It's based on the idea that most people are very very stupid. Most people who think that never include themselves, interestingly enough.
You should probably include yourself... I certainly include myself. None of us know as much as we think we know.
To really understand how useful such manipulation is, you have to think about it philosophically. How do you really know anything? You can do scientific experiments, and then calculate the statistical probability that bad luck gave you incorrect results (and so even then we still don't really know anything), but we don't have that luxury for most of what we know. Most of the time we're limited to simply accepting what other people tell us, and filtering it according to how plausible it sounds and how well it fits what we already know.
For example, you probably believe in the theory of relativity. Have you seen anything moving at near the speed of light, thus that you've witnessed that it doesn't work like Newtonian physics? Do you even know anyone who claims they have? Do you truly understand relativity, or do you just have a rough idea of how it works due to some examples of pool balls on a rubber sheet, and some explanations about time and space contraction that at least sound plausible? Do you understand those formulas and all of their implications, or do you just trust them because they're math? Do you even personally know anyone who truly understands relativity whom you know from experience wouldn't lie to you about it?
This is where most people, even intelligent people, are at when it comes to most issues. You may be quite intelligent, but what do you really know about some random political issue other than what you've heard on random news shows, read on Wikipedia, and perhaps found via a random web search? You may think you've studied an issue, but there are people who've spent decades studying economics. Have you learned something in your hour of web searching that they didn't learn in that decade?
We couldn't be elitist even if we wanted to. How do we know that those economics experts are really economics experts. They could just be some douche on Wikipedia linking to made-up articles written by made-up people with made-up credentials. Without some means to verify that our elites are actually elite, elitism can't work. So what do we do? Build up a web of trust, where we trust a friend who trusts one of his friends who trusts one of his friends who trusts that this economics expert is really an economics expert? Then wait for some clever person to find a way to subvert that?
The end result is that every issue has to be dumbed-down so that everyone can understand it, but in doing so, it's possible to make any side of an issue sound like the correct side of the issue simply by choosing what information to include and what information to exclude. The obvious solution to that is for the opposition to publish a rebuttal, explaining what was omitted in order to better inform the voters.
It may not be possible to fool everyone, but they only have to fool 50%, and since 50% are of below-average intelligence, a little manipulation of search results can go a long way.