When is 'free' really free?

Judging by the number of e-mails I received in reply to last week's column it is clear that the free PC and free Internet access topic is, well, topical. To top it all, this morning's Business Day reports that a "Free Internet service is on the cards". Apparently the service, based on UK company X-Stream's software, will be launched next month in sunny South Africa. It will derive all its revenue from highly targeted advertising, based on detailed user profile information.

The Internet is, after all, a catalyst for revolutionary change.

A number of the e-mails I received drew my attention to the targeted advertising (or direct marketing) business model for free PCs, à la the Free-PC.com and shopss.com model. (My column last week, by the way, was focused on the alternative model, ie PC leasing). The advertising model has come under considerable attack, ranging from ethical concerns ("invasion of privacy") to practical doubts ("it just will not work, at least not with only so-and-so number of initial users" and "so many of the schemes in the US have already died a natural death"). This is particularly interesting, considering the local launch of X-Stream will rely on a similar business model.

Robert Cringely (http://www.pbs.org/cringely/pulpit/pulpit19990211.html) goes further with the cry of "foul play", by saying that this sort of scheme is actually nothing short of a scam. His analysis is that Free-PC.com is only offering 10 000 users free PCs, which alone will attract insufficient advertising revenues to pay for the service. The real motivation is to obtain the detailed profile information of hundreds of thousands of applicants, most of whom will not "qualify" to receive a free PC, but whose information will in any case be used for the direct marketing programme. Free-PC.com is also very rapidly (virtually instantaneously) creating a household brand name for itself, which could be leveraged in numerous other ways.

Whether or not this is unethical, it is certainly not illegal. But what is interesting is the question of whether it could work in SA. The global trend towards direct marketing spending has barely taken off here, with traditional advertising still enjoying the vast majority of total expenditure. However, if a scheme such as this takes off on a large scale, it could serve to significantly accelerate the conversion rate. And the Internet is, after all, a catalyst for revolutionary change.

What makes the local X-Stream project just that bit more feasible, is the use of Telkom SAIX's Internet backbone. This allows a rollout of an additional, say, 100 000 users with no additional network infrastructure. Or so the Business Day reports: "... as the user base grows the network can be expanded without capital investment." This argument sounds just a tad flawed to me, since clearly someone (ie SAIX) would need to do something special to accommodate the additional network loading, and there are support costs to add in as well.

These costs were put into context by one of the e-mails I received from an ISP owner in response to last week's column. After factoring the cost of international bandwidth, this fellow reckons, along with other costs, an additional R30 million a year per 100 000 additional users would be required. And satellite bandwidth would apparently not be much cheaper.

As a quick sanity check I looked at BMI-T's Internet Access report, to see what the revenue-to-subscriber number ratio is for dial-up access services. Assuming the dial-up ISP market is operating at a fairly thin gross margin, R30 million could be a very conservative estimate of incremental costs! Perhaps R60 million to R90 million would be required.

Could this be raised from the direct marketing business model alone? Quite possibly, considering the traditional advertising industry is worth close to R10 billion a year, and direct marketing perhaps another R2 billion on top of that. Even more importantly, the direct marketing sector is set to explode locally, as has already happened elsewhere in the world. Still, to achieve a 0.5% to 1% share of the total pie would take something very special. Then again, "very special" is something that the Internet has proven itself to be.

Should it work, the net effect on the existing ISP market would, at first glance, be disastrous. What would have to happen, in reality, is for ISPs to differentiate their services offerings based on quality of service. And few dial-up users will disagree that there is a difference between slow access and even slower access. In the end it would probably still boil down to the old adage that there is no such thing as a free lunch!

As always, I would be interested to hear your comments. E-mail me at brian@bmi-t.co.za and share whether you think free Internet access could work in South Africa.