On this page
How we calculate broadband speeds
The ACCC runs the Measuring Broadband Australia program.
This program relies on volunteer households across Australia. These households receive a Whitebox that tests the speeds of their fixed-line broadband services. The Whitebox won't monitor or record any personal information or browsing history.
The speeds we report are direct measures of what our volunteers can achieve over their connections when downloading and uploading data.
Grouping of results
We run hundreds of thousands of speed tests and then calculate the average speeds achieved by different groups.
Group by type of connection
We group results by volunteer households on the NBN and volunteer households still on ADSL broadband.
Group by ISP and retail plan
We also group results by internet service provider (ISP). We narrow this further to specific retail plans when there are enough volunteers to do so.
Some volunteers will experience better or worse speeds than the average, even if they are on the same retail plan.
The results show how close, on average, the ISP gets to the full speed of its retail plans. This is why these results are presented as percentages.
The more often volunteer households can achieve speeds closer to the full speed, the higher the Measuring Broadband Australia speeds for the ISP will be.
Speeds in busy times and over the whole day
We highlight speeds for the busy evening hours between 7 pm and 11 pm. This time period reflects the speeds consumers are likely to achieve when most consumers are using their service.
We also show speed results for the 'busiest hour’ of testing as well as for the whole day.
The main factor that affects these speed results is the amount of capacity that an ISP has installed in its network, including capacity it has purchased over the NBN.
The busiest hour speed measure roughly indicates the lowest speeds recorded over a thirty day test period during the busiest time for each ISP.
To ensure this measure does not include out of the ordinary events that may distort the results, we use the fifth busiest hour over the entire test period.
A higher busiest hour speed could mean that web applications are more likely to run to an acceptable quality even when the ISP network is under significant stress. However, this is a single hourly measure with fewer test results to calculate it, so the margin of error will be a little higher than other results.
The whole day
The average speeds achieved across the whole day will typically be higher than the busy hour speeds, as ISPs may need to ration capacity across their networks due to high demand during the busy hours.
The greater the difference between the average speeds over the whole day and the busy hour speeds, the more likely it is that the ISP relies on rationing its network capacity during high demand periods rather than delivering full speeds to its customers.
The ISP speed dials on the Broadband performance data page show a shaded area with a percentage figure directly below.
The shaded area indicates the boost to average speed if the services not able to achieve maximum plan speeds were excluded from the overall results. This boost is calculated by excluding test results made over poorer quality connections to the NBN. That is, we remove results for volunteer households who never come close to achieving the full speeds of their selected plan at any time of the day.
Underperforming services impact the overall results. For example, some customers who are on an FTTN connection with a plan speed of 50 megabits per second or faster presently receive much less than the full speed of their plan even outside the busy hours. This can be due to a localised problem at the customer’s house, which can be relatively easy to resolve, or another problem requiring more complex work in the NBN network.
How an ISP can boost its results
An ISP can improve its Measuring Broadband Australia results by:
- resolving connection problems for its customers
- offering an affected customer an alternative plan with a speed closer to what the customer’s connection can support (we would expect this to be at a lower price).
We suggest affected customers speak with their ISP to see what can be done to:
- fix their connection, or
- compensate them for the reduced speeds that are available until the connection problem can be fixed.
If ISP1 records a busy hour download measure of 85 per cent and ISP2 records 80 per cent, this means, on average, ISP1 has download speeds that are 5 percentage points closer to full plan speed than ISP2 when viewed over all of the busy hours.
This would not necessarily mean that the download speeds of ISP1 and ISP2 are always 5 percentage points apart for each busy hour.
It could be that for a large number of the busy hours, the speeds are actually quite similar, while in the remaining busy hours the speeds offered by ISP1 are significantly faster than those offered by ISP2. These may actually be the hours that a customer wants to use high speed applications on the NBN service.
The table presents a set of results for each busy hour. There are 4 busy hours in the test period to illustrate this:
Busy hour 1
Busy hour 2
Busy hour 3
Busy hour 4
Measuring Broadband Australia speed measure
In this example, the download speeds in the last busy hour for ISP1 are actually 20 percentage points higher than those of ISP2.
When this is averaged out over all 4 busy hours, the difference is only 5 percentage points.