Unlike this board, no one's telling P&G, "don't do any research, introduce whatever product you want"
Based on the rate of success, radio managers are brighter and better than those in many other fields. Procter & Gamble is renowned for having a new product success rate of nearly 50%, while competitors average less than half if this. But that means that half of P&Gs new products don't meet the sales and ROI goals set for them.
So, to be as good as the best in the world, more than half of all format changes would have to be successful. I'd say that the success rate is higher than that.
Also consider that to get that 50% success rate, P&G spends $2 billion a year in product research and $400 million in market research (all P&G stats come from the June, 2011 Harvard Business Review). A radio station generally would spend less than $100 k in format research and a music test... and in smaller markets, not even that.