The carrier’s PRISM moment: A wakeup call?

Wireless operators have long had some of the most valuable customer data opportunity at their fingertips, but will NSA open their eyes to it?

The recent reluctant coming out party of NSA’s PRISM and other surveillance details courtesy of Edward Snowden, highlights the potential uses and value of data, and in particular that of carrier data.

One of the early responses from the government to the PRISM revelations was that it’s all “harmless” as it was simply metadata.

So what is the government’s definition of metadata and what can be done with it?

The metadata consists of: Caller ID, called number, IMSI (the SIM card Identifier), IMEI (the individual device identifier), trunk identifier, time and duration of call. According to the order it is not limited to this, so one can assume that they are asking for a location of the device as well as similar activity for any data sessions (for context an average smartphone pings the network every few minutes).

The short path from anonymous to identified

The claim that this data is anonymous and simply not usable without the actual phone number and account information details is patently false.

For data to be re-anonymized and converted into personally identifiable information, one simply needs access to other data sources that can complement and corroborate some of this data and suddenly the data identifies the individual to whom a single item referenced. The low threshold of the corroborative data required was highlighted in this study performed by researchers at MIT.

This was highlighted by the Netflix million dollar challenge which was initiated for research purposes and was highly regarded, yet by researchers at the University of Texas simply cross referencing it with data freely available from IMDB’s database the researchers were able to accurately re-identify the identity of users in the sample data.

The carrier data is far more insidious as there is simply no better tracker than a cell phone, as the cell phone has been shown to be bound to us even more than our key and wallet. What’s more; the cell phone is both an active tracker for all the activity we manually perform like application usage and phone calls, and a passive tracker that constantly interacts with the wireless network to identify our locations and activities in real-time.

The scary capabilities

Now if researchers were able to do this with freely available data, imagine what the NSA and other intelligence agencies can do with their proprietary data to de-anonymize the carrier metadata with extreme accuracy.

To get a grasp on what the NSA can do with the data once they are de-anonymized, one needs to simply look at this study conducted with similar carrier data by researchers at the University of Birmingham which I referenced to in an earlier post. The researchers were able to “predict” with 94% accuracy where an individual was likely to be in the next 24 hours within 20 meters of accuracy, and further research by Microsoft allowed fairly accurate predictions 285 days in the future.

The goldmine for carriers

To date the carriers have dipped their toes into the possibilities enabled by the rich data opportunity with programs such as Verizon’s Precision Market Insights or Sprint’s Pinsight Media and now AT&T, or abroad in the UK with the cross carrier Weve joint venture that provide agencies with anonymous user data to improve targeting. However the real opportunity is for carriers to create their own businesses utilizing their data to provide rich and contextual customer experiences and application that leverage these types of analytics.

While the NSA does not require any opt-in or privacy mandates, any commercial application will require these types of stringent controls.

Yet the NSA’s insight into the value of carrier customer data – while distasteful, can hopefully be a catalyst for the carrier community to recognize the opportunities presented by that “collateral” side of their business and take its possibilities seriously, while proceeding with thoughtful caution.


The Programmable Mobile Network Cloud

In the ongoing discussion about how carriers can keep up with their customers’ speed and coverage requirements, a larger question is being neglected: How do they pay for all of the new technologies and capabilities required to address those concerns?

Operator pricing reflects the cost structure of core networks, including 50,000-75,000 cell sites for a national carrier in the US and lots of spectrum. This model served the carriers well through the migration to 2.5G and even through the early stages of 3G. Yet, with 3G, the new generation of data-hungry smartphones started showing their impact, best demonstrated by some of the early issues for AT&T’s network after the iPhone’s launch in 2007.

A realization started taking hold that a new way of deploying networks will be needed to keep up with the realities of wireless growth and use. The issue became more about capabilities than coverage, and the need to be closer to the user and create a denser network became obvious. That closeness improves the data capacity and throughput in heavily used areas.

The new capability is a merger of various technologies that fall under the small-cell moniker. These small cells are generally miniature versions of cell sites, but in size and deployment, they look more like WiFi hotspots. The reduced footprint allows them to be positioned in plain sight but obscurely in areas previously unsuited for a full cell site. This opens up the possibility of bringing the network close to the users, such as in a store or on streetlight posts. This new style of network is commonly referred to as the heterogeneous network (HetNet).

Unfortunately, this new style of network adds a tremendous burden to the capex and opex of the carriers. The cost and complexity of deploying these small cells is significant and will only add to cost of the network. However, these new costs (the cost of running today’s wireless networks) are not reflected in any changes in charges to the customer. This in effect means that the carriers are subsidizing improvements in service with little reward, aside from the all-important customer satisfaction requirement.

This has led to acrimonious talk about OTT applications. The carriers feel they should be compensated for creating a network that allows these applications in the first place. One can argue about the overall merits of the carriers’ position, yet the reality remains that these OTT applications create the need for a costlier network, which the carriers provide at great cost that they must eat.

The question, therefore, is how carriers are supposed to benefit from the ongoing and increasing investments in their networks. The answer lies in those very same investments.

