Last week Dutch Telecom giant KPN held its shareholder meeting at the Mayfair Hotel in London. All four hours of the presentation are viewable online here along with the slides used. Those who watch can see a startling revelation about the use of deep packet inspection (DPI) by a mobile operator.
The presentation documented the rise of smartphones and the growing replacement of voice services by data. This is a well-known story in the mobile world, to be sure. But the KPN presentation was unusually explicit about how this was impacting their revenues – and how they planned to respond. Company spokespersons noted that KPN's “Hi” service, which is targeted at younger people, had seen an increase in the penetration of the instant messaging application WhatsApp from 0% in August 2010 to an eye-popping 85% of their users in April 2011. The growth of WhatsApp was correlated with an 8% decline in their highly profitable SMS (text-messaging) traffic. Moreover, applications like WhatsApp were frequently used as a substitute for the voice services offered by KPN mobile, leading to fewer users going over their usage bundles, and thus a decline in associated revenues charged to customers for exceeding their limits.
KPN spokesperson Marco Visser brought the point home by comparing the bill of a customer who moved from an early smartphone (Nokia N95) to a more recent Android-based smartphone: in one year, voice usage went from 200 minutes to 175 minutes, SMS messages went down from 200 to 160, and data usage went from 37 Mb to 200Mb. Her average total bill thus dropped from EU 62.45 to EU 48.95.
This is a great example of how Internet-based data communications, with its plethora of user-selected applications and services, creates a consumer surplus; i.e. a more efficient and productive use of their time and money. But of course, to the telephone companies it represents a major disruption of expected revenue streams.
Knowing full well that such substitution will intensify over time, KPN's Visser then outlined a fairly systematic strategy for responding to it. In the short term, they would engage in “damage control” by upselling customers to “lock them in” to higher-priced bundles. Beginning in the summer of this year, however, they would no longer sell voice and data as separate bundles; they would instead bundle them together, thereby reducing the ability of consumers to trade one off against the other. They also plan to reduce the size of data bundles at the lower prices levels, and try to participate in the app market themselves, using their mobile service as a platform to compete with the Apples and Androids. (Good luck with that!)
Most controversially, Visser said that KPN would start charging for mobile voice over IP capabilities. Noting that KPN had installed the Deep Packet Inspection capability to detect whenever their customers were using specific applications, he said bluntly, “we will not block new services or applications; we will monetize them.” So it sounds as if customers who use VoIP and possibly also streaming video will be charged as if they were using minutes on KPN's network, regardless of whether that VoIP service is KPN's or someone else's. This is preferable to blocking, of course, and this information contradicts some reports which indicated that KPN was interfering with WhatsApp. But a decision to charge customers for using applications offered by third parties, without the end user specifically contracting with KPN for that service and without the app provider contracting with KPN either.
Mr. Visser seemed blissfully unaware of any regulatory implications. But of course, once this was publicized (by WebWereld and others) the Minister of Economic Affairs promised to start an investigation, and one can only imagine how the impacted service providers and app developers, as well as its consumers, will respond.
In general, this illustrates a disturbing tendency of the phone companies to attempt to capture for themselves the consumer surplus generated by the mobile Internet., by using vertical leverage of its network capability. Granted that the growing use of applications requires more investment in bandwidth, and that this bandwidth must be paid for, it seems that KPN should concentrate on competitive and sustainable pricing of bandwidth and not on discriminating among applications or on discouraging particular kinds of user substitutions.
It remains to be seen whether regulation or competition is the best response to this problem. The mobile market in the Netherlands seems to be reasonably competitive. KPN may be flat wrong in their assumption that they can monetize the use of applications that are not theirs and thereby drastically limit the utility of users' smartphones without suffering severe market share losses. KPN's mentality as a corporation is still not fully extracted from the old days of PTT monopoly, according to insiders. On the other hand, if they are able to implement such tariff changes with impunity – i.e., if other mobile operators follow suit and imitate their tactic – it is very strong evidence that their market is not competitive.
In the meantime, KPN joins the casebook of our DPI and Internet governance research project and we will watch its development and outcome with interest.