Alongside Melbourne and Sydney, Perth is the third Australian capital to have rival 4G LTE services from both Telstra and Optus. When it comes to performance, there's actually a clear winner, but the main lesson continues to be that 4G services vary hugely in the performance they can offer -- even within the radius of a block.
This shouldn't be news, of course. Mobile networks are inherently variable; even moving from one side of the MCG to the other made a massive difference to the speeds and networks I got in a pre-Grand Final test. Yet many people remain firmly convinced that a given network will always be a better choice than the other in speed terms. The reality is rather more complex.
Perth provides an excellent example of how much performance can vary even with very slight changes in location (as well as a test bed for more complex 4G projects). In my hotel room in the Perth CBD, I was utterly unable to scare up a 4G signal from Telstra, and the 3G speeds it offered were low. They weren't so low as to be unusable, but there was a bigger problem: the signal would regularly drop out altogether. A slow network is annoying, but a network that drops out on a continuous basis is arguably worse. Optus' numbers were also fairly low, but it did achieve a 4G connection.
Assuming that demand might be a factor and that the building's architecture might be an issue, I performed the same tests very early the next morning (like many Perth visitors, I largely worked on east coast time, which meant I was hard at work at 4am). The speeds for Telstra were slightly better, but the problem remained. Optus' numbers were somewhat higher than the night before.
I'd have been convinced that Telstra 4G wasn't happening at all in this corner of town, but when I went downstairs after checking out and walked one block, Telstra happily offered me 4G. The download speeds were still healthily outgunned by Optus (also on 4G), but Telstra won out for upload speeds.
Having walked another block or two to Esplanade station, I decided to repeat the test again, since there wasn't yet a clear pattern. Telstra performed even better, again winning the upload speed test and performing at an extremely similar level for downloads (16.73Mbps versus Optus 17.48Mbps).
So my CBD experience suggested that Telstra performed well if you could get a 4G signal, but that was almost impossible to predict. It was a different story at Perth airport on a Friday afternoon. The airport was filled with fly-in fly-out mine workers, and the Telstra network was struggling to cope. There was no sign of 4G signal at all, and the 3G performance was incredibly slow. How slow? Optus' connection was 90 times faster for downloads than Telstra's.
Here's the full set of results. I've indicated where Telstra defaulted back to a 3G network rather than offering 4G; for all other tests, a 4G connection was achieved. Reminder: being on a 4G network is not in itself a guarantee of better performance; in some of my previous testing Telstra 3G has managed to perform better than Optus 4G in the same location. However, that wasn't the case here. For ping times, lower numbers are better; for upload and download speeds, higher numbers are better. Tests were performed using the Speedtest.net app, averaged over three results.
|Provider||Ping (ms)||Download (Mbps)||Upload (Mbps)|
|Hotel PM: Telstra (3G)||218||1.25||0.36|
|Hotel PM: Optus||92||4.00||3.93|
|Hotel AM: Telstra (3G)||118||1.94||3.74|
|Hotel AM: Optus||125||16.47||13.48|
|CBD AM: Telstra||77||3.61||19.31|
|CBD AM: Optus||125||18.30||14.03|
|Rail AM: Telstra||104||16.73||9.42|
|Rail AM: Optus||126||17.48||7.23|
|Airport PM: Telstra (3G)||329||0.17||0.24|
|Airport PM: Optus||75||15.49||5.02|
Is there an explanation for this evident difference? Very possibly. Several Perth residents I chatted with while on this trip mentioned that the Telstra network appears to be struggling to cope right now. Fly-in fly-out workers are potentially a major factor here. Telstra's coverage of mining and resources sites is, by most accounts, far broader than Optus'. If you're working in those areas, a Next G phone is likely to be an essential requirement. But that means there's a high concentration of Telstra phones in the state, and virtually all those workers will pass through Perth regularly. When they do (as at the airport), the result isn't pretty.
None of those workers are likely to switch to Optus in the short term. If I was living in Perth, not travelling out into the regions and concerned with data speeds, Optus would look like a much faster and more reliable choice. But if I'd stayed in a different hotel, my conclusions might have varied again. Right now, I'd suggest the biggest priority for Telstra in Perth is improving airport coverage: that's a big potential audience getting done over for coverage.
Lifehacker Australia editor Angus Kidman isn't prepared to pay for domestic in-room broadband anymore. His Road Worrier column, looking at technology and organising tips for travellers, appears each week on Lifehacker.