The pattern of usage of Smartphones continues to evolve quickly. What tools and metrics should operators be tracking to understand better how and where their customers are using wireless services? How will these feed through to determine where and when to deploy new network capacity?
Network performance and planning tools can deal with activity currently on the network, but may not see or consider Wi-Fi use at all. How can these two be reconciled and expanded to determine where and when to deploy Small Cells and/or Wi-Fi?
Cellular data is only half the picture
About half the wireless data consumed on smartphones today is sent over the cellular network. The remaining half is sent via Wi-Fi, of which a small proportion (say 2%) is handled by Carrier Wi-Fi. These proportions vary by country and by consumer behaviour.
While network operators have extensive performance management and reporting tools, these only show activity from within their own network. They don't monitor activity off-net, such as when using Wi-Fi in your home or office. There may also be very limited visibility of roaming Wi-Fi other than the data volume consumed.
The only single point in the network which does see the total picture is the smartphone device itself. Several companies have developed smartphone apps which track data usage; a few actively drive traffic to/from Wi-Fi hotspots where available. These can give wider insights into the overall behaviour of end users.
Painting an unsavoury picture
They can also be a concern to operator's marketing departments. A recent survey report from Rootmetrics was slammed by Vodafone for ranking it as the worst performing UK network. They'd much rather point to the findings of Ookla (speedtest.net), which painted them in a better light.
For end consumers, such measurement Apps often bring other benefits such as data compression or Wi-Fi hotspot connection. These encourage wider use (and better data collection) although there may be a latency penalty, or little extra benefit for already heavily compressed video streams.
OpenSignal provides consumer data for operator consumption
OpenSignal was founded 4 years ago by four Oxford physicists. Their App has achieved 6 million downloads to date, of which 1 million were active in the last month. After a one time activation, the App quietly runs in the background collecting data and runs an active test every 10 days. A side benefit to end users is that it helps find their nearest Wi-Fi signal.
Their website provides freely available comparative data between mobile networks in numerous countries.
The commercial objective is to analyse the data and sell reports in a suitable format for different departments across an operator. There is no "one size fits all" here, because each team have different sets of KPIs to aim for. What helps with this approach is that the same baseline is used throughout the organisation – before, it has been common to track their own targets using different data sources.
Some interesting findings
As LTE is progressively rolled out, their recent findings include one useful metric of how much of the time each smartphone is within 4G coverage. Unsurprisingly South Korea wins that at 91%, closely followed by Sweden at 88%. North America and Japan are both around the 70% mark with European countries still at about 50%. Actual LTE data rates are also reported with many countries at or near 20Mbps – something that matches the cautious advertising claims rather than the theoretical peak.
An organisation stumbling block
Ahmed Nazish, Head of Business Development at OpenSignal, told me that sometimes the drive to meet KPIs used by different departments in large network operator organisations can be counter productive. It's common for senior management bonuses to be closely tied to meeting specific KPIs, leading to entrenched behaviour. Sometimes these KPI targets conflict between different departments, making it difficult to achieve a common purpose.
At this stage, their system doesn't track whether usage is indoors or outside. They have been constantly asked about indoor usage but today have no way to resolve that. One option is to detect when near to home Wi-Fi signal and base it on that.
In answer to my question, OpenSignal aren't claiming that this data is used to determine where to position or deploy Small Cells but their data has been used to identify potential hotspots for Wi-Fi. I would expect a different set of geo-location tools to be used to identify traffic hotspots, and metrics to be fed from there.
Summary
Capturing a full dataset of user device behaviour and using that to drive the network operator's business should align it more closely with satisfying their customers and making the right investments.
Crowdsourcing a common and wider set of KPIs based on the same data, combined with sensibly aligned targets for different departments should lead to improved decision making.
While these crowdsourcing apps provide a much clearer picture of the end user's experience, it doesn't displace other performance management and planning tools used to determine where/when to deploy new sites. Those need to take into account a range of factors invisible to the smartphone (e.g. availability and cost of new sites, total traffic across the area etc.). However, this may be another useful input into the overall assessment and resulting performance improvement.
