• Andreas Jörbeck

Customers measuring QoS – Salvation for mobile operators?

Ever since CarrierIQ got revealed as “spy-ware” running unknowingly on hundreds of thousands of mobile handset, the idea of measuring QoS on the handset has been a hot topic. Today there are many solutions on the market, some who even claim that this is the only way to measure true E2E quality and customer experience.

Is this the solution to replace probes and other OSS systems for measuring quality of service; and can mobile probes really feed customer service representatives with enough information to help an operator’s whole customer base with their trouble tickets? My response is that they can’t and that it will never be a replacement for other BSS/OSS-systems. Nevertheless it can be a good complement.

Let’s take a closer look of what can be achieved with “mobile based probes”

1) It will always, at least on paper, be easy to argue that the only way to capture true user experience is to measure it on the handset However there’s a flip-side to that coin, all mobile terminating traffic, such as mobile terminating calls, SMS:s and data push notifications, are initiated from the network side with a paging. If the handset has no coverage these pagings will fail and all that can be logged on the phone side is missing coverage, not how many calls and SMS:s that failed to be delivered during this period of no coverage. Additional core network problems can occur that a phone probe will not cover either.

2) Cheap to deploy since you use the device’s computing power to do your logging Save for licensing cost which is hard for me to estimate, there will definitely be provisioning cost for the operator as well as battery power cost for the subscribers. CPU power and battery life seem to go hand in hand. You barely have enough power for one day’s normal usage of a smartphone, so if the software drains more than 5-10%, of the battery per day I believe it will be a hard sell to the consumer.

3) Penetration of smartphones is ever increasing, so in theory you should be able to reach full penetration over time In a short term perspective penetration will definitely be a problem if you intend to try to support all your customers, or do meaningful corporate reporting. You will simply not get all data. In my view the “mobile probes” will give you a sample of what kind of service you deliver to your customers, which can be used to try to tweak your network to improve your service delivery. This can be a very valuable complement, and it could save you some drive tests etc. However, if you want to cover all your subscribers with QoS metrics you need other systems.

4) You can log things that aren’t possible from the network side – like battery status, installed apps and GPS based coordinates How much do you have access to for logging purposes? How does this vary between different models and OS versions? In essence how many useful and platform independent KPIs will you be able to derive? In a multi-vendor network it is often hard to get comparable KPIs between vendors like Ericsson & Huawei, just imagine the challenge for hundreds of less documented handset models.

5) Customer profiling, based on more parameters that can be captured in the network Integrity could be a show stopper for penetration altogether. For the noble cause – improved service quality, it might be acceptable, but there might be legal issues in some countries around having software logging GPS positions, DPI and other user information. In addition, customers might not accept having such software running on their devices.

6) In the larger picture, mobile probes can be an excellent way of bench-marking different operators There are already several speed-testing and coverage measuring apps available for consumers. Crowd sourcing these metrics can both be a valuable source for consumers comparing quality between mobile service providers and a great new source for an operator to get an additional quality metrics for internal use.

In my opinion mobile based probes can and will be an additional valuable asset when tweaking your mobile network to deliver the best possible quality. But, it will not replace probes or other OSS system for a very long time. These are still the best sources of information when you want to cover your whole subscriber base with CEM solutions, for customer service, service assurance, predictive churn analysis or capturing profiling data for one-to-one marketing.