Speed rankings and Australia’s fixed-broadband performance

With the rollout of the NBN now complete, Australia is among a small group of large countries with universal, affordable broadband service availability. Today almost every home and business in Australia can order a service of at least 25Mbps and over 8.4 million already have.

Significant public investment has been made in the rollout of the NBN and it is important the performance of Australia’s fixed broadband infrastructure is known and comparable to our international peers.

However many of the common international speed rankings or comparisons have limitations – such as their methodology, data availability, sampling bias and an absence of ‘like for like’ country comparisons. As a result many of these existing measures do not provide an accurate picture of Australia’s relative broadband performance.

Different Rankings have different methodologies

Most international comparisons focus on comparing countries based on ‘average experienced download speed’. For instance the well-known Ookla Speedtest publishes its ‘Speedtest Global Index’ each month which includes a result for median download speed for fixed broadband per country based on the tests undertaken via speedtest.net in the previous month.

While the Ookla test methodology uses local servers in most countries enabling some consistency in measurement – it has several limitations. Tests are necessarily impacted by a user’s in-home equipment, their subscription speed, and even the strength of their Wi – Fi connection. In addition the sample is limited to users who initiate tests via Ookla’s testing website and may have specific biases, such as lower than average or higher than average speeds.

Most critically Ookla doesn’t provide a like for like comparison - it considers one metric alone – the median speed experienced by users who click on its website. This means that countries like Paraguay can score higher than Australia, despite have much lower availability and take up of broadband. Speed doesn’t mean much if you can’t actually get connected.

Ookla’s global Index also only requires a very small number of unique results – 300 – for a country to be ranked which can skew the results.

The OECD speed rankings methodology does more closely consider like for like comparisons – by factoring in subscribers per capita. However when it comes to measuring speeds it uses the speeds advertised by retailers rather than actual speeds of the network. It also skews heavily against Australia in terms of measuring fibre penetration as it only considers fibre to the premise connections (FTTP) and fibre to the building (FTTB) as fibre – this is despite the fact that Fibre to the Curb (FTTC) and Fibre to the Node (FTTN) connections can deliver similar speeds.

Some key issues and challenges

There are a range of other biases which can impact these online speed rankings. This include:

• Population Density: these tests favour countries with small highly concentrated population centres like Korea and Japan over those with more dispersed populations like Australia.
• Geography: The sheer physical size and terrain of a country can have a huge impact on its ranking. This is one of the reasons we see small ‘City State’ nations like the Netherlands, Singapore and Luxembourg perform well.
• Population Size: It is far easier to achieve a high take up of broadband services per capita in a small population.

What are we actually trying to measure?

So these rankings almost always do not tell the full story - they don’t properly take account of the many factors that actually influence the quality of broadband: take-up, network capability, population density, geography, market structure and so on. So we need to consider what we are actually measuring. Important factors to consider are:

• The Availability of or access to a broadband connection;
• The take up of broadband services;  
• Speed – in terms of both upload and download, but also how a network performs during both peak and off peak times;
• Technology type; and
• Fixed versus mobile connection.

Key Features of Australia’s Position

Our position is characterised by both ubiquitous access to and a high take up of broadband.

Australia is one of only five comparable countries where at least 99.9 per cent of the population has fixed broadband access, with near universal coverage. The proportion of households with a fixed broadband connection increased to 90.2% in 2021 (up from 88% in 2020).

Despite what we often hear Australia is not lagging behind on speed. 76% of NBN subscriptions are for services of 50Mbps or greater and 72% of Australian premises in the NBN fixed-line footprint can access speeds of 100Mbps or greater. Today 42% of the fixed line network is able to achieve speeds of up to 1 Gbps. What’s more under the current $4.5 billion network improvements 75% of the fixed line network will be able to achieve speeds of 1 Gbps by 2023.

And our broadband services are also affordable. In September 2021, NBN Co published a report commissioned from Accenture comparing retail pricing in real terms and found Australia has the sixth most affordable broadband services compared to 12 other OECD countries it sampled. Broadband has become more affordable and accessible during the NBN rollout, with ABS data showing that communications prices have fallen by over 26 per cent between 2013 and 2021.

Further Information

As you can see meaningful comparisons of broadband speeds are not straightforward. That is why my department’s Bureau of Communications, Arts and Regional Research (BCARR) has undertaken detailed analysis in this space – and is working to provide a more accurate and robust assessment of Australia’s fixed broadband performance.

BCARR has published a dashboard on Australia’s fixed line broadband performance measured with a group of comparable countries.

Further background information on the work and previous releases can be found at this link.