Wired on Saturday: It is often terrifying how quickly one has to face the question of scaling in a successful internet business. Audiences can leap from a few hundred to many hundreds of thousand in a matter of weeks.

A hobby website run on a spare computer either has to be forgone in the face of popularity, or fed many more servers, reared and cared for in proper "server farms" - those endless shelves of humming, blinking pizza-box shaped computers that fill chilled rooms in all the developing world's major cities.

But there's more than just the simple technological scaleability of a website.

And, even if you have enough money to pay for all the servers in the world, there's still many ways that a web business can become too popular for its own good, and for the internet as a whole.

Is Yahoo! too popular? The company clearly doesn't think so, as it has determinedly continued to buy up small, bright websites in an attempt, at least in part, to transfer some of their celebrity status online to its own somewhat mundane branding.

Still, those sites - sites that we've mentioned before, such as photography site Flickr, social bookmarking site del.icio.us, and calendaring site upcoming.org - are tiny compared to the audience of Yahoo!

The prices paid by Yahoo! for them dwarf the investment in hardware made by the start-ups but clearly the intention is to scale up these products to a mass market.

Has it been a success? Only Flickr has been owned by Yahoo! for long enough to be able to make any decision about this one way or the other.

It has definitely passed through strained moments, as the service slowed or stopped entirely under waves of new members.

Internally, the challenges of scaling up a complex, high-bandwidth site are said to be considerable. But the experience itself hasn't changed.

The site doesn't feel "spoilt" by being popular.

Del.icio.us was bought by Yahoo! last week, and seems to be headed for a choppier ride.

The purchase - said to be valued in the tens of millions - unhappily coincided with several days of "unscheduled interruption in service", as a hardware failure occurred while the staff were on the other side of the country, finalising the deal.

That said, del.icio.us is a much simpler site in many ways to Flickr. From a purely technical standpoint, scaling up to many users should not be an issue.

Except, talking informally to del.icio.us users, one realises that one of the services that del.icio.us provides is as a communal resource.

The site offers, for instance, a list of the most bookmarked new links among its audience. Many of its users read this as a "Top 10" guide to potentially intriguing new websites.

But looking at that "Top 10", it's clear that del.icio.us's community is relatively narrow: a "pantone to hexadecimal converter", a guide to "practical programming", and several programming tutorials head the current list.

Sites like Flickr, and blogging site MySpace (a recent purchase by Rupert Murdoch's News International media empire) were already appealing to a wide range of people before they were bought.

But more importantly, few of their features actively relied on having a narrow demographic, as del.icio.us's popular list does.

Back in September, one of the early adopters of del.icio.us openly worried about "the del.icio.us community being diluted with non-geeky type people".

The sentence was picked apart as being elitist, but the underlying point was a practical one.

The site just wouldn't, under its design, work as well when its popular list no longer pointed to the quirky sites the majority of its users were looking for.

If you want a Top 10 list for all of us, you can go to Yahoo! itself.

Del.icio.us was bought for its ability to serve its audience in a more individualistic way: if all Yahoo! got from scaling it up was Yahoo! again, then it would have wasted its money.

Now, as it happens, I think that del.icio.us can make the transition but it'll involve a subtle rethinking of the site, not just throwing more computer servers at the problem.

A wider criticism might be that Yahoo! has spent millions on a site that its employees and nearby West Coast circle of geeks liked, rather than a site that was guaranteed to appeal to everyone, and that Yahoo! didn't understand that they were buying a narrow community because they were members of that community themselves.

The insularity of California's Silicon Valley geek culture is something that has been noted before, many times, and over many decades.

As a byproduct of simple geographical and sociological closeness, it doesn't seem to have done the success of Silicon Valley companies that much harm.

But perhaps that is more luck on the Valley's part. In the reception to Google's Mountain View offices, a projector famously outputs samples of the search queries being typed at that moment into the search engine.

Hypnotised visitors watch as (suitably worksafe) searches like "engine parts" and "quick cure for sore throat" float by.

Curious, I made a rough estimate of how many of the queries were of irredeemably geeky topics - the sort of subjects that appear in the del.icio.us popularity list.

It was about one in 10. Google remains, after all these years, a geeky search choice and that hasn't affected its multibillion bottom line.

Narrow communities certainly don't preclude scaling, but they may leave a weakness for others, who consciously appeal to wider groups, to leave the Valley standing.