In an ideal world, carbon offsetting would work something like this: if there was a company that manufactured, for example, heavy machinery that would, in a given fiscal year, end up producing several thousand metric tons of carbon dioxide beyond the mandated cap in order to remain profitable, then the company would pay a certain amount of money to cover every ton they would have to go over the cap, and that money would be invested in renewable energy. Eventually, the renewable energy produced would exactly offset the amount of carbon the company went over the cap, and everyone would go home happy.
However, we do not live in an ideal world, and more often than not carbon offsetting does not work so perfectly. The problem is that once the money has changed hands, the company in question is entirely off the hook, regardless of whether or not it actually achieves what it was intended to achieve. For example, if the money from the hypothetical company above went to help fund a wind farm that spent more of its money trying to keep up the appearance that unrealistic goals were be already being met rather than using it to actually make the facility more efficient, the company would still be credited for the tons they paid for, whether or not any reduction in emissions actually took place.
In order to make carbon offsetting a viable tool for tackling global warming, there must be rigorous regulation of the projects it creates and accountability on the part of those who invest in it in. As long as companies face no real consequences whether or not really make up for the amount they go over the cap, there is no reason for them to put in more than a minimal effort to solve the problems of climate change.