Making the network work for you
When one looks at the 3GPP standards that are the foundation of the wireless network infrastructure and service, one sees a network aimed at providing reliable wireless services. On closer inspection, the standards are merely the sum of their parts, which contain a host of minute capabilities that allow the wireless service and were designed with that primary goal in mind. Yet those minute capabilities can be reconfigured and used for purposes other than wireless service.

Take this research from Microsoft that aims to gather real-time information on a given location. The current setup requires a user to check in at the location and provide the required information. Yet the optimal way to do it is by leveraging the registration requirement that allows the network to know where the user is upon entering the facility and using the control plane to activate the device and gather the required information.

Vodafone is using that registration information in Greece to provide free service in participating retailers. The control plane device activation is currently used primarily for location discovery in the event of an emergency call.

Carriers already mix and match some of these capabilities as needed for various M2M applications, yet these implementations are extremely costly and time consuming, so they are limited to carrier-sanctioned products. It is incumbent that we turn the core of the carriers into development environments — with the obvious stringent privacy, security, and veto controls — by simplifying and exposing the network to outside developers.

OTT companies and app developers will always look to take the quickest, most efficient route to market. We can all agree that is currently not with the carriers if they have a choice in the matter. Yet developers and their investors will recognize the value of unique and game-changing capabilities, and they’ll be more than willing to deal with the reasonably surmountable yet clearly defined roadblocks to partake in the opportunity.

By opening up these core capabilities and making them easy to use as individual modules, carriers remove the need to limit which services get deployed based on resources. This will open up profit avenues and scenarios that leverage the new and unique underlying capabilities of HetNets, including new business models.

For carriers, recognizing that they do not have all the answers, and that they cannot exploit every opportunity, is the first step in preventing their networks from morphing into an unsustainable dumb pipe.

This blog dated 4/17/2013 was originally edited by Sarah Reedy from Light Reading and published on the now defunct community website.

Mining the Last Mile for Data Gold

Wireless operators have long had valuable customer data at their fingertips, but they are just now starting to explore what that data can do for them and their customers.

The new way of thinking is based on the fact that the more an end customer uses services on an operator’s networks, the more data points the operator can collect and use for a new breed of services that can be deployed flexibly and monetized rapidly.

recent University of Birmingham study hosted by Nokia Corp. (NYSE: NOK) highlighted the value of some of the ever-increasing amounts of data accumulated during the regular use of a mobile phone. By using sample user data from across a wireless network, researchers were able to predict where a user would be heading over the next 24 hours to within 20 meters. This data included ordinary location updates and call logs, which were layered and analyzed with similar data from the relationships that were established by the call log.

Imagine the opportunities if all this data was used as a real-time platform for creating services. The types of services this data can enable — everything from law enforcement apps to location-based advertising and payment apps — are truly groundbreaking and can make today’s average mobile app appear frivolous.

Take a simple example from day-to-day life: Your spouse asks you to pick up your child, who is staying over at a friend’s house. Instead of running the risk of forgetting, you could set up a reminder based on a calculation of the time it takes to get there — live traffic is also carrier data, after all. Your spouse could also see or be preemptively notified by a choice of voice, SMS, MMS, or push notifications when you will be arriving with updates in real-time. Using an operator-powered app, your spouse could set this up the night before simply by adding the notice to your calendar app that supports this functionality.

If the app environment has showed us anything, it’s that the mobile phone is an always-present companion. If you allow services to be created uniquely for it, the result is a highly successful and profitable app store for its enabler, whether that be Apple Inc. (Nasdaq: AAPL) or Google(Nasdaq: GOOG) or the wireless operators.

The weakness of the current phone-based app environment is the need to program individual applications for specific operating systems and the need for the application to be activated for use. If a developer wants to create a passive application that runs in the background without requiring the user to launch it or provide input, it burdens the battery and the network.

The carriers face no such burdens. Wireless network data is already constantly monitoring the mobile devices. If the data were used in real-time, it could power a new generation of easily created, cloud-based, contextually aware applications.

As luck would have it, the technologies and resources to deploy these types of services are already in widespread use in various divisions of the carriers and their landline siblings. For example, AT&T Inc. (NYSE: T) has AT&T Interactive, which is all about ad targeting, and AT&T Labs, where analytics are a core competency.

Roadblocks to data divinity

However, more can be done. One thing holding back innovation around contextual apps is the wireless operators’ focus on optimizing the radio requirements, rather than putting them to work in new ways. If they factored data collection and storage into the network’s design, the new data they would collect would be even more valuable and useful.

The second factor stopping carriers from allowing this to happen is a reluctance to share this data, due to privacy and regulatory concerns. But a tight and granular set of opt-in controls can actually improve the privacy factor significantly over today’s extremely limited uses.

The opportunities for carrier apps that use their last-mile knowledge are immense. Instead of sitting on the sidelines and viewing mobile devices solely as a burden to be managed, carriers ought to join the fray and view the devices as an opportunity to create an extremely positive and profitable relationship with their customers.

This blog dated 11/7/2012 was originally edited by Sarah Reedy from Light Reading and published on the now defunct community website.