The first video is an interview of two poll workers expressing their concerns regarding the unusually high number of voters that were removed from the voter roll. There aren't any apparent reason why their testimony should not be considered as credible.
The second video shows a voter being shown a document she allegedly signed and that had for effect to switch her party affiliation from Democrate to Republican. However, it seems the signature on this document has been forged. There have been multiple