Britain's first next-generation mobile phone network, 3, launches next month. So it seems like a good time to look at the lessons from the Japanese market, where 3G has been available for 18 months, and ask when we'll really be enjoying the benefits of 3G in the UK.

At the end of last week, the two leading Japanese mobile operators announced their subscriber numbers to the end of last year. NTT DoCoMo, the market leader in providing advanced mobile services on a 2G network (today's technology), has a 58% market share in Japan. At the end of last year, it had 42.8m subscribers, of whom 36.2m use its iMode interactive services. But DoCoMo, a significant shareholder in 3, has managed to lure only 152,000 subscribers to its 3G service, called Foma, which launched in mid-2001.

KDDI, the number two player, has a different story to tell. Despite launching six months after its bigger rival, KDDI had 4.76m subscribers - and at the end of January, it passed the 5m mark.

Furthermore, KDDI was more successful in getting users to pay than DoCoMo. Average revenue per user per month (ARPU), the number that pretty much determines the profitability of a phone company, was 9,240 yen (around £50) for KDDI and only 7,750 yen for DoCoMo.

Additionally, ARPUs are dropping for both KDDI and DoCoMo. KDDI saw a 4% drop in the quarter to December, and DoCoMo experienced a 25% drop over the year.

What lessons can we draw from Japan?

For one thing, first movers don't always win. By winning market share quickly, you build incentives such as cheaper calls to other users of your network.

This should have given DoCoMo the advantage. It launched its 3G network first and had a vast iMode customer base. It was far and away the world leader in advanced, interactive wireless phone services. So how did KDDI overcome DoCoMo's first-mover advantage?

First, KDDI launched using a technology that worked with existing 2G masts and antennae. This meant that from day one, it provided almost nationwide coverage. Customers moving to KDDI's 3G service lost none of the benefits of 2.5G. DoCoMo eschewed backwards compatibility in favour of a new technology that had very limited coverage at launch. While this may have impressed the geeks, it won over few consumers who were more concerned with making calls.

Second, DoCoMo's original marketing proposition was built on speed. Speed (bandwidth, in other words) is an elusive benefit to all but geeky early adopters.

In practice, while DoCoMo's network was technically available first, KDDI's 3G service was beneficially available first. It provided all the "must-haves" people wanted from a phone service, such as 100% coverage.

The four UK operators must focus on selling the benefits of 3G rather than the new technology. Introducing new services that happen to require a new handset (and coincidentally run on their new network) will focus their minds on selling the end user benefits rather than the gizmos in the network.

The second lesson is that there is a novelty effect. The decline in ARPUs over time suggests that, after an initial burst of interest, users get bored with applications. This means operators will need to find compelling applications. To supply these new applications, the operators should turn to their users because, as the web has shown, nowhere is innovation richer than at the grassroots.

The success of SMS should be a useful precedent. SMS messaging only took off when the operators grudgingly opened their networks a few years ago. Now the bulk of SMS revenues are based not on over-priced information services, but on user-generated content. The SMS boom has created addicted and spendthrift users creatively using a technology with the most meagre specification.

The success of 3G will depend on usage. If the UK operators are to recoup the £22.5bn they spent on licences (and a similar investment on their networks), they need to get customers to use the service. The key factor for success will be user-driven and group-driven applications.

Enough of us need to use the system and this means the operators will strive for interoperability. As a customer, I will want to know that my 3D video mail can be received by my friends, and they can use their phones to vote on my new haircut.

So while 3 launches on 03/03/03, I suspect we'll be waiting until 06/06/06 before we really feel we're part of something new.

· Comments to online.feedback@theguardian.com