Crystal ball gazing

"Where market predictions become critical is in competitive bid situations, such as bidding for telecoms licences, or equity in a company," Andre Wills, managing member of Africa Analysis, tells iWeek. Wills, who provides market perspective and predictions for selected clients, concedes that "on an international level, some research and analyst agencies are definitely more credible than others".

Recently, international analyst firm Juniper Research said it expects the global market for mobile TV to be worth $11.7 billion by 2010. When considering how much faith to put in such a bold prediction for a technology that has barely gotten off the ground, it`s worth looking at some of the areas where the forecasters have been wrong in the past.Local analysts BMI-Techknowledge, for instance, released statistics in July 2003 that predicted South African IT revenues would be at about R57 billion by 2006. At the time they were hovering around the R44 billion mark. If we fast-forward three years, to just a few weeks ago, BMI-T released its latest market sizing survey results and forecasts, with senior analyst Natalie Bryden announcing IT spend in 2006 would be about R49 billion - billion less than was forecast three years ago.

In BMI-T`s 2003 report, a statement alludes to some of the problems with accurately forecasting the IT market: "Exchange rate fluctuations have wreaked havoc on the economy between 2001 and 2003, with a corresponding impact on the IT market, which is highly sensitive to imported components, especially hardware."

TELECOMS EVEN HARDER TO GAUGE

On the telecommunications side, reliably predicting future trends can be even more difficult. With 11 years experience in the analyst business, Wills firstly points out that there are two kinds of predictions: financial analyst forecasts and market analyst forecasts.

In general, he says two large telecoms markets have been wrongly forecast: the growth in the mobile market, and the growth in the 3G market. "The single biggest issue in the mobile market was the introduction of prepaid. No analyst that I am aware of got the mobile market forecast correct after the introduction of prepaid," he says.

This underestimation can be contrasted with an overestimation of the growth in the 3G market, adds Wills, who remembers how a few years ago many analysts drove up the 3G market with predictions of significant customer migration to 3G through the adoption of new services, saying: "Telecoms operators were caught up in the feeding frenzy and paid very large sums of money for 3G licences, before it all came crashing down due to a lack of market performance."

ERP MISCALCULATION

Enterprise resource planning (ERP) is arguably another area where analysts have failed to foresee key developments. A Computerwire research paper from 2003 notes that since the refreshment cycle for enterprise-class software is about five to seven years, "with the big ERP push of 1997 to 2000, the expectation was that the core ERP market would be ripe for a major renewal program between now [2003] and 2005". It adds there is little evidence this expectation is being fulfilled because while vendors` maintenance revenue streams remain strong, licence income continues to weaken.

IT FORECASTS

"A lot of pressure is placed on research companies to be right in their predictions, which is fine. But what about the IT functions predictions? They predict certain benefits, and unlike the research houses, it is within their power to make their benefits predictions come true," comments Terry White, business and technology advisor at MarketWorks. Highlighting the importance of analysts` forecasts, and what can happen when those predictions are wrong, Craig Terblance - also a business and technology advisor at Marketworks - points to a Computerwire report published two months ago, just after ERP giant SAP released its second-quarter results. "Shares in the company dropped almost 8% on the Frankfurt exchange after licence revenue fell short of analysts` expectations," reads the report.

"Looking at Butler Group, Datamonitor and Computerwire predictions from 2001 or thereabouts with a view to assessing the extent to which they have come to pass, it seems to me the rapidity of uptake of the trend is often over-stated. So while the trend is seen to carry weight over time, the extent to which the trend has penetrated the market is often less than predicted," notes White. He says reports from 2004 predicting 52% of vendors will use subscriptions as their primary sales model during the next two years has clearly not yet come to pass. "But who`s to say it never will?" he asks.