The importance of QoE for CSPs

In today’s highly competitive digital world, network service performance and user satisfaction are key to Communication Service Providers’ (CSP) business success and profitability. As data consumption continues to grow exponentially while revenue per bit keeps dropping, CSPs can leverage Quality of Experience (QoE) for network operations and planning to optimize their investments in infrastructure and improve their NDR (net dollar retention). CSPs should maximize their networks’ value by implementing intelligent and dynamic bandwidth management to eliminate congestion by prioritizing critical applications over those that hog bandwidth and by investing in network expansions at the locations that generate the maximum increase in their subscribers’ QoE.

CSPs must deliver optimal QoE to attract and retain subscribers, as high QoE increases their satisfaction and enhances the CSP’s brand reputation, preserving and perhaps increasing revenue as most CSPs anticipate that there are subscriber segments that will spend more for a QoE-based SLA. In addition, CSPs can execute marketing campaigns by identifying these subscribers’ usage patterns and defining additional tiered services that they can up-sell to their subscribers.

Preventing customer churn is very important to CSPs since It’s much easier for a CSP to maintain existing subscribers than to acquire new ones. And satisfied subscribers don’t abandon their CSP.

But how do we know if subscribers are happy? For this CSPs must appreciate the difference between two terms that sound very similar – QoE (Quality of Experience) and QoS (Quality of Service).

QoE captures subjective quality of experience, the metric that captures subscriber happiness. QoS parameters are easily measured network metrics, that may or may not be relevant to the application a subscriber is currently using. However, for many applications, we actually know how degraded QoS impacts QoE, allowing a CSP to gauge subscriber satisfaction, to monitor satisfaction trends, and hopefully to prevent churn.

What is application QoE?

QoE for any particular application may be derived from readily measured QoS parameters, such as bit-rate (in Mbps), Packet Loss Ratio (in percent) and Round Trip Time (in milliseconds). But each application is somewhat different. For example, The QoE of generic browsing is mostly determined by the page stabilization time (in seconds), while the QoE of video streaming is influenced by bitrate and whether there were “stalls”, and the QoE of gaming is strongly dependent on the RTT.

But all QoE is inherently end-to-end. The web page stabilization time, video streaming bandwidth, and RTT depend on the local network, but also other networks, the data center environment, the web server, and other factors, as well. Thus, accurately predicting true QoE depends not only on QoS parameters of the local network, but those of the complete round-trip path. So QoE degradation may or may not be due to factors under control of the network operator.

In order to simplify things, we can introduce a new level of metrics situated between QoS and QoE, which relates solely to the network in question. We call these network qualifiers. The following network qualifiers suffice to determine the degradation of application QoE caused by the local network:

1. Interactivity – how responsive the network is, i.e., the connection latency (based on RTT)
2. Speed – how fast the network is (based on the connection bitrate)
3. Consistency – how frequently the data-flow stalls e., line “chokes” (based on available bit-rate variability, i.e., the speed variability)
4. Reliability – how much information is lost (based on packet loss)

These network qualifiers are not simply the QoS parameters, i.e., interactivity is not equal to the internal RTT in millisecond and the speed is not the available bitrate in Mbps. Rather, they are normalized values between 0 and 1 that compare the network QoS to what is needed for demanding applications.

Each application category’s QoE degradation may be derived from the generic qualifiers, but their relevancy and importance vary according to the application categories.

For example:

• Online shooting games depend strongly on the interactivity measure, but not at all on the speed.
• Cloud gaming depends strongly on the interactivity measure but also on the speed and its consistency.
• Video streaming depends strongly on the speed, but not at all on the interactivity.
• Video conferencing depends strongly on interactivity and consistency, but only very weakly on speed.
• Browsing depends equally on interactivity and speed.

The network qualifiers enable a service provider to perform trend analysis over service types and discover potential service degradation. Since the qualifiers are for the local network, they provide actionable warnings, allowing the service provider to take appropriate action before customers discern any decline in service quality. This empowers the CSP to maintain subscriber satisfaction and prevent churn.

Moreover, for each application category, a mean opinion score (MOS) value can be estimated to indicate the current quality as well as to enable trend analysis by comparing the MOS over time. The MOS is calculated by approximation formulas for each category depending on the relevant network qualifiers to the specific application category.

However, these MOS estimates are not those of any specific session nor do they take into account service degradation due to factors beyond the service provider’s control. They are typical measures that indicate the experience that the service provider could provide were such an application to be running over the network in question, assuming everything outside the network performs perfectly.

By analyzing MOS estimates the CSPs can identify degradation in the service they are providing, optimize their network investment to maximize their subscribers’ QoE for the minimal capital investment in network expansions, while maintaining their subscriber’s satisfaction by improving their service without their subscribers being aware of the temporal service degradation, to preserve and even increase their revenues and profitability by preventing churn.

About QoE and Allot Smart solutions

Allot Smart traffic management solutions enable CSPs to apply default shaping policies for different apps and services at different times of day. We gather information, both historic and real time. When we detect degradation, we can automatically implement techniques such as prioritization of critical apps and services, buffering, expedited forwarding, limiting traffic for less sensitive apps, optimizing RAN schedulers, video optimization, service plan prioritization, and more.

To learn more, we invite you to join our webinar about The ROI of QoE where you’ll get info from the field about how CSPs can use QoE measurements to save significant money by deferring the CAPEX and OPEX of network expansion.

100 CSPs were independently surveyed about congestion management and QoE and we’ll be sharing the results during the webinar.

Be sure to save your seat!