For a European perspective of how mobile operators are actually running their networks, their business drivers and technical focus, you can't beat a small, niche event like this. At Mobile Network Performance Management in London, we heard of a growing shift towards the customer experience (and their perception of it) and away from simply delivering ever higher peak speeds.
Relatively small, niche events such as this one allow you to network and discuss more directly with other delegates, contrasting with the hype of some of the larger conferences. Delegates are often less senior, more approachable and open to discuss issues frankly, resulting in a better measure of what's really happening and of concern within network operations departments. Participation was European including from the Middle East.
David Chambers of ThinkSmallCell chaired the event and (at short notice) presented the keynote address. Having participated in this event over several years, it's clear there has been a continuing shift from purely resource-based analytics across to end-to-end service performance. Capacity planning has evolved to include performance assessment.
Most European network planning departments don't even consider Wi-Fi
Network performance reports are no longer restricted only to operators but are now in the hands of the consumer, enabled by RootMetrics, OpenSignal and others who freely publish detailed and comprehensive analyses. This has changed the behaviour and priorities of the entire business, from marketing departments through to operations.
What surprised me most is that (in general) operators still weren't really considering or monitoring customer Wi-Fi usage, whether private or public. They expect to have to take that into account in future, but it's still early days. It was still termed "offload" rather than being an integral part of their offering. While most had heard of Google Fi, that business model wasn't on their radar as an imminent threat.
Shifting from measuring QoS to Quality of Experience
Despite many industry headlines shouting about ever higher peak data rates, most delegates felt that 4G speeds often exceed what makes a noticeable difference to the smartphone experience. You can only browse the web at a certain pace. Customers are looking for a more reliable, consistent service that "just works anywhere" and so planners are more focussed on ensuring minimum speeds are met for everyone. Network reliability is the key factor of customer churn. An Ericsson report was quoted from 2014, stating that 52% of users are dissatisfied with their mobile service. The question is how to improve that perception while staying within the budget it justifies.
Operators haven't been able to charge a premium for the higher speed and performance of 4G. It has led instead into higher usage (in Gbytes/month), which results in higher revenue. But that isn't proportional – many users didn't use much of their data allowance anyway and so don't move up to a more expensive tier; high volume users would take out a better value package that has a lower price per GByte. One case study indicated that doubling traffic might raise revenues by only 30%.
One multi-national operator told me they wanted to refine their choice of which KPIs most affected customer perception and focus on those. Few customers can tell what their actual speed is (or even how much cellular data they consume per month). Many more will notice the dropped call, slow or unusable data session or network outage.
Drive testing with Apps
Widespread use of smartphone apps for performance monitoring continues to grow. This isn't just Rootmetrics or similar standalone Apps. Functionality can be embedded in paid content Apps, so that if you buy a film or sports TV replay but it can't be shown, then the reasons for that (e.g. network too slow, film server offline) can be determined and explained to the customer. Results are more accurate than can be determined purely within the network using probes and system reports alone.
Another concern is the inceasing us of data encryption (HTTPS) which means that operators are losing sight of the services being used. This may make it difficult to prioritise and shape traffic flows in the future. Apps on the smartphone would have visibility of the type and purpose of each data session.
Drive testing, which used to involve specialist engineers driving around in vans bristling with antenna and technical equipment is still alive and kicking. But in a different form. Case showed their small cube shaped standalone network tester (smaller than a mains power adaptor) which can make and report test calls. Standard hardware reduces the price to about $250. Hundreds have been installed in taxis in several cities. The taxi driver benefits from small payment, or it can even provide a free in-taxi Wi-Fi service for their customers on the side.
Correlating usage with revenue
There's still a view that a relatively small percentage of users are consuming a substantial portion of overall network capacity. Whether that's 5% consuming 30% or a different mix, there's no doubt it's still unbalanced. That seems to be more of the case in networks where unlimited uncapped tariff plans remain in use.
Three UK revealed that they still carry more data traffic than any other UK network, (nationally 45% of total data from 10% of total subscribers). EE is catching them up, but still some way behind.
Du (the 2nd operator in Dubai, grew to 45% market share in 10 years) explained how they had correlated statistics for the top 50,000 customers by revenue and the top 50,000 by dropped calls. Perhaps surprisingly, there's not much overlap. They are drilling down into the underlying technical reasons and dealing with them. This includes are actively trialling small cells (for example at Emirates corporate HQ) to address specific issues such as in the lift shafts and car park areas. This allows them to redress any competitive advantage from other operators who have more spectrum.
It was a surprise to me that the peak busy hour(s) for data usage in several European networks seem to be between 22:00 and 24:00 daily.
The round table discussion on 5G was relatively brief. All thought the requirements were unclear and timescales unrealistic. Few felt that super fast speeds were what the industry needed today. Some cynics thought it was more a way for the telecom industry to sell upgrades and more equipment. A timeframe of 5-10 years was thought realistic, by which time much will have changed.
Few operators seemed to be looking forward to deploying VoLTE (Voice over LTE). UK network EE has done so, but it was technically difficult and involved. Giovanni Romano, Telecom Italia and also 3GPP RAN vice chair, said it is being rolled out slowly worldwide but thought it will eventually become more widespread. It doesn't uniquely offer a major benefit to the user. For example, while there are 132 networks offering HD Voice, only 16 are using VoLTE and the rest use 3G. Many are concerned it would only lead to dissatisfaction where problems arise. The primary benefit is to the network operator through increased spectral efficiency of 4G compared with 3G.
LTE Broadcast potentially offers an efficient way to deliver multicast video, but you need to allocate large parts of capacity resources which might be wasted. There is no commercial service using it yet.
Most networks now operate 2G, 3G and 4G technologies in parallel. Telenor have just announced they will shutdown 3G first (in 2020) and 2G by 2025. However, many networks have a strong legacy of M2M (machine to machine) devices which rely on 2G and you can't easily switch those off. The general thinking was that 3G would sunset first, but would need excellent LTE coverage to be in place. Although VoLTE can fall back voice calls to 2G when outside LTE coverage, it's not very good – data sessions drop and it can take 10 seconds or more to make the handover.
Telefonica gave a technical exposition on CoMP (Co-ordinated Multi-path) as a technology which could deliver very high capacity for a few targetted areas. The first step might be to use this for the uplink, which would only impact the network equipment. Downlink would need replacement hardware in the smartphones. It seemed to me that today's planners didn't have a great appetite for either in the short term, but want to understand its potential for the future.
Business realities are moving network operations away from high speed hype and relentless cost cutting, instead focussing more on the customer experience perception. It's less about peak performance and more about consistent, reliable and adequate service delivery.
Yet network operations teams remain inwardly focussed on 2G, 3G and 4G cellular service delivered primarily from outdoor macrocells. Increased tools to measure actual performance, including directly from the smartphone, give then much better insights into the quality of experience being delivered.
Some present were trialling or actively deploying in-building solutions, including small cells and even Wi-Fi, but these weren't considered mainstream yet. It seems to me that it is this more mature business approach to customer satisfaction and value that will drive adoption of disruptive technologies and business models, including Small Cells and quality Wi-Fi.