One of the most disconcerting aspects of the business is the way vendors treat forecasts made by research houses of parameters such as market size three and five years ahead as if they were gospel when in reality, even extimates of what happened last year and the year before are subject to huge margins of error: La Jolla, California-based Computer Intelligence InfoCorp has owned up to a huge 15.7% underestimate in its assessment of the number of personal computers shipped in 1993 – it now reckons that 43.8m were shipped, an enormous 6.9m more than it originally suggested; it has also revised its forecast for this year upwards by 15.2% to 49.4m machines; such huge discrepancies are understandable when one considers how difficult it is to collect accurate data, but it is an awful warning that everybody should take all such numbers with a large pinch of salt – after all 15% represents the difference between comfortable profits and bankruptcy.
