Taking Free-to-Air TV online in Australia: opportunities and challenges

Full Report

Taking Free-to-Air TV online in Australia: opportunities and challenges

Already a Subscriber? Login here to view Full Report

Taking Free-to-Air TV online in Australia: opportunities and challenges

Taking Free-to-Air TV online in Australia: opportunities and challenges

Introduction

Project context and funding

This report addresses some of the issues associated with migrating free-to-air television to online delivery. It was written by David Kennedy of Venture Insights and commissioned by Swinburne University of Technology as part of an Australian Research Council funded Discovery Project, ‘Spectrum After Scarcity: Rethinking Radiofrequency Management for a Connected Society’ [DP 150100887]. The research was overseen by Jock Given, Professor of Media and Communications [[email protected]] and Giles Tanner, Senior Research Fellow [[email protected]]. Any views expressed are the author’s.

The aim of this report

The report is designed to enumerate the key technical, commercial and equity issues that would arise in the course of such a migration, providing insights and identifying problems that would need to be addressed. Amongst other things, this requires an examination of Australia’s digital infrastructure and its capacity to support television distribution, the media and telecommunications business models that would need to be adapted, and the response of households to the changes in consumer technology that would be required. A household survey was also undertaken to inform this report, and details are included in the Appendix. The report is not a proposal that free-to-air television should entirely move to online delivery, nor is it a strategy to get there. Nor is it a cost-benefit analysis.

In summary, this report concludes that it is difficult to foresee a complete migration of free-to-air television to online delivery on current industry technological and commercial settings. These settings would need to shift significantly for this assessment to change, and could require significant policy intervention to accomplish. But there is significant scope for a partial migration to be commercially driven over the coming years. The challenges to a complete migration identified in this report might be taken as a checklist of areas where the online solution might pragmatically need to be supplemented with other technical or policy approaches. Moving all existing free-to-air viewers into the online environment would meet some significant hurdles, and could only be accelerated with a planned, complex and potentially expensive intervention, in contrast to the organic rise of the “born digital” SVOD services which already reside there.

Might free-to-air television eventually migrate to the online world? One benefit of a shift to online delivery that is often touted is that it would free up spectrum for other purposes in the UHF and VHF bands, with a windfall gain to the government and the mobile telecommunications industry. From television’s perspective, it would support the medium’s gradual transition from a traditional mass market form to one that improves advertising yields through programmatic and targeted advertising, deploying a mix of traditional linear television and a widening range of on-demand services.

None of the new SVOD services have the centrality that free-to-air TV has enjoyed, and the progress of media fragmentation suggests they probably never will. It is true that the COVID crisis of 2020 demonstrated the capacity of Australia’s digital infrastructure to support a surge in online usage, including rising levels of digital video consumption. But it is equally true that the crisis also drove an increase in the usage of free-to-air broadcasting, reversing a downward trend. This reflected the enduring appeal of free-to-air television for collective participation in events of national significance. It was the free-to-air media that were the most accessible source of news as the crisis unfolded.

Assessment of the prospects for a migration of free-to-air television onto online networks must take into account this centrality, and the other features of broadcasting services and regulation that distinguish it from other forms of video services. On a digital platform, all video services are just digital applications, but from a policy perspective some services are more equal than others.

Any judgement on the desirability or likelihood of a migration of free-to-air television to online delivery first requires an analysis of the issues that this migration would generate. More generally, it would require the reconciliation of many differing interest groups: the nbn Co, the broadband retail service providers who resell nbn services, free-to-air broadcasters themselves, and of course viewers. Free-to-air national/commercial and metropolitan/regional broadcasters are all starting from different points and have different interests. Central to these interests is the question of who pays for the shift to online delivery, and who benefits from it.

Given this complexity, and the importance of the issues that a migration to online distribution would raise, what should be avoided is any rush to a “simple” solution that may have unintended consequences.

Executive Summary

The migration of free to air television to the online environment is an option for the future distribution of advertising and government supported television services. This option is attractive for several reasons. It improves utilisation of Australia’s digital infrastructure, and potentially frees spectrum for new uses. It also opens up new opportunities for the free-to-air television industry to improve viewer targeting and also advertising yield (in the case of the commercial broadcasters and SBS). Even online viewers who restrict themselves to streamed linear television would do so through an app that allowed broadcasters to harvest first-party information, opening up new possibilities for broadcasters to develop the marketing relationship.

However, any migration also faces significant challenges that would need to be addressed. The job of migrating free-to-air television to the online environment is as much about migrating viewers as it is about migrating television services. Both the demand side and the supply side of the migration need to be considered. The task of migrating all free-to-air television services (not just some), is complicated by the availability of Internet-connected television sets, variety of broadband access technologies, the disparate geographies of Australian communities, and the commercial model for broadband in Australia that includes payment for capacity, not just connection.

The supply side is more complex because it involves several supplier groups with differing interests. These include the nbn Co and other wholesale infrastructure providers like independent fibre companies and wireless ISPs (WISPs), broadband retail service providers (RSPs), and the free-to-air television service providers themselves. Nbn Co seeks the highest wholesale revenues it can get, while its retail service provider clients seek the lowest, particularly the lowest capacity charges which are their primary variable cost. Free-to-air television broadcasters want to minimise their own cost of carriage, but retail service providers want to avoid bearing the costs of additional video traffic. And viewing households want to avoid additional costs for either devices or broadband connection.

Despite the progress the nbn has made, speed remains an issue for a significant minority of households. Netflix recommends 5Mbps for viewing HD quality video, and its recommendation for 4K is 25Mbps. Assuming other applications (and potentially other channels) might be used simultaneously, a household might easily need a minimum 50Mbps service to ensure sufficient speed if multiple HD streams or a 4K stream were watched.

nbn Co’s latest Corporate Plan states that currently 66% of the network customers can receive speeds of 100Mbps, and half of those can get speeds of 250Mbps or above. But this leaves a total of 34% (or roughly one third) of current nbn connections that cannot support speeds of 100Mbps. Half of those cannot support 50Mbps (and this only when legacy ADSL services are switched off in 2022).

nbn Co has committed to invest a further $3.5b by 2023 to lift the availability of speeds in excess of 50Mbps to 85%. However, this program does not address any limitations in the fixed wireless and satellite user base, and will still leave 15% of total connections on 50Mbps or less, and 5% on 25Mbps or less. This leaves a significant minority potentially with insufficient broadband access speed.

Other issues and potential controversies are related to the sheer volume of data that would be shifted onto online networks if free-to-air were migrated there. At current usage and penetration levels, nbn capacity virtual circuit (CVC) capacity per user would need to roughly double to accommodate free-to-air television viewing, in an OTT mode (an important qualification). This would eliminate the free-to-air service providers’ current transmission costs, but would cost the RSPs around an additional $4 billion per annum, around 5 times more than current free-to-air transmission costs and about two-thirds of the broadband RSPs’ revenues. This is clearly impractical.

However, the use of more efficient distribution technology than OTT offers an alternative. The nbn network also supports a ‘multicasting’ technology that allows multi-channel linear video to be ‘broadcast’ to large numbers of viewers across the nbn. This approach can significantly reduce the capacity burden even at current nbn wholesale prices. Adopting this approach could bring the total industry cost of nbn distribution into the vicinity of current broadcast transmission costs. This solution would require the free-to-air television industry (preferably both national – ABC and SBS - and commercial) to take a coordinated approach to using the nbn, but there are precedents for this cooperation in the operation of existing broadcast multiplexes used for digital terrestrial television broadcasting. Over time, advances in compression technology will also help to reduce costs.

Non-nbn networks like independent fibre operators and wireless ISPs (WISPs) lack the multicast technology nbn has. Similar capability could be built to mitigate capacity costs in any migration scenario, but this would add to the cost base of relatively small telecommunications operators and would raise barriers to entry into that market. There is a scale below which such investments would not be practicable, and this would hit small WISPs hardest. Some kind of policy provision would need to be made to cover these exceptions.

Network resilience is a separate consideration from network capacity. There is a headline question whether it makes sense to consolidate Australia’s telecommunications and broadcasting services onto a single infrastructure, and this is most apparent in emergency situations. The experience of the 2019-20 bushfires is instructive. Both telecommunications and broadcasting sites were damaged by fire, but in most cases, outages were caused by power failure. Unfortunately, no comprehensive analysis of these outages has been undertaken to determine the sources of fragility. It would be necessary before conducting any migration to understand how the redundancy between telecommunications and broadcasting networks contributes to communications resilience in emergency situations.

One approach would be to recognise radio as the alternative infrastructure. This has some appeal as radio has greater utility in emergency situations than television. In any case, it is likely that a resilience strategy would be needed to ensure that both telecommunications and broadcasting were not compromised in a variety of emergency scenarios. This strategy could include site hardening against natural and criminal events, improved local generation backup arrangements, and undergrounding of power infrastructure to key infrastructure sites. Clearly, this would add considerably to the expense of any migration, but could equally be regarded as an overdue investment in emergency management.

Finally, the key geographical coverage issue regarding nbn’s network is the satellite service. No unlimited plans are available on the Sky Muster satellite service. If nbn satellite is excluded as an option, then nbn’s fixed and fixed wireless networks would have to be relied upon to provide coverage. This would be straightforward in urban areas where fixed network technologies can deliver both high speed and unlimited capacity plans, and the same is true in regional areas where nbn’s fixed and fixed wireless network technologies are deployed.

However, in any given regional location, there is no guarantee that the non-satellite nbn networks will be available everywhere that free-to-air television is currently available. Broadcast and nbn networks use different technology and are optimised for different use cases, and inevitably do not reproduce each other’s coverage. There will be areas with poor broadcast reception and good nbn coverage, and vice versa. The question of whether nbn connectivity was available to convey free-to-air television could only be determined on a locality-by-locality basis.

For this reason, a purely nbn-based migration would not be practicable for all current broadcast viewers, particularly those in regional areas. The only practicable solution would be to combine nbn delivery with something like an expanded VAST service, which is currently used to deliver a selection of free-to-air television outside the regional broadcast licence areas.

On the demand side, viewers face two main issues. First, we estimate over 20% of households do not have a fixed broadband connection, and this share has not changed much in recent years. Of those who do, a significant share still face issues with capacity. A survey conducted to inform this report found that 30% of households worry about their data allowance when watching online video, and this figure will only increase if free-to-air television was migrated to online delivery. Ensuring that all of these households could continue to watch free-to-air television after migration would probably require significant subsidy of broadband connections, and potentially also an ongoing subsidy for recurring monthly costs in some cases. Targeting such a subsidy to those who really need it would be difficult. This is probably the biggest issue that a managed migration would need to address.

The issue of getting an Internet-connected television set was less of a concern. 69% of respondents to our survey told us they already had an Internet-connected TV. If only those with an existing fixed broadband connection are considered, the figure was 83%. Most TVs sold in the last decade are either natively Internet enabled or can be connected using options such as Chromecast devices. These devices are relatively inexpensive, so enabling television sets to view online video is a relatively minor issue financially. However, it might be necessary to educate some viewers on how to install and use these devices. Over time though, this issue will fade as older television sets without native Internet connectivity are gradually replaced by new ones.

In summary, there are considerable obstacles to the idea of migrating all free-to-air television to the online environment. This is not to say that online networks have no role to play in the distribution of free-to-air broadcasting. This is already happening. These obstacles arise from trying to migrate all of the viewers, on all of the broadband technologies, in all of the locations where television is currently broadcast. In each of these cases, there are significant minorities of viewers who will be costly to migrate.

Of the issues identified, the commercial ones are the least concerning. Nbn Co’s Multicast pricing is already in the vicinity  where a commercial discussion could take place between nbn Co and free-to-air broadcasters. Advances in network efficiency and video compression technology will only ease pressures, although future 4K demand remains a wildcard.

In contrast, there are some significant technical and viewer obstacles. The main obstacle is the number of households who are still unconnected to fixed broadband, which remains a significant minority. These households have low interest in online activity, and migrating these households to online delivery will be difficult and potentially costly. The next most important issue is the number of nbn households that lack sufficient speed to consistently support television along with other applications, though this issue is gradually being addressed by nbn Co and can be expected to be resolved over a number of years.

Other issues include nbn coverage and network resilience,  but these problems are more manageable within the existing framework of programs (e.g. the VAST service) and policies (e.g. network hardening for emergency preparedness). Other issues like non-Internet television sets will fade naturally over time.

In conclusion, it is difficult to foresee a complete migration of free-to-air television to online delivery in the immediate future. Some of the obstacles identified will take time to resolve and could require significant policy intervention. But even a partial migration could achieve some important benefits. These include a digital marketing opportunity for free-to-air television amongst online viewers, and savings on transmission in high cost regional areas where terrestrial nbn has already been deployed.

State of the free-to-air television industry

Free to air (FTA) broadcasters have traditionally played a dominant role in aggregating television content from multiple content owners and creators, despite competition from pay TV. However, over the last five years, the growth of subscription video-on-demand (SVOD) has driven a significant transformation in TV viewing habits. Audiences today have unprecedented choice when it comes to content and how to watch it.

What is distinctive about free-to-air television?

Free-to-air television has national, collective character that is hard for other services to replicate. The high levels of household participation in free-to-air TV consumption and the ease with which this system handles major surges of viewing in a national occasion or crisis, are closely related to its current delivery arrangements. These include public support for a national broadcast delivery infrastructure, the availability of cheap and easily available television devices, and an advertising-supported and/or publicly supported revenue model which lowers or eliminates price barriers to participation. These features arguably account for the robustness of television over the last two decades despite rising competition from pay TV and then SVOD.

It is important to recognise that these technical features have been major factors in the formation of this national, collective character. Once this is recognised, it is clear that the impact of a move to a different mode of delivery is something that would have significant implications for the future of free-to-air television.

For policy-makers, the distinctive features of broadcasting services that need to be considered are summarised in the 21 objectives set out in Section 3 of the Broadcasting Services Act 1992. The Act sets out a framework of objectives more complex than, for example, the framework for telecommunications. Many of its objectives are inter-related, sometimes reinforcing one another and sometimes being traded off. Amongst the objectives set out in the Act, concepts that stand out are:

  • Availability
  • Efficiency, competition and responsiveness
  • Diversity of both control and content
  • Australian content and content of local significance
  • Community standards

Of these, the first two concepts are most directly at play and are the focus of this report. As already noted, the current delivery arrangements for free-to-air television are a key support to its current high levels of availability. In addition, efficiency considerations, of the kind that impact business decisions, are at play as broadcasters look to manage the costs associated with content production, programming and distribution as advertising and government-sourced revenues fall relative to costs. A migration would also add new complexity to ownership arrangements, with broadcasters becoming dependent on digital infrastructure owners to deliver their services.

But even local content rules and community standards come into play at some point. Does it make sense for free-to-air broadcasters and SVOD providers sharing the same digital platform to have different responsibilities for Australian content? And how are broadcasters and narrowcasters, with their different levels of responsibility for community standards, to be distinguished when they share the same online environment? And how could local content be preserved, particularly regional news and current affairs, if and when Australia becomes a single online broadcasting market?

The focus of this report is on the technical and commercial challenges of a migration, but this brief discussion is enough to set aside any notion that migrating free-to-air television to online delivery would simply amount to a change in the technology of delivery. Current free-to-air regulatory and industry arrangements are integral to the achievement of current broadcasting policy objectives, and any change in the former has implications for the latter. These issues cannot be identified outside the framework of the policy and industry objectives that frame broadcasting regulation.

Rising competition

The delivery of these policy objectives is taking place against a backdrop of rising competition and falling television industry revenues. Free-to-air television has faced competition from pay TV since the 1990s. However, household penetration of pay TV in Australia never exceeded 30%. But competition for “eyeballs” has intensified as new SVOD service offerings have emerged in the 2010s. Venture Insights estimates that SVOD subscriptions overtook pay TV subscriptions in Australia in 2016, led by Netflix which dominated the early SVOD market. The free-to-air television industry has responded by making more of its content available online through catch-up services (broadcast VOD, or BVOD), though viewer minutes are still dominated by traditional delivery.

The result is fragmentation of the industry, with a move away from a single, homogeneous bundle of channels to a much more diverse set of content choices. Rising competition has had a significant impact on average free-to-air television viewing. Average viewing time per day per user fell from 155 minutes in the year to December 2017 to 138 minutes in the year to June 2020. Viewing of free-to-air catch-up services (broadcast video on demand, or BVOD) grew rapidly over this period, but this was off a small base and was not enough to offset viewing declines on the traditional delivery platform. The June 2020 result was actually an improvement on the previous year, reflecting the return of viewers to free-to-air television as the COVID-19 lockdowns played out. However, even this improvement did not bring total free-to-air viewing (traditional plus BVOD) back to the levels of June 2018. It is interesting to note that pay TV also experienced a decline over this period, suggesting that SVOD was the culprit.

The negative trend in free-to-air television viewing has led to a decline in advertising revenue. The growth of free-to-air advertising revenue has been declining since the December 2017 half year. Venture Insights has forecast free-to-air television ad revenues to recover strongly as the economy bounces back from COVID-19 in 2021 and 2022, but for decline to resume after the economy stabilises in 2023. Overall, we expect FTA ad revenues (including BVOD) to decline CAGR -2.3% over FY19 to FY24. This is mainly driven by underlying decline in time spent with TV. The other major source of funding for Australia free-to-air television is allocations to the national broadcasting sector, which have stagnated over the last few years.

Current arrangements for free-to-air broadcasting transmission and distribution

Free-to-air television is broadcast in dedicated spectrum in the VHF and UHF bands. Since the introduction of digital television broadcasting in 2001, analogue television has been phased out and digital transmission based mainly on the MPEG-2 standard has been used.

Free-to-air television transmission (for both commercial and national networks) is generally outsourced to specialised  infrastructure companies, led in Australia by the multinational BAI Communications, but with important contributions by TX Australia (owned by SevenWest Media and Nine Entertainment), and Axicom (another multinational). There is no official tally of the full cost of transmission to the free-to-air television industry. The ABC estimates its total transmission and distribution bill (television and radio) was $187 million in 2018-19 and Venture Insights estimates that the total annual television industry cost is around $800 million.

Over the last few years, free-to-air networks have also begun making most of their content available as BVOD; either as live streaming, or as a catch-up service online in an over-the-top (OTT) Internet mode similar to Stan and Netflix. Viewers must own an Internet-enabled television to use these services, just as they must to view SVOD services. In this case, the free-to-air television providers must pay for their own Internet connectivity, but they do not pay for the viewers’ connectivity. This means that the bulk of the cost of video delivery on the Internet is borne by the broadband retail service providers, who must purchase data capacity from the nbn Co. These are basically the same commercial arrangements that Netflix and Stan enjoy online. In particular, it is the RSPs that must bear the cost of any surge in capacity usage as a consequence of higher video viewing. These costs are passed on to broadband consumers.

Issues in the migration of FTA broadcasting: industry

The speed and capacity requirements of online networks obviously constrain the technical ability to deliver free-to-air television online to households. Different fixed and wireless networks have different abilities to deliver these services. In particular, Australia’s largest and most ubiquitous network the nbn has some specific features that affect the way video is delivered:

  • The nbn uses a variety of access technologies in different locations, some of which constrain the speed available to any given household
  • The performance of nbn services depends on both the speed of the access virtual circuit (AVC, the “tail” into the premises) and the size of the capacity virtual circuit (CVC, the capacity provisioning into the nbn aggregation network). Speed and capacity are therefore largely determined independently (with nbn satellite being the notable exception)
  • The nbn also offers different traffic classes with different quality performance. Household services are Class 4 services, which are asymmetric and “best efforts”. In contrast, Class 2 services are symmetric and “business grade”. Class 1 services typically support low bandwidth voice services. Nbn has also implemented a separate class of traffic for video under its Multicast video distribution service, but this is not currently in use.

To determine what limitations these networks may place on video distribution, it is first necessary to determine what technical capabilities are required. Then we turn to how networks can deliver these capabilities.

Technical requirements for online distribution of television

Rising screen resolutions are driving demand for higher definition television, though this demand is not always met. The resolutions available for entertainment video include:

  • Standard Definition (SD, 576i), resolution 720×576
  • High Definition (HD, 720), resolution 1280×720
  • Full HD (1080i or 1080p), resolution 1920×1080
  • Ultra High Definition (UHD, 4K, 2160i or 2160p), resolution 3840×2160
  • 8K (4320i or 4320), resolution 7680×4320.

Australian free-to-air broadcasters, which are limited to 7MHz spectrum each to broadcast digital television services, currently deploy a mix of these resolutions depending on the type of content, balancing resolution against the number of channels. SD tuners cannot display HD signals, so free-to-air broadcasters typically simulcast their main channel on HD and SD to ensure availability to both HD and SD households, with the remaining channels in SD to maximise the number of channels. Free-to-air television broadcasters are required to provide a minimum of SD quality, but they must also offer a minimum of 1040 hours per year of Full HD content. 4K channels are not broadcast on free-to-air principally due to lack of bandwidth.

Foxtel broadcasts dedicated Ultra-HD (4K) channels for sports and movies on its satellite services. The Foxtel Now set-top box allows up to full HD streams for on-demand titles, but only HD on its linear channels. Some 4K content is also available on Netflix, Stan, YouTube, Google Play, Apple iTunes, and Amazon Prime. Of course, physical media such as Blu-Ray are also available in 4K.

The survey conducted by Venture insights shows that availability of 4K viewers is something of an obstacle to the deployment of 4K content, but a diminishing one. We asked households the highest resolution TV set they had. Of those who responded, only 10% had an SD television, while 65% had an HD television. 26% had a 4K or better television, which is a substantial share. This is an important issue for the free-to-air television industry, as reach drives both advertising revenue and government funding. This is particularly important for the commercial television broadcasters, who are already seeing advertising revenue decline.

There was also evidence that this share will grow. 66% of respondents said they had plans to buy a (or buy another) 4K set, with 20% of respondents planning to do so within 12 months. Pay TV viewers were somewhat more likely to have these plans, with 26% planning a purchase in the next 12 months, probably due to the greater availability of 4K pay TV content. This shows that there is now a significant and growing addressable household market for 4K content.

The larger file sizes of higher definition television also consume more data capacity. For households on limited broadband plans this is an issue. These may be households using mobile networks or nbn satellite where unlimited plans are either unavailable or subject to fair use restrictions that limit usage. This is less relevant to households with an unlimited plan, but has material consequences for retail service providers (RSPs) that must provision this data capacity and pay for it.

Online bandwidth requirements escalate rapidly as screen resolution increases because digital file sizes for full HD and 4K video are higher. There are also innovations like High Dynamic Range (HDR) that improve colour resolution as well. Video compression technologies are an important offset to the trend to larger video file sizes and the need for higher broadband speeds. Video compression techniques, embodied in software or hardware codecs, are advancing all the time, and dozens of different codecs are available. Codecs like H.266/VVC , which are optimised for the large file sizes of 4K video, already exist.

However, deployment of new codecs typically lags their development because there is always a subset of TV sets (and/or set-top boxes) that are turned over by households infrequently. We asked our surveyed households whether they had a main TV set where most viewing took place, and if so, how old it was. The average response was 4.4 years, but there was also a long tail of older sets, with 21% between 6 and 10 years old, and 5% of sets were 11 years or older.

Netflix, which uses the H. 264/AVC compression codec for SD and HD video and HEVC for 4K video, recommends the following for online viewing of a single channel:

  • 0 Mbps - Recommended for SD quality
  • 0 Mbps - Recommended for HD quality
  • 25 Mbps - Recommended for Ultra HD (4K) quality

These estimates are a rough guide, but the actual bit rate a video stream requires depends on several factors, including how much loss of quality is tolerated in compression and decompression of the data. Higher tolerance of picture degradation would lower bandwidth requirements.

The bit rate for an uncompressed 4K video stream can reach up to 6Gbps, clearly beyond the nbn at this time. But Netflix’s 25Mbps recommendation for 4K shows that while compression has not completely offset the need for higher bandwidth, it has dramatically reduced the impact of rising resolutions on both broadcast and broadband networks. The compression technologies deployed are generally lossy, meaning that there is some loss of picture quality due to compression. Compression ratios vary depending on the amount of quality loss tolerated, but typically reduce the required data speeds by factors in the 100s.

In Australia, free-to-air television broadcasters currently use MPEG-2 compression in almost all cases. MPEG-2, which uses the H.262 and H.232 codecs, is now fairly old but it is widely used in broadcasting both here and overseas because it is backwards compatible with existing hardware and software. There are certain Australian television channels that use more advanced MPEG-4 compression, which can incorporate more advanced codecs like H.264/AVC (used by  Netflix for SD and HD) and H.265/HEVC (used by Netflix for 4K). These offer around a 50% reduction in bit rate over MPEG-2 video. The recently announced H.266/VVC offers another 50% reduction in bit rate for high resolution video such as HD, 4K and 8K. However, none of these more advanced codecs are compatible with older television sets that only employ MPEG-2 compression.

Any television set or set-top box sold in the last ten years will most likely support MPEG-4, but there will still be some that do not, contributing to a reluctance by broadcasters to fund a systematic upgrade of the broadcast transmission infrastructure they lease. This reluctance comes at a cost in bandwidth. A full HD channel compressed using the MPEG-2 codec requires around 20 Mbps, and more up-to-date codecs would reduce this significantly. But in practice, the free-to-air television offer is not principally limited by bandwidth, but by the supply and cost of good quality content. Freeing broadcast spectrum by adopting newer codecs would therefore yield little benefit, as it would not be practicable to fill up the extra digital broadcasting channels that would be created. Some SD channels could be converted to HD, but this would be a relatively marginal improvement. A more attractive option would be to use MPEG-4 compression to reduce spectrum needs, allowing more channels to be broadcast off a single transmitter multiplex. So far, this advantage has not been enough to overcome concerns about stranding older MPEG-2 compatible television sets.

Overall, the trend for video data demand is for the required bandwidth to increase with each new iteration of screen resolution, but at a much lower rate than the rising data requirements would suggest due to advances in video compression technology. It could be argued that it is therefore possible to foresee a ceiling for household video bandwidth requirements. This is because higher screen resolutions offer diminishing returns in the standard use case of a household loungeroom. On a typical 52 inch screen, the difference between full HD and 4K resolution is difficult to perceive if the viewer is more than a metre away from the screen. Other use cases like home theatre, where the resolution difference is much more noticeable, are much rarer and likely to stay so for the foreseeable future. Of course, a household may require more than one channel at any given moment, but a 4K set could represent something of a final destination for most households, with the 25Mbps speed per channel at that resolution being a benchmark for the foreseeable future. But the word “foreseeable” is doing some work here; some past estimates of the maximum bandwidth required by households have underestimated current demand, so caution is required making such predictions.

Network capabilities: speeds

First, consider data speeds. The nbn’s fixed footprint comprises FTTH, FTTB, FTTC, FTTN and HFC technologies. Of these, FTTH (20% of the nbn footprint) already offers 1Gbps speeds. FTTC can comfortably deliver 100Mbps+ services that easily accommodate multiple 4K video streams. FTTB’s performance depends on in-building wiring, but these connections are typically very short so similar performance to FTTC is the norm. HFC currently offers up to 250Mbps, with an established technology roadmap to 500Mbps and beyond.

FTTN’s performance is much more variable. Nbn Co’s Corporate Plan states that 36% of connections are FTTN. Once legacy ADSL services have been switched off in 2022 (allowing nbn Co to use vectoring technology to improve speeds) FTTN performance will improve. Of the 36% of total connections that are FTTN, 10% will be able to get 100Mbps, but 17% will only be able to get 50Mbps and 9% will only be able to get 25Mbps.

Fixed wireless and satellite have specific speed and capacity issues. Fixed wireless is used by 5% of connections, and speeds are currently capped at 50Mbps. Satellite is used by 3%, and is capped at only 25Mbps; in addition, unlimited plans that include video are not available for satellite.

nbn Co’s latest Corporate Plan states that currently 66% of the network can receive speeds of 100Mbps, and half of those can get speeds of 250Mbps or above. But this leaves a total of 34% (or roughly one third) of current nbn connections cannot support speeds of 100Mbps. Half of those cannot support 50Mbps (and this is in 2022 when legacy ADSL services are switched off).

nbn Co has committed to invest a further $3.5b by 2023 to lift the availability of speeds in excess of 50Mbps to 85%. However, this program does not address any limitations in the fixed wireless and satellite user base, and will still leave 15% of total connections on 50Mbps or less, and 5% on 25Mbps or less.

The availability of high speed fixed broadband has improved dramatically in Australia over the last ten years. However, the persistence of a significant minority of households that cannot access 50Mbps, let alone 100Mbps, is a significant commercial and policy obstacle to any migration of free-to-air television to the online environment. It is a significant commercial obstacle, since free-to-air television providers do not want to abandon up to 20% of their audience (and a corresponding share of their advertising revenue). It is also a significant policy obstacle because it would be a significant blow to the objective of broadcast service availability.

There would also be a significant geographical disparity in the impact. The bulk of these connections will be fixed wireless and satellite that are unique to regional areas, so regional audiences and regional television broadcasters would be disproportionately affected. In addition, some FTTN connections would still be unable to achieve acceptable speeds, and these would be distributed across regional, outer metro and some inner metro areas depending on factors such as the quality of the remaining copper and distance from the node.

Traffic class is not a major factor in the delivery of household video. In practice, “best efforts” means that the service is susceptible to occasional slowness, and this aspect of quality is largely determined by the level of capacity provisioning offered by the RSP at peak traffic times. Although this has proven acceptable for “born digital” services like SVOD, it is debatable whether this would be acceptable for free-to-air television. Current free-to-air broadcasting technology makes it easy to support as many viewers as wish to tune in at any time without any reduction of quality, and this quality does not depend on their choice of service provider. While occasional SVOD quality issues may be acceptable, free-to-air over digital networks suffers from any comparison with traditional broadcast technology.

We asked households in our survey on a scale of 1 to 5 whether the prospect of moving free-to-air television was not at all concerning (1)  to very concerning (5). The average level of concern was high, at 3.3. Those concerned raised a range of issues, among them the need to upgrade a broadband service to cope with extra usage (21%), failure of connection in an emergency situation (33%), and reliability during live events (34%). This suggests that some viewers perceive a significant quality difference between traditional broadcast and online technologies, and that this perception would need to be addressed if any organised migration were launched.

Non-nbn networks should also be considered. Certain parts of the Australian market are served, sometimes exclusively, by non-nbn infrastructure. The main examples of fixed infrastructure provision are TPG (which operates some fibre access infrastructure), and Uniti (a wireless broadband operator that has acquired fixed providers Opticomm and LBNCo). We estimate that TPG has over 100,000 active fibre customers, while in 2019 Opticomm reported 60,000 and LBNCo reported 21,000.

There is also a range of wireless internet service providers (WISPs), many of which operate in either urban or regional areas. There are no official estimates of the numbers of WISPs in Australia, but ACMA reported that 17 WISPs held point-to-multipoint licences in the 3.6GHz band. Uniti is one of the largest WISPs, and operates in Adelaide and Melbourne. WISPAU, the industry body for WISPs, estimates that over 200,000 premises are currently served by non-nbn and non-mobile fixed wireless broadband. To these must be added growing numbers of users who are adopting 5G fixed wireless services from mobile operators Telstra and Optus (TPG has yet to launch a 5G fixed wireless offer). The number of 5G fixed wireless customers has not been announced, but is growing fast.

FTTH and FTTB predominate in the Opticomm and LBNCo networks, so speed is not an issue for their users. The situation with fixed wireless operators is more complicated because there are many of them, and they increasingly offer a mix of WISP, nbn fixed wireless and nbn satellite services. Uniti, for example, offers three WISP broadband plans at 25Mbps, 50Mbps and 100Mbps with unlimited data. All of these speeds are capable of supporting free-to-air streaming services, but it is difficult to generalise about the WISP industry as many providers are small and/or localised by the standards of the telecommunications industry. Many now offer nbn plans, including satellite plans which suffer from the speed and capacity limitations discussed above.

Current 5G fixed wireless offers are still emerging and subject to change. Telstra launched the new 5G home broadband service in October 2020, offering download speeds of between 50 Mbit/s and 300 Mbit/s and a 500GB per month data cap. Optus offers two 5G fixed wireless products, both with a minimum 50Mbps product and unlimited data. Speed is therefore not a limitation for these products, though there are capacity limitations which are discussed below.

Network capability: capacity and reliability

Video is the largest contributor to global Internet traffic and drives the traffic management challenge. Traffic capacity is separately priced in nbn’s charging model, so the management of traffic also has significant commercial implications.

Nbn’s standard pricing construct for its “best efforts” household product has three elements, corresponding to the nbn’s main network elements:

  • An Access Virtual Circuit (AVC) which determines the speed of the service, is dedicated to the individual user. Higher speeds cost more per month. The more users the RSP has connected, the more AVCs they need to buy
  • A Capacity Virtual Circuit (CVC) which determines the amount of bandwidth allocated to the RSP’s users. More capacity costs more per month. This capacity is shared by all of the RSP’s users in a given nbn service area, so RSPs must buy enough CVC capacity to carry traffic at peak times.
  • A Network to Network Interface (NNI) charge for connection between the RSPs backbone network and nbn’s Point of Interconnect (PoI). Nbn Co operates 121 PoIs around the country, each corresponding to a service area.

The ACCC found in its September 2020 quarterly wholesale indicators report that average CVC capacity per nbn user is 2.6Mbps. Video traffic dominates consumer traffic, so it is reasonable to just assume this capacity is fully committed to video traffic for the purposes of an estimate.

ACMA, in its latest Communications Report (2018-19), found that online video (including SVOD, catch-up TV and other online video, and excluding gaming) accounted for 9.0 hours per week of viewing per viewer in May 2019, while free-to-air television (live and recorded) accounted for 8.7 hours per viewer. In other words, online and free-to-air viewing were on a par for individuals. Of course fixed broadband connections are connected to households, not individuals, but assuming that household viewing roughly tracks individual behaviour, shifting free-to-air television to online delivery alongside SVOD would roughly double video traffic and so double the CVC capacity required. This assumes that online and free-to-air viewing occur at the same time of day, which is highly plausible. This would cause a dramatic one-off jump in traffic (one-off because total viewing time is unlikely to change in the long run).

This “over-the-top” (OTT) delivery arrangement would be highly inefficient because it requires the RSP to buy sufficient CVC capacity to deliver a separate digital video stream for each nbn user. The technical implications of this inefficiency are very significant, but the commercial implications are too. If free-to-air television migration doubles the required capacity, then the total extra capacity needed would be equal to the existing CVC capacity per user (2.6 Mbps) times the number of active nbn accounts (7.3m in June 2020), or 19.0 Tbps. At nbn Co’s current standard price for CVC capacity ($17.50/Mbps per month), this would cost RSPs around an extra $332 million per month, or around $4.0 billion per annum.

This does not include the extra transit costs that RSPs would incur to carry this additional traffic from the nbn’s PoIs to their own Internet PoIs, or the additional connections that will come on line over the next eighteen months as the last legacy services are transferred to the nbn. Nor does it take into account the cost to households that might need to take up a higher speed AVC to cope with extra online video consumption.

An added complication is that it is RSPs and not free-to-air television broadcasters who would bear this cost. To put this $4.0 billion cost into perspective, the RSP industry’s entire fixed broadband revenue in FY20 was $5.6 billion. In contrast, the ABC’s entire (television and radio) transmission and distribution budget in 2020 was only $187 million, suggesting that the total free-to-air sector bill for television transmission might be around $800 million, or one-fifth as much as the estimated cost of nbn delivery using this form of OTT delivery.

The RSPs would need to pass this cost on somehow. RSPs have been unsuccessful in extracting payment for carriage from content providers all around the world, so their alternatives in Australia would be to pass this cost on to households in the form of significantly higher broadband prices or onto taxpayers by seeking lower rates for nbn capacity. Neither appears to be feasible, as they would certainly encounter significant consumer and/or political resistance.

An OTT approach to free-to-air online delivery is clearly commercially impracticable. However, there is a technical alternative that can address the inefficiencies of an OTT approach. The nbn provides an alternative point-to-multipoint product (called Multicast) that allows multichannel video to be injected into each PoI and transmitted to every nbn user connected to that PoI. To put it simply, a single video stream is injected at the PoI, sent through the capacity part of the nbn, and is then copied to each viewer access line. This ‘copying’ means that there is no need for each viewer to have a separate video stream pass through the capacity part of the nbn.

The commercial model is different to normal OTT video delivery as the broadcaster, rather than the RSP, pays for a Multicast connection. Apart from paying for nbn Multicast capacity, broadcasters would also be required to convey their video streams through transit networks to each of the 121 nbn PoIs. However, using Multicast would also reduce these transit network costs because they could deliver a single stream (or streams for a multi-channel television service) to each PoI, rather than delivering a separate stream(s) for each viewer.

The wholesale price structure for Multicast video distribution is similar to the price structure for nbn’s normal broadband connections but is cheaper, reflecting the use of a single stream to deliver video across the network rather than a separate stream to each viewer. Multicast pricing has three main components:

  • A Multicast Access Virtual Circuit (MAVC) is required for each user. The MAVC needs to be dimensioned to accommodate peak speeds for the total number of streams to made available simultaneously
  • A Multicast Domain (MD) charge at each PoI. The MD needs to be dimensioned based on the total size of all of the media streams being simultaneously injected at the nbn PoI. The MD can be dimensioned in increments of 100Mbps up to 1000Mbps.

If the number of media streams exceeds 200, there is an additional charge, but this will not apply to free-to-air services which only have around 5-6 streams individually and around 40 collectively.

The most efficient way to organise a Multicast delivery of free-to-air television would be for the free-to-air industry to share a single MAVC to each household, operating alongside the existing AVC. The industry would then bundle their total 40-odd channels into a single multichannel stream which would be delivered to each nbn PoI and passed through the network to each household MAVC. It would be necessary to have some infrastructure for the free-to-air networks to insert advertising appropriate to the area covered by each PoI. Viewers would be free to choose which free-to-air channel(s) to watch at any given time from the selection on offer.

Under this scheme, the free-to-air industry would need to collectively provision (say) a 20Mbps MAVC at $5.00 per month on the nbn to each user (7.3 million in June 2020). This assumes that the free-to-air broadcasters adopted a more advanced video compression scheme than MPEG-2 such as MPEG-4. In that case, the MAVC could deliver one HD channel at a bit rate of around 10 Mbps, plus some additional SD channels, at any given moment. The cost to the industry of provisioning this access capacity for the current 7.3m nbn connections would be $36 million per month, or $438 million per annum. This would rise somewhat as more nbn connections come on line over the next eighteen months.

In addition, the free-to-air industry would need to pay for a Multicast domain at each PoI where it offered service. The entry Multicast Domain size is 100 Mbps, and can be purchased in additional increments of 100 Mbps, up to a maximum of 1000 Mbps. Allocating a very generous 100 Mbps to support each existing station line-up of one HD channel and several SD ones, it would still cost only $1,250 per month per PoI. For all fixed stations, at all 121 PoIs, this would add up to only $151,250 per month, or around $1.8 million per annum.

Using Multicast, the capacity-oriented Domain charge is much lower than the capacity charge for OTT delivery. The access-oriented MAVC charge is still over $438 million per annum, but this is lower than our estimate for the current costs of television transmission.

These calculations are approximate and not intended to reflect exact costs, nor do they take into account future developments like the growth of 4K television. Also, the free-to-air television operators could only set up a MAVC to the household if an nbn connection was already in place. In this case, the household would continue to bear the monthly cost of their “best efforts” broadband traffic while the industry would bear the cost of their free-to-air video traffic. However, these calculations are enough to show that online delivery of free-to-air television on the nbn is commercially feasible for viewers with suitable access technology if free-to-air television industry cooperated on a Multicast approach.

It is important to note that, so far, nbn’s Multicast product has never had a customer. The product was launched early in nbn Co’s life before the rise of SVOD players demonstrated that satisfactory service could be achieved on nbn’s “best efforts” household service. Subsequent to that, free-to-air television providers built their own BVOD services in an OTT mode.

However, if free-to-air was delivered exclusively online, the “best efforts” attitude would need to be reconsidered. As we note below, free-to-air television viewers cite reliability as an issue for any migration to the online environment, and it would be incumbent on both policy-makers and broadcasters themselves to address those concerns with a technical solution that made service reliability comparable with traditional broadcast performance. The Multicast service specification provides an avenue to address those concerns. The Multicast product currently offers a better quality of service than nbn’s Traffic Class 4 “best efforts” service now used by most households. If taken up, it would have improved free-to-air quality and reliability compared to the OTT offers of SVOD players.

The nbn would reserve the right to review Multicast service definitions and pricing if the product were taken up at scale. If a deal for nbn distribution of  free-to-air television could be established, the broadcasters would still be subject to any future price variations, including variations that will emerge from regulatory intervention in nbn pricing. This adds a layer of regulatory risk that does not exist for current television transmission infrastructure, and this is something the free-to-air industry would need to learn to manage.

Non-nbn network providers such as independent fibre operators and WISPs lack the Multicast technology nbn has. Similar capability could be built by these operators to mitigate capacity costs in any migration scenario, but it would add to the cost base of relatively small telecommunications operators and would raise barriers to entry into that market. There is a scale below which such investments would not be practicable, and this would hit small WISPs hardest.

Finally, it is impracticable for mobile networks, even 5G networks, to play a leading role in any free-to-air migration to the online environment. Mobile networks are optimised, unsurprisingly, for mobility not capacity. According to the latest ACCC figures, Australia’s fixed networks carry over 9 times more data annually than Australia’s mobile networks, and this ratio is rising[1].

New data traffic from free-to-air television migration would roughly equal current fixed traffic levels, and there is no scenario where mobile networks could carry this much data. Nor would mobile operators want to, as a large step-change in mobile traffic would seriously compromise the quality of existing mobile services. Setting aside the capacity that mobile operators are committing to 5G fixed wireless, the remaining gains from 5G are already needed to cope with the growth of existing mobile video services. Mobile networks, even 5G networks, could technically only make a contribution to free-to-air migration for a small segment of households, and even this would be questionable from a commercial point of view.

Network resilience

Viewer concerns over the resilience of online networks in comparison to broadcasting networks have already been noted. The experience of the 2019-20 fire season is instructive. Broadcast, mobile and nbn infrastructure were all affected by the fires. In submissions to the Senate Standing Committee on Finance and Administration, which inquired into the fires in 2020, Telstra submitted that power failures were the primary cause of outages in its network, and Free TV submitted the same regarding broadcasting outages[2]. It is important to note that there was significantly more concern about radio broadcast infrastructure than television, reflecting radio’s utility in a range of situations, including at home and on the road.

Unfortunately, the Senate inquiry and the similar State inquiries did not systematically catalogue communications failures in the fires, so a direct comparison of the resilience of online and broadcast infrastructures is not available. This would be a useful area for further inquiry. There were failures on both sides. A number of broadcast transmitters sustained significant damage, but in most cases of failure sites were not damaged but lost mains power as the electricity grid was damaged, and had to be powered by backup diesel generators which ran out of fuel after a few hours.

In its submission, BAI Communications (which operates most broadcast transmission sites in Australia) recommended significant hardening works, particularly for smaller regional transmission sites. Free TV recommended that the Department of Infrastructure, Transport, Regional Development and Communications undertake a separate inquiry to determine what steps the Government could take to improve the resilience of broadcasting infrastructure.

Although a direct comparison of resilience between telecommunications and broadcast infrastructures was not undertaken, the prospect of migrating television to the online environment raises a specific and important issue. Resilience requires redundancy, and shifting all communications onto a single online infrastructure would therefore have significant consequences for resilience. In any given situation, telecommunications infrastructure may fail while broadcast infrastructure operates, and vice versa. At least some communications capacity is then preserved. But a single, shared telecommunications infrastructure would expose television to a complete blackout in the event of a telecommunications failure.

One approach would be to recognise radio as the alternative infrastructure. This has some appeal as radio has greater utility in emergency situations than television. In any case, it is likely that a resilience strategy would be needed to ensure that both telecommunications and broadcasting were not compromised in a variety of emergency scenarios. This strategy could include site hardening against natural and criminal events, improved local generation backup arrangements, and undergrounding of power infrastructure to key infrastructure sites. Clearly, this would add considerably to the expense of any migration, but could equally be regarded as an overdue investment in emergency management.

Network coverage

The final major technical consideration is whether online networks can reproduce free-to-air television broadcasting’s existing geographical reach. The Telstra, Optus and TPG mobile networks cover 99%, 98% and 96% of Australia’s population respectively. The nbn network covers effectively 100%, as satellite provides connection across the entire Australian geography that nbn’s fixed and fixed wireless networks do not reach.

We have already excluded the mobile networks, including 5G mobile networks, as a general option for migration of free-to-air television to the online environment.

The ubiquity of the nbn suggests that digital networks capable of carrying free-to-air television services are widely available. That is true for most of the population. However, there are residual issues that would need to be addressed in any migration scenario.

The key issue regarding nbn’s network is the satellite service. As noted, no unlimited plans are available on the Sky Muster satellite service. Sky Muster “Plus” services, which were launched in 2019, provide unmetered capacity for web browsing, videoconferencing and application updates (amongst other things), but this specifically excludes OTT video.

Satellite plans typically include a significant night-time allowance, which is exploited by applications such as Play On Cloud, which can be used to download selected videos from certain SVOD services overnight for daytime watching. However, the night-time allowance is still limited, and the overnight downloading of video requires some additional equipment.  But most importantly, such applications are not suitable for the real-time linear streaming that free-to-air television would generate if moved to the nbn.

If nbn satellite is excluded as an option, then nbn’s fixed and fixed wireless networks would have to be relied upon to provide coverage. This would be straightforward in urban areas where fixed network technologies can deliver both high speed and unlimited capacity plans, and the same is true in regional areas where nbn’s fixed and fixed wireless network technologies are deployed.

However, in any given regional location, there is no guarantee that the non-satellite nbn networks will be available everywhere that free-to-air television is currently available. Broadcast and nbn networks use different technology and are optimised for different use cases, and inevitably do not reproduce each other’s coverage. There will be areas with poor broadcast reception and good nbn coverage, and vice versa. The question of whether nbn connectivity was available to convey free-to-air television could only be determined on a locality-by-locality basis.

For this reason, a purely nbn-based migration would not be practicable for all current broadcast viewers, particularly those in regional areas. The only practicable solution would be to combine nbn delivery with something like an expanded VAST service, which currently used to deliver a selection of free-to-air television outside the regional broadcast licence areas. Broadcast satellite has great advantages over nbn satellite in this regard. The DTH satellite broadcast systems that underpin VAST are optimised for broadcast television and deliver it efficiently, unlike the nbn satellite that is optimised for two-way data exchanges. While it would be necessary to upgrade VAST capacity to provide the full range of existing free-to-air television services outside the nbn’s fixed network footprint, this can be done without any significant technical challenges. The primary challenge would be funding the cost of terminal equipment and additional DTH satellite capacity to carry all current SD and HD free-to-air channels.

Alternatively, there are areas where the nbn extends beyond regional broadcast areas where it might displace VAST. This, again, would have to be determined on a locality-by-locality basis, depending on the availability and cost of the different delivery platforms.

Another consideration is the role of regional broadcasters in this scheme. Regional broadcasters in the competitive equalisation markets created back in the 1980s act effectively as subsidiaries of the metropolitan networks from a content perspective. They do not stream their feeds, rather their metropolitan partners do. It is not clear what role regional broadcasters would play after a complete online migration when the metropolitan broadcasters would have national reach. Regional advertising slots could be maintained, as different ads could be injected at different nbn PoIs. But there are some significant implications for local content, particularly local news and current affairs in regional areas. Regional broadcasters are the mechanism that delivers the localism that currently exists, and there are no obvious incentives for metropolitan networks to maintain this local coverage in what would be a homogenised national market. Preservation of localism would almost certainly require policy intervention.

Alternatively, in a commercially-driven, partial migration, regional broadcasters would keep their role, but could look to the nbn as a delivery alternative in areas where traditional transmission costs were high, but nbn had deployed fixed or fixed wireless infrastructure.

Some transitional considerations

There are some transitional issues that would need to be considered, particularly in regional Australia. Nbn distribution of free-to-air television would allow for different advertising to be inserted into channels at each separate PoI.  There are many more PoIs than broadcast licence areas, so overall this would introduce more geographical granularity in television advertising. However, the effect varies across the country.

In metropolitan areas, there are generally several PoIs in each of the metropolitan television licence areas. This could allow for more geographical granularity for distribution of television advertising.

But in regional areas where there are small licence areas such as South Australia and northern Victoria, multiple regional licence areas could be subsumed into a single PoI area. Their licence areas would be effectively absorbed into neighbouring licence areas. While their traditional viewers would be unaffected, they would lose viewers who moved their television onto the nbn. This issue would be exacerbated if the metropolitan networks used their ownership of their content to bypass regional broadcasters to deliver services directly over the nbn. The transition from a system of licence areas to a new network geography would have significant implications for the future of regional broadcasters and local advertising, news and current affairs that policy-makers would need to address.

Finally, the sequence of migration from broadcast transmission to online delivery would need to be considered. The best approach from a cost saving perspective would be an “outside-in” approach, with regional viewers being migrated first (supplemented by an expanded VAST program) and the expensive regional transmission system being decommissioned. In contrast, metropolitan audiences are served much more efficiently from fewer station sites, and this second phase of migration could occur at a later stage when the value of VHF and UHF spectrum had risen. The economics of a hypothetical move of TV to nbn multicasting would look more attractive if the migration was phased in this fashion. Clearly, this approach would require significant viewer preparation and some policy intervention to ensure a smooth migration.

Issues in the migration of FTA broadcasting: viewers

The free-to-air industry has already migrated some of its viewing to the online environment in the form of streamed linear TV and on-demand catch-up TV. But the task of moving all free-to-air television (not just some) is complicated by the availability of Internet-connected television sets, variety of broadband access technologies, the disparate geographies of Australian communities, and the commercial model for broadband in Australia that includes payment for capacity, not just connection. Any and all of these factors can come into play when considering migrating any particular viewer. The importance of availability of broadcasting services means that each of these issues must be addressed in the case of each viewer. In fact, the challenge is not really to migrate free-to-air television to an online environment. Rather, it is how to migrate free-to-air viewers to the online environment. A survey of viewers was undertaken as part of this project in the last week of November 2020, and informs the discussion of viewer attitudes below. Details of the survey methodology and demographics are in the Appendix below.

Installed base of broadband connections and Internet TVs

The first consideration is whether a household can view online television at all. This requires both a broadband connection and an Internet-enabled television set.

It is difficult to arrive at a precise figure for household broadband penetration in Australia, as published connection numbers generally include small business connections. If these are included, household penetration is around 83%, but the actual number will clearly be lower. It is fair to say that the number of households without fixed broadband exceeds 20%, and this figure has not changed much in recent years. Some of these households will have access to mobile broadband though smartphones, connected tablets and Mi-Fi style devices that utilise mobile network capacity. A small number of fixed wireless connections, less than 100 thousand, are also active though this number is growing as the mobile network operators promote 5G fixed wireless as an alternative to nbn broadband.

It is clear then that there is a significant segment of households that lack the basic broadband connectivity to participate in a migration of free-to-air television to the online environment. The persistence of this segment despite the nbn’s expansion suggests that this is not just due to availability. Affordability will be an issue in some cases. It is also likely that a lack of interest plays a role, particularly amongst older Australians who are consistently less engaged with online activity. As it happens, this is the group who most frequently watch free-to-air television[3]. This is consistent with our survey data, which shows that only 20% of 25-34 year olds were “very concerned” about the prospect of relying on the Internet for free-to-air television, while a much higher 48% of 65+ respondents were.

Clearly, these unconnected viewers could only be migrated to online free-to-air viewing if they arranged a broadband connection for their home, but this would still leave these unconnected households with the monthly costs of connection to cover. Our survey data suggest that it is the monthly cost, not the connection cost, that is the focus of concern. Overall, only 14% of respondents said that the installation of a new broadband connection was a source of concern in the event free-to-air was only available online (this is close to the 17% of the survey respondents who did not have a fixed broadband connection). In contrast, 47% of all respondents said they were concerned they may have to pay for events they currently get for free. It is possible that some respondents thought that shifting free-to-air to online delivery might require them to pay for content, but this does not seem to be able to explain such a high level of concern about paying. Concern about the monthly cost of a broadband connection is the most likely explanation.

This concern would be mitigated if the free-to-air television industry adopted a Multicast solution for online delivery because the industry would carry the cost of the Multicast connection to the household. However, this would not cover the cost of the original nbn connection (if one was required).

In addition, capacity issues are a factor for a significant share of households. We asked our survey respondents whether they worry about their fixed broadband data allowance when making decisions to watch Internet video services, and 30% replied yes. Interestingly, this anxiety was significantly higher for younger age groups (46% of 18-24 year olds) than for older groups (only 18% for 65+ respondents). This may reflect higher levels of use of online video amongst younger people. 62% of 18-24 year olds reported that their households watched SVOD services, while only 20% of 65+ respondents did so. Also, 52% of 18-24 year olds reported that their households watched free-to-air catch-up services, and only 39% of 65+ respondents did so.

As expected, the penetration of Internet-connected television sets broadly reflects the penetration of broadband, but the issue of non-Internet sets is easier to manage. The survey results show that television sets are already widely connected to the Internet, with 69% of respondents’ main TV sets already connected, compared to 83% who reported having a fixed broadband connection. Of these Internet-connected television sets just over half were connected directly, while the other half were connected through another device such as a Chromecast device or a Telstra TV box.

The two top reasons for not being connected were “We’re happy with the broadcast TV programming that we’re already viewing”, and “TV [set] is too old”, which were both nominated by 42% of unconnected respondents. The latter response suggests that a significant number of respondents were unaware that most non-Internet TVs can in fact be connected relatively inexpensively by using an appropriate device e.g. a Chromecast device. But only 15% of unconnected respondents agreed that “It’s too complicated to connect a TV to our broadband connection”, suggesting that “technophobia” is an obstacle for relatively few.  The issue seems to be that these respondents are just not aware of the options, rather than regarding them as too complicated. Migration of television to the online environment would need to systematically address awareness of these options to encourage adoption, but the low cost of the solution would avoid any need for a significant subsidy.

Appendix: household survey methodology

Venture Insights’ household television survey was conducted in the last week of November 2020, with a sample size of 1003 respondents. The survey was conducted online.

  • 83% of respondents reported they had a fixed broadband connection at home (somewhat higher than our estimate of connected households), and 17% did not
  • 72% of respondents lived in metropolitan areas, and the other 28% lived in regional areas
  • 53% of respondents reported themselves as male, while 46% reported as female.

The respondents showed a variety of age groups, weighted somewhat towards the middle age groups.

The most common household income distribution was $50,000 to $99,999, followed by $25,000 to $49,999.

[1] ACCC, Internet activity report, June 2020. Downloaded at https://www.accc.gov.au/system/files/Internet%20Activity%20Report%20%28June%202020%29.pdf on 4 December 2020.

[2] Submissions can be downloaded at the Committee’s website, https://www.aph.gov.au/Parliamentary_Business/Committees/Senate/Finance_and_Public_Administration/Bushfirerecovery

[3] ACMA, “Communications Report 2018-19”, p.95.