A whole lot of people with open minds listen to her. They don't vote in lockstep with her suggestions (McCain and Hatch come to mind), but they do look into candidates she endorses. A substantial proportion of conservatives, after checking on those endorsed politicians, find that she was right about someone they overlooked. It's more "nuanced" (to use a word that so many Sarah-haters love) than changing their minds because of Sarah, but her endorsements result in huge jumps in polls and in votes. She takes good candidates who would have been ignored by early deciders because no one ever heard of them and then would have been ignored on election day because voters want the best among the front runners and her picks are so often too obscure to be considered competitive before receiving her endorsement.
The Washington Post tracks Palin endorsements - a sign that people care enough to attract clicks. This woman, who endorses far more relative unknowns than front runners, and VERY few incumbents, had a pretty good record in 2010: 33 wins, 10 primary losses, and 20 general election losses. That's just barely over 50%, but non-Palin endorsed challengers have about a 10% success rate against incumbents. We can't prove that they won because of her - they might as Palin haters claim just be good candidates who would have emerged from obscurity without her, but their polls jump almost immediately, and their election outcomes are startlingly positive.
If palin were to endorse early (which she doesn’t) and if she were to campaign for those she endorses (I don’t think she does) then it would be easier to make a case for the impact of her endorsements. But, jumping in at the eleventh hour with nothing but an announcement does not create a model where one can say her action changed minds and put people over the top.