There is nothing new in this post. But the issue keeps coming up.
I have written many posts about this before.
My last one was Advice to undergrads giving research talks.
Perhaps the following basic point gets lost in all the suggestions.
Keep it simple.
In almost every situation, most of your audience knows very little about your specific research topic. In some cases, they know virtually nothing about your actual research field.
Thus, you need to cut out almost all the technical details and give plenty of background and motivation.
But again, you need to realistic. Don't kid yourself that in 5 minutes you are going to teach them Density Functional Theory or Two-dimensional NMR spectroscopy.
Have modest goals.
Teach the audience some interesting science.
Convince them that your topic/field is interesting and important.
Show them you have achieved something concrete and interesting.
Don't bore people.
But perhaps, these are ambitious goals because most talks I hear don't achieve them!