In part 1 I explained my recent switch from O2 to Vodafone, and why I made the move. This week, my flatmate Dave and I have been comparing network speeds on our iPhones, his on O2 and mine now on Vodafone.

It's a somewhat unscientific test, but the results are nonetheless interesting. We've been doing regular speed tests at various times of day in our flat in Battersea. Clearly this means our results are only relevant for one specific area, but hopefully they give some indication of the strength of the two networks in London. We have a good strong 3G connection on both networks. We've done at least 3 tests on each occasion, and measured 3 things:

• Download speed - how fast does data get to my phone? For example when loading a webpage or getting the latest data in an application.
• Upload speed - how fast does your request for information reach the server? How long does it take to send information out, for example when sending an email.
• Ping time - a very rough indication of 'round trip time'. In other words, how long does it take to send a message and get a response.

In each case the server that the phone is talking to has a big effect, so the absolute numbers should be taken with a pinch of salt. But the key is that we used the same application on both phones, so the comparison between the 2 is valid. We used Speed Check mostly because we're cheap, and it was free!

Here's the download speeds:

You can see that they both vary quite a lot, and there's not particularly an obvious winner. Vodafone seems to be in the lead a little, and that's verified by the averages; O2's average download speed was 1083 kbit/sec to Vodafone's 1425 kbit/sec. That's 31.5% better on Vodafone. They also had a better peak value (2881 to 2363) and a less bad minimum value (368 to 145).

What about uploads?

Here there's a clear winner, and it's Vodafone. Their average (317 kbits/sec) is not only 53% higher than O2's (206 kbit/sec) but it's higer than O2's peak of 284 kbits/sec.

Now for the ping...

Here you can see a more even match again, with both networks returning some worryingly high ping times occasionally. The maximum and minimum values are similar across networks, but once again Vodafone's average of 2.61 secs beats O2's 3.05 secs. That's a 17% improvement.

So if you take all the averages and average them together (still with me?) you get a figure that says Vodafone beat O2 by 34%. Conclusive then, right? Well maybe, but there's more to it than that.

Vodafone clearly provide a faster network service where we live. But both networks do fine, and in the real world I don't think these differences are so noticeable - remember that Vodafone put in some pretty shoddy results sometimes as well as O2. The key thing that these figures don't show is times where we couldn't get a connection at all. Dave had this on O2 on 3 occasions, whereas I had 4 such problems on Vodafone. Granted, trying again a few seconds later has always sorted it out, but if you factor those wasted seconds into the numbers above, the results might differ. Would it change the balance between the networks? Probably not by a lot.

There's other factors to consider too, like available network services and coverage. In part 3 I'll finish off by considering some of these.

Oh, and if you want all the gory detail, below is a spreadsheet of the full results.

I've had an iPhone since they first came out in the UK. I got a 3G, and then a 3GS. The one thing that hadn't changed all along was the network; I had no choice but O2. Problem was, while the service started off pretty good, the speed and connection reliability seemed to be getting worse and worse. It seems that the popularity of the iPhone, and the amount of data iPhone users tend to consume, meant that O2's network was grinding to a halt.

Once O2's iPhone exclusivity deal ended, I was able to get my iPhone unlocked, and with my contract ending this month I seriously considered moving network. I wasn't too sure though - would any other network really be any better? Perhaps the final straw came in December, when O2 CEO Ronan Dunne apologised for the problems they were having and said that they were working hard to fix the problem. Seemed like a good statement, but my issue was with this:

Dunne said that users in and around London should have already noticed improvements during the month of December

The problem was, for me December was the worst it had ever been. I had whole days where I couldn't get a connection at all! So that was it; I went looking for a new network. I needed someone who'd give me a large amount of data usage (I usually use between 300 and 400MB per month), a fair few texts (about 300 is plenty for me) and a moderate amount of calls (200 minitues as a minimum). Having paid off my phone subsidy with my O2 contract, I wanted to pay less, too.

Vodafone's £20 SIM only tariff fit the bill (unlimited SMS, 500MB data, and 600 mins calls - or 900 mins if you sign up for a year). And of course, when you switch network, you should never accept what you're given - always look for special offers or cashback. I used Quidco to get a little extra back in return for signing up.

It's worth noting that if I'd stayed with O2, I would of course have been able to pay less. I was on a £35 a month iPhone tarrif, and if I'd have signed for another year (or 18 months?) without getting a new phone, I'm sure they'd have reduced that. Their website offers 'up to' £150 bill credit (which works out as £270 over a year, compared to £240 with my Vodafone SIM before any cashback from Quidco), though they offered to halve Dave's £30/month contract to £15 (but that doesn't have enough minutes & texts for me). Your milage may vary...

But with a desire to switch and a confidence that I certainly wasn't losing cash by doing so, I made the leap. So, with a new SIM card I'm happily enjoying life on Vodafone. The question is: is it actually any better? I set out to put some numbers to it all, with the help of my flatmate Dave, who still has his iPhone on O2. We've spent the last week speed testing the networks regularly, and the results are interesting. More in part 2...