There's a power law in startup success, and people who seriously study charity (such as GiveWell or Peter Singer) say there's a similar phenomenon in philanthropy: the best charities are often more than 100x as effective, per dollar, than average. One popular example: the Seeing Eye foundation spends more than $50,000 to train one Guide Dog[0], GiveWell estimates that the Against Malaria Foundation can save more than 10 lives with the same amount of money. So it's extremely unlikely that donating more will outweigh the impact of donating to the right charities.
[0]: http://www.guidedog.org/content.aspx?id=564#trainingcost