From Common Stocks and Uncommon Profits
The impact of this sort of thing on investment can hardly be over-stated. The cost of this type of research is becoming so great that the corporation which fails to handle it wisely from a commercial standpoint may stagger under a crushing burden of operating expense. Furthermore, there is no quick and easy yardstick for either management or the investor to measure the profitability of research. Just as even the ablest professional baseball player cannot expect to get a hit much more often than one out of every three times he comes to bat, so a sizable number of research projects, governed merely by the law of averages, are bound to produce nothing profitable at all. Furthermore, by pure chance, an abnormal number of such unprofitable projects may happen to be bunched together in one particular span of time in even the best-run commercial laboratory. Finally, it is apt to take from seven to eleven years from the time a project is first conceived until it has a significant favor-able effect on corporate earnings. Therefore,even the most profitable of research projects is pretty sure to be a financial drain before it eventually adds to the stockholder's profit.
But if the cost of poorly organized research is both high and hard to detect, the cost of too little research may be even higher.