Inaccurate data can sometimes be better than no data at all because they at least increase the probability that you are headed in the right direction. When I worked in the long term planning section of a telephone company we had to rely on subscriber forecasts which were little better than projection lines drawn on graph paper. It all evened out in the long run.