an inverse price-to-value equation. We made consumers expect to get their most valued calls for less.

In some markets, such as western Europe, through the success and ubiquity of mobile telephony we've also allowed the so-called "mobility premium" to become eroded. There are perhaps still opportunities to resurrect this through additional "context" awareness, but the challenge remains to demonstrate consumer value in these new approaches.

Similarly for data access, we've had techniques for selling "coloured bytes" for years, but on closed networks we have been fighting the consumer expectation that the network should allow them to connect and communicate with anybody, whilst on open networks the difficulty has been overcoming the consumer perception of the Internet as a "free" medium.

Multi-sided business models allow operators more flexibility in matching revenues and costs, but fundamental principles still apply on each side.

As Keith points out cost of traffic is not purely proportional to volume, as it is proportional to three main variables a) volume, b) capacity incl. spectrum licensing, new fibres, new centrals and,
c) subscribers.

However, in this context it is only *incremental* costs we need to look at. Capacity incl 3G spectrum are sunk costs and can be excluded if no new spectrum etc is needed. And subscribers translate to volume of traffic. Therefore, although somewhat simplistic, costs of traffic are proportional to volume. (especially when considering marginal costs)
In the end it is all about the positive cash flow.

Hence. I believe cost and pricing need to be geared to lower price per unit of performance. This could be actually higher prices for users in a particular context (indeed the inverse of 'friends and family' plans). It is exactly the *lower cost per unit of performance* paradigm that leads me to conclude that we need to both lower the cost per unit of performance (e.g. volume) and create new business models (including 2-sided platforms) to increase the performance.
This may require different, disruptive, technology.

Post a comment

(To prevent spam, all comments need to be approved by the Telco 2.0 team before appearing. Thanks for waiting.)

Telco 2.0 Strategy Report Out Now: Telco Strategy in the Cloud

Subscribe to this blog

To get blog posts delivered to your inbox, enter your email address:

How we respect your privacy

Subscribe via RSS

Telco 2.0™ Email Newsletter

The free Telco 2.0™ newsletter is published every second week. To subscribe, enter your email address:

Telco 2.0™ is produced by: