I agree that the $100,000 is very small for election ad spend. However, I think Cambridge Analytica's alleged ability to target voters should not be discounted: it's possible that they've optimized for specific enough characteristics that effectively make Facebook's ad cost algorithm(s) think that certain campaigns are not as valuable as Cambridge Analytica does.

I'm linking a few general things about Cambridge Analytica here, but nothing specific to my hypothesis: [0] https://en.wikipedia.org/wiki/Cambridge_Analytica [1] https://cambridgeanalytica.org/ [2] https://twitter.com/CamAnalytica [3] https://www.nytimes.com/2017/03/06/us/politics/cambridge-ana... [4] http://www.techrepublic.com/article/cambridge-analytica-the-... [5] https://holidata.cambridgeanalytica.org/

I don't think that one ad campaign on Facebook swung the election, but I also know that there is a difference between dumb money and smart money in a lot of domains, and ads could just very well be one of those things.