You don't (really) have to cover what you promised

Imagine you propose an awesome talk idea and it gets accepted.

Later, when actually trying to create it, you realize your idea doesn't work so well in the harsh light of the projector. Perhaps you don't have nearly enough material to cover. More likely, your proposal has bitten off far more than you can chew in forty minutes. What can you do?

Fortunately, you're not really in a tough spot.

First of all, no one expects you to deliver exactly the talk you promise in your proposal. You've got some wiggle room if you want to cover (or cut) things in a way that doesn't match your original description.

This works fine if you need to make minor changes to your talk, but what if you've got deeper issues? What can you do if there's just no way this talk is going to work?

Even here, you still have an out. I've contacted organizers on three different occasions asking for if I can change my topic, and all have granted it without complaint. Even when the new talk was wildly different (from a soft-skills talk to full-on live-coding, for example). My theory is that once you're through the selection process, the organizers are confident you will deliver something of quality, and are willing to trust you.

Even so, if you do ask to change topics, don't forget a little salesmanship for the new talk you're swapping in. Set them at ease by conveying that this substitute talk will be as good or better as the one they originally accepted.