Alongside paid industry analyst reports, there has been an increasing number of freely available cellular industry activity reports with five year forecasts published by major vendors and industry associations. There's a mixture of detailed statistics, differing interpretations and somewhat biassed perspectives. We offer a guided tour highlighting key attributes.
Cisco’s VNI (Visual Network Index) has been the benchmark for 12 years. It provides substantial depth and detail including many statistics for the Internet overall, both fixed and wireless. You can filter the data model to suit your own purposes with a sophisticated tool
Published annually in February with data for complete calendar years.
- LTE accounts for 69% of mobile traffic
- Mobile offload to Wi-Fi 60% (down from around 80% a few years ago)
- Average mobile connection speed tripled to 6.8Mbps
- Mobile video 60% of total mobile data traffic
- Top 1% of users consumed 6% of traffic, down from 52% in 2010.
Predictions for 2021
- China’s total mobile traffic will surpass that of the USA by end 2017
- Strongest growth will be in Middle East and Africa
- Global data traffic growth will be 47% annually between 2016-2021
Details include breakdowns for monthly usage, the impact of unlimited data plans, family shared tariffs.
Cisco highlight themes close to their heart, such as the Internet of Things and the potential for IPv6. You’d expect it to have a good handle on current internet traffic patterns and at least short term trends.
Ericsson Mobility Report
Ericsson has published a more focussed report since 2011, based on extensive measurements from live cellular networks worldwide. You can read the standalone illustrated report and now also create your own graphs using a traffic exploration tool.
Their forecast period is a couple of years further out, to end 2023.
Ericsson are pushing 5G very strongly, forecasting that 5G will account for 37% of all subscriptions in North America by 2023, ahead of North East Asia which includes China, Japan and South Korea. I feel this skews the longer term forecast. This may be partly due to including many early 5G deployments desgined for fixed wireless access rather than only those delivering a fully mobile cellular service.
3GPP has expanded the scope of 5G to encompass 4G evolution (anything from 3GPP Release 15) so this doesn’t necessarily mean that it’s all about 5G New Radio.
The detailed breakdown of traffic for 2016 and 2017 is very useful and granular, differentiating by region, device type and technology generation.
One interesting insight is that total cellular data traffic is dominated by smartphones (11 out of 14 Exabytes/month) with tablets consuming just 3.3%.
Rather than just looking back, Huawei takes a longer term view through to 2025. Their Mobile Broadband Report paints a picture of life at that time, predicting rapid take-up of augmented and virtual reality amongst other applications.
They examine the home, workplace, outdoors and inside connected cars, predicting growth in all four areas. They believe that mobile devices are gradually becoming the “primary screen” for consuming video content, taking over from the television. In the workplace environment, they anticipate that mobile operators will have to provide highly reliable cellular service indoors.
They forecast average traffic consumption of 0.9GB in 2016 will grow to 30GB by 2025, with cell edge data rates increasing from 5Mbps to 30Mbps.
The chart below from American Time Institute illustrates how Americans spend time during an average day.
Huawei believe we’ll spend much more time using our mobile devices, growing from 2 hours per day to 5.5. If you think this is high, a recent study by Dscout determined that a typical user touches his/her phone 2,617 times per day spending 145 minutes spread over 76 phone sessions. Apple confirmed that its customers unlock their iPhones 80 times per day – that’s six or seven times an hour.
While I couldn't uncover a market statistics report branded directly by Nokia, there is a Bell Labs Mobility Traffic Report from 2016 which has some thought provoking context for how we use (and will use) data. [Nokia acquired Bell Labs along with Alctel-Lucent last year]
For example, the chart below highlights the five primary categories of traffic consumption and what times of day they each peak at:
The GSMA, originally the Association of GSM Operators but now includes 3G/4G/5G, has increased its publication of industry data and insights over recent years.
The Mobile Economy 2017 report isn’t just a single annual report, but now includes a series of regional sub-reports released throughout the year. I can imagine these could become aligned with the GSMA’s regional events, which are replicating Mobile World Congress in each continent worldwide.
Rather than purely focus on the numbers, these reports have taken a wider view of economic developments. They cover the
Key themes in their 2017 report include:
- 4.8 billion unique subscribers growing to 5.7 billion by 2020
- 7.9 billion SIM cards
- $1.05 trillion operator revenues
- 28.5 million jobs directly and indirectly in the mobile ecosystem
- 55% of subscribers are mobile broadband 3G or 4G, growing to 75% by 2020.
- Operator revenue growth outlook “remains modest”. Somehow operators are expected to deliver much more data traffic for similar top line income. Capex is in slow decline but expected to be boosted by 5G after 2020.
- Consumer engagement through mobile has reached a scale that allows monetisation of a wide range of new services. The question is how much of that new economic benefit will be captured by operators rather than content and OTT providers.
- GSMA has frequently championed the substantial value that cellular brings to the wider economy. They believe its running at 4.4% of GDP today and will grow to 4.9% by 202, as countries benefit from improved productivity and efficiency.
Their vision for 5G rollout is beginning to crystalise:
- The pre-emptive phase providing fixed wireless access in millimetre wave spectrum (effectively bypassing/replacing/emulating) the “last mile” copper or fibre to individual premises. There will be a few headline sporting events.
- First commercial launching late 2019/early 2020 built on evolved LTE (Release 15)
- Second phase based on LTE Release 16 delivering mobile data over millimetre wave frequencies (around 30GHz).
The report documents activities and timelines in various parts of the world.
So the forecast for 5G adoption seems to primarily relate to future LTE evolution (Release 15) rather than 5G New Radio (Release 16).
It could be easy to overlook or misinterpret these forecasts.
Small Cell Forum
The Forum has commissioned analyst firms to provide current and future shipments of small cells. Analysis has been provided in turn by Informa, Mobile Experts and now Rethink Research.
The latest June 2017 edition can be downloaded here (registration required).
In the early days, forecasts were very optimistic and talked the industry up. In recent years, estimates have become more accurate and stable. Residential shipments are expected to peak in 2019, when they will be overtaken numerically by Enterprise units. Volumes for urban small cells are modest by comparison, but expected to ramp up around 2020.
One issue is that the volumes are heavily dominated by residential femtocell shipments, where market value is in the higher priced Enterprise and Urban segments. Thus variations in numbers in the millions may affect different vendors and suppliers in quite different ways.
Another issue has been the changing definition and scope of a small cell to include other architectures. The latest report includes not just standalone femtocells but also distributed radio systems such as Huawei Lampsite and Ericsson RadioDot, and remote radios with a virtualised controller.
As you’d expect, part of this report sells the value of the Forum’s work in promoting small cells over the past ten years, and asks the question of whether the technology would have been introduced if it hadn’t existed. There are also pointers to other Forum documents including case studies and its wider Release Program.
Independent Analyst Reports
In the past we have tracked a number of independent reports from analyst firms. The number of dedicated small cell analyst reports and refresh periods have dropped off, or they have been included in more comprehensive RAN studies.
Rather than pick out any individual analyst, I’d simply comment that the great value in these reports isn’t necessarily the figures alone – it’s the accompanying commentary that explains the reasoning and assumptions made to estimate them. There is a variation in forecasts, partly down to differing opinions but also partly down to differing definitions/scope of what consitutues a small cell or market sector.
There is now quite a wide range of available statistics and market forecasts freely available which a decade ago would have attracted a premium price.
Even when free, the rule of caveat emptor (buyer beware) applies.
I would be very confident that the traffic figures reported for current years is fairly accurate. But as any financial advisor has to inform you, past performance is not a guaranteed prediction of the future.
This is where its worthwhile comparing alternative forecasts for the 3-5 year period ahead (anything further than that is mostly guesswork), and focussing on the underlying reasons/critical factors involved.
- Residential femtocell shipments will peak in the next year or two and then slowly fall back.
- Enterprise in-building systems could grow rapidly. CBRS in the US will be one driver, but most of the volume has been forecast for Asia Pac.
- Urban small cell growth could ramp up over the next 2-3 years. There are many potential pitfalls but also good opportunities.
- 5G rollout will include quite a bit of marketing hype and will initially be related to future enhancements to existing 4G/LTE
- The majority of Indian and sub-Saharan users are still on 2G/GSM. There are still billions who aren’t connected or only have basic voice/text service.
So my forecast is that LTE will capture the greatest investment/spend over the next 5 years, with GSM or 3G plugging the gap where VoLTE isn’t available. In-building small cells have the most promising opportunities for growth, especially if CBRS takes-off and regulators elsewhere recognise its potential. 5G will be mainly hype for the next 5 years for full mobile cellular service and so risks just being a distraction.