SA?s mobile subscriber base has been routinely overstated by at least 20 percent, and sometimes more, over the last five years, according to research findings released Thursday.

According to "The State of the Mobile and Wireless Services industry in SA 2005", produced by World Wide Worx, the size of the local mobile subscriber market, widely assumed to have reached the 18 million mark in March, was in fact only around 14.5 million. The 20 million mark, thought to have been passed at the end of September, will not be breached until well into 2005.

"It is still an astonishingly successful industry, but that success can be undermined by over-hyping it," says Arthur Goldstuck, MD of World Wide Worx, who led the research project.

The study is the first of a six-phase World Wide Worx research project entitled "Mobility 2005", the broadest study yet of mobile and wireless technologies in SA.

The first three phases are sponsored by mobile network Cell C and a new mobile strategy think-tank, the Mobile Institute. The research is designed to produce a better understanding of the industry, and covers cellular, wireless and radio technologies, as well as mobile commerce.

The next three phases, sponsored by FNB and Sentech Pty. Ltd., will include market research among corporations, small and medium enterprises, and consumers.

The first phase of the study shows that there has been widespread controversy within the industry around the question of what constitutes "active customers".

Most networks use the definition of "Active 3", for subscribers who have generated activity on the network in the past three-month period. However, networks have been "fluid" in their choice of definition, hopping between Active 3, Active 6 (subscribers who have been active in the past 6 months), Active 7, and even Active 9.

"The networks are very honest about their churn rates, as well as their definitions of active customers, but industry observers and analysts tend to ignore these for the sake of talking up the numbers," says Goldstuck. "At the same time, though, we have also identified numerous other factors behind the over-counting -- by both networks and analysts."

Subscriber numbers are distorted because:

-- Pre-paid starter packs cause churn through "arbitrage" - buyers spot the discount between the cost of starter packs and the cost of recharging existing accounts. Once they use up the airtime on such starter packs, they discard them - but are still counted as active subscribers;

-- Active contract subscribers purchase new contracts when they spot a bargain in the price of a contract versus the combination of free phone and free gift, such as shopping vouchers or TV sets - they continue to use their existing contracts, but are counted twice by the networks;

-- The top echelon of prepaid customers tends to match the handset upgrade behavior of contract customers, who upgrade their phones every two years. These upper-income prepaid customers prefer to keep their old numbers, but take out new prepaid bundles every time they want an upgrade, and are again counted twice by the networks. As many as a million "new customers" a year may already be subscribers;

-- Analysis of churn patterns -- as many as 45 percent of prepaid customers change or drop networks every year -- indicate that many prepaid subscribers are counted on more than one network during their network hops;

-- Retailers believe that a high proportion of phones are being bought by foreign visitors, especially from other countries in Africa -- many tourists routinely purchase a handset with prepaid airtime when they arrive;

-- According to a recent World Wide Worx research report on Smart Cards, around 10 million new SIM cards for cell phones were expected to be issued in 2004 -- three times the amount of new subscribers expected by the networks. This alone indicates huge movement between and within networks, and casts significant doubt on how many of the new subscribers really are ?new?.

Interviews with representatives of the mobile networks -- all three mobile networks cooperated with the research -- suggest that these findings are not seen as all bad news for the industry.

Says Jonathan Newman, head of strategy at Cell C: "The number shows that there is in fact more growth to be had out of the market before it reaches saturation."

The total market potential continues to be estimated at between 26 million and 33 million mobile subscribers, depending on whether industry representatives are making conservative or aggressive forecasts.

"For the market to continue growing at its present rate depends on economic factors, such as SA maintaining the present Rand/Dollar environment, favorable inflation and economic growth, and market factors, such as the networks introducing very low denomination recharge vouchers," says Tristao Abro, head of Segment Marketing at Cell C. "Cell phone innovations per se will not drive growth in users, but are more about increasing penetration among existing subscribers," he adds. The results of phase two of the research, into mobile commerce in SA, will be released shortly.