Monday, September 10, 2012

[Guest Post]: 1-800 Apps Concept: Superficially Appealing but Unworkable

By Dean Bubley*, Founder, Disruptive Analysis

A number of vendors and operators have recently suggested the idea of a “freephone” model for mobile apps. The concept is that accessing specific services would not count against the user’s bundled MB quota of data, because the cost would instead be paid for by the app’s publisher or the content provider.

Various terms like “1-800 apps” and “sender-pays data” have been dreamt up to describe this model, evoking familiar ideas from other industries, such as freephone numbers or “post and packaging included”. A number of telcos have already mentioned that it is under consideration, especially in the US.
  
Disruptive Analysis view this as (largely) a mix of wishful thinking and poor conceptualisation. While it is superficially compelling and has some “elegance” at first sight, it is yet another attempt to come up with an “app-specific”, “service-aware” or “personalised” approach to mobile data pricing. Most DPI and policy vendors push such ideas for obvious reasons, even though they have seen little success in either fixed or mobile broadband domains. 
  
Like most app-specific charging propositions, “1-800 data” falls foul of many countries’ net neutrality laws – but this doesn’t really matter, as the idea has so many practical flaws anyway. A further refinement of this idea is that these apps would not only be free for the end-user, but also have specific network QoS or prioritisation applied to them – although that is even harder to get right. (Note that this is subtly different from what is called “zero-rating”, where nobody pays a bill for transport of certain data).
 
We don’t really know what an app is. The network understands it even less. And it’s getting ever more difficult to draw the boundaries for an “app experience”. 
   
Network-based DPI is far from perfect when it comes to separating apps. Although it can detect certain “app signatures” or “port numbers” – and certainly also see web addresses (URLs) – these do not map 100% to what the user perceives as an app. The classic example is a YouTube video shown inside a Facebook app – is that YouTube, or Facebook? While that particular example is now covered by some  DPIs, plenty of other variants are not. 
  
Many apps include external elements that are outside the control of the developer or content provider, and which may or may not be included in any toll-free arrangement. Many apps incorporate web components, often using the native web rendering engine in the smartphone’s OS. You can see external web pages inside Facebook and Twitter apps, and for example, the author has used an IM function, inside a Zynga game, embedded in Facebook, in a browser. 
  
Conversely, you may get a Facebook “Like” button, a “Logon with Twitter” function or third-party analytics tool embedded. This then makes it very hard to define what is the “app” which is supposed to be free – and how and where the overall “sponsored” volumes of data might be measured. 
  
It may mean that app developers have to have protracted discussions with their partners/suppliers about data consumption. It may also mean that savvy users start deliberately browsing the web inside other apps, to save themselves using their own quota.

What is an app, and who is responsible for its data use?
Source: Disruptive Analysis

For example, consider a game which features adverts between each level. These may be provided by a third-party ad network, and might vary between static, animated or even video-based content. The main app owner may well have no control over this (or even visibility) – yet will still want to know if they are liable for picking up the transport bill. Now imagine the costs and IT systems needed to resolve a support call, when a disgruntled user find that a supposedly “free” app has resulted in them paying charges for its adverts’ data. 
  
Other content may be based on a mashup (eg an embedded Google Map), or delivered from a separate source (eg a CDN such as Akamai) with its own URL. There is no easy way to stop a user using the app-embedded version of a map to navigate for free, rather than the standalone one. The “My Vodafone” app still charges end-users for mapping data needed to help them find a suitable WiFi access location for offload.
   
Now much of the idea is that certain apps will be “self-declared” by their publisher to be heavy data consumers – and therefore attractive for the user to be taken outside their bundle, and for the operator to be able to charge meaningful sums to the developer for transport. Most are probably thinking of charging heavy media download-centric apps such as movies (eg Netflix) or music (eg Spotify). They are also thinking of easy standalone use-cases such as buying a book on an Amazon Kindle, where the user doesn’t pay for data transport themselves. 
  
But even there, while the “big chunks” of data might be monolithic and relatively easy to identity, we are seeing apps evolve towards much more hybrid “experiences” – perhaps social sharing functions around a movie or TV programme, or Spotify’s integration with Facebook, to see what friends are listening to. We are also seeing “user journeys” which involve using multiple apps for a single “session” – perhaps commenting on a piece of content via a social network, or completing a transaction in a separate payment app, for example.

Complexity - Multi-app user journey but a single “experience”
Source: Disruptive Analysis

Now it is true that the policy infrastructure can work out approximations of what apps are being used in the user’s mind. And in some markets, or for some users, that might be enough. But there are enough grey areas to worry lawyers, regulators, customer-service agents and marketeers. The network – and even the operator – doesn’t really know what an “app” actually is, unless it’s one they’ve developed themselves, or worked on in very close collaboration with a partner. The pace of app evolution – and user behaviour - means that the gap is widening, not narrowing.
  
No doubt we will see continued evolution of this, especially with the ongoing improvement in functionality of device OS’s and the advent of HTML5, which will further blur the lines between app and web, and permit on-the-fly customisation of apps. All of these conspire to keep the “essence of an app” changing quickly. In future, we will likely see more interaction of apps on the device so the app-traffic that a network box sees might actually be destined for another piece of software once on the phone. 
  
Then we have related problems such as the fact that the same-named app might work very differently on different devices or OS’s. This makes the commercial arrangement between app provider and telco for “toll-free” much harder to put in place. You couldn’t offer a 1-800 freephone service with costs dependent on the type of handset callers were using. 
Furthermore, the app might even behave differently on different networks. If content is compressed/optimised, this may impact the data transport – and therefore costs – for the provider. It may be configured to work differently on 3G, 4G and WiFi.
  
These and other problems – arbitrage risks, OSS/BSS challenges and more – mean that the original idea of a 1-800 model for mobile apps’ data usage is simply unworkable in the “general case” of massmarket apps for smartphones or tablets. Where the concept might work is for in-house apps the operator themselves develops, or those developed through extremely close partnerships with external companies. 
  
Disruptive Analysis’ new report “10 Reasons Why the "toll-free" 1-800 Apps Concept Won't Work” is available from here.


__________
*Dean Bubley is the Founder of Disruptive Analysis, an independent technology industry analyst and consulting firm. An analyst with over 20 years’ experience, he specialises in wireless, mobile, and telecoms fields, with further expertise across the broader technology industry.
  
His present focus is on wireless technology, especially the evolution of mobile broadband, service provider business models, next-generation voice & messaging services, mobile device architecture & software, applications ecosystems and enterprise mobility. He provides clients with advice and analytical opinion on topics such as business models validation, technology innovation and go-to-market strategies, "addressable market sizing", planning and due diligence.
  
Mr Bubley has extensive experience in both published analytical research and bespoke consultancy, and regularly speaks at industry conferences and events, and writes the well-known Disruptive Wireless industry blog. He also works collaboratively with various partner organisations, including the Telco 2.0 Initiative and Martin Geddes Consulting. He holds a BA in Physics from Keble College, Oxford University

No comments:

Post a Comment