Managing Customer Experience

In my previous blog I mentioned how we in telecoms have progressed throughout the years in determining subscriber experience. Now comes a much challenging task, how do we use all this sources to improve customer experience, network performance and ultimately the financial benefits of the network.

The first and hardest challenge here is how to support all data sources, secondly is how to correlate all data sources so they can be combined and makes sense together, thirdly is how to extract valuable information, analysis and reports out of the combined and correlated data. Lastly is how do you then use the data to improve customer experience, network performance and financial benefits.

Each and every data source there is almost a different vendors, not even the system vendors like Ericsson and Huawei have all the solutions. Outside of the system vendors there are a few companies that might have a complete solution such as Tektronix and more recently DingLi, a fast growing player in this field that have drive testing, autonomous testing, probing solutions, device testing, test handsets, mobile agents and supporting for recordings from system vendors. But how do you address the challenges presented by such numerous data formats, high volumes and continuously being created. Storage in itself pose another set of challenges.

The solution to the challenge lies into having a common platform that can support all this data sources, correlate and combine them then generate analysis and reports. Of course it should not stop there it should also be able to do automated diagnostics and generation of the most optimum solution to improve customer experience and network performance. Financial benefits to the operator comes when subscribers are happy and satisfied… they will stay with the same operator and most likely be tempted to increase their usage, especially with so many cool applications for various mobile devices….

In our own little world we have a solution called the Wireless Explorer…

Extracting Customer Experience

As the old saying goes ”there more ways than one to skin a cat”, in the world of telecoms there are more ways than one to extract customer experience. In the old days we used drive testing and propagation analysis to simulate customer experience, additionally we look at call data records and performance counters to know what they have experienced. We also analyzed protocol stacks to see where the problems that user experience lies, later we migrated this to a full probing solution. And recently we started looking at network recordings and started to look at actual usage directly from each and every subscriber mobiles.

Extracting user experience directly from the subscriber mobile is the best way to know their experience, herein lies the information not captured by any network recordings, probes or taps, the state of the subscriber handset. By knowing how, where and why the subscriber is using his or her mobile phone answer the question of what is the subscriber experiencing when using the serviced provided by the network. Knowing that allows the network operator to improve its services where it is needed and wanted, thus satisfying the customer and increasing both usage and retention.

Working with HUGE Data Volumes

With the advent of high speed mobile broadband, comes the intricate problem of handling very high volume of data to be used for network management and optimisation. It is common to have 20 TB of network or probe recordings everyday, the question is how do you process and analyze this large volumes of data to be useful? It is a known fact that even the vendors of test and measurement equipment that supplies the systems to capture and gather this data cannot cope with the volumes in terms of processing, analysis and reporting. Now considering that you add to that all other data sources that can be used for network management and optimization such as drive testing, call records, event records, performance counters, probes, provisioning, customer support and complaints… how do you handle all this to make the most out of it?

It’s simple really, we have the answer and have proven that it is possible to handle such a high volume of data easily and very fast, it is forgetting about schema’s and tables. When we design a solution we do not think of reformatting data instead we simply keep the same formats as they were in their original form, so we end up with several data formats being loaded into our systems. Why is this effective and efficient, firstly we do not need to convert and store the data in its new format. Secondly the original format where it was created is probably the best for this data since the vendor would have spent several man years to get it right for their equipment. And lastly, we simply have clever solutions...

OK the last part was pushing it too much as a sales pitch, however it is fair to say we think differently than most companies in our field and industry and we have proven that it works well.