Which KPIs should be followed to show a carrier that the network is not working properly?

Hello Robert,

Here are some major KPIS that can measure QoS for MBB service.

- HTTP download speed – higher number is faster
Measured by initiating the download of a 2GB file and downloading for 30 seconds. After 30 seconds, the connection times-out and the download speed metric is calculated by dividing the amount of data received by the 30-second period, to produce a metric of megabits per second (Mbit/s). This indicates the speed at which a network has downloaded data using HTTP.

- HTTP upload speed – higher number is faster
Measured by initiating the upload of 100 MB of data and uploading for 15 seconds. After 15 seconds the connection times-out and the upload speed metric is calculated by dividing the amount of data sent by the 15-second period. As with download speed, this produces a metric of megabits per second (Mbit/s).

- Web browsing speed – lower number is faster
Measured by loading an ETSI-standard reference page and recording the amount of time that this page takes to load. This produces a time in seconds (s).

- Latency – lower number is faster
Measured using ICMP ping. Five packets are sent to a dedicated ICMP server and returned, and the average time that these packets take to complete this round trip is recorded. This produces a metric in milliseconds (ms).

I am attaching a sample case study for MBB QoS.

Session_12_QoS_Regulatory_Framework_-_Case_Studies_R1.pdf

Hi Robert, Haider gives a complete reply regards technical KPI’s. But I do believe one needs two perspectives here, as in my experience - regardless of how ‘good’ n/w KPI’s seem - I’ve seen many customer focus groups still complaining about NW performance and giving lower than average NPS scores for this - and NW speed/reliability /etc is often evidentially cited in the top 3-4 key drivers for loyalty. Investigation then often reveals geo areas of latency etc etc, time of day latency and so on, frustrating key customers. Ive also seen situations where the NVO - sharing exactly the same NW as the carriers - scoring vastly higher overall NW CSAt scores, so any view including customer perspective has to be examined to see the key drivers for the customer segments and their types of usage. Ping me if you want more info.

In order to provide good service and support to your customers, there must be a certain level of transparency between the MVNO and the MNO. Obviuously, you could do some end to end testing to verify the reports you get from the MNO, but in principle the MNO should provide full reports on all services they are providing. Some user centric metrics may also be subject to issues in the interface between the MNO domain and the MVNO domain. Both voice services and data services may be affected by these problems. A user centric approach will normally be recommended for a end to end monitoring and reporting. You could check with https://medux.com/ if their solutions and approach fits your needs.

Guys are right - Customer Experience KPIs are the most important but at the same time not so easily available. I would suggest to check counters at the packet core level (SGSN/MME + S/P-GW or GGSN) as a quick-dirty solution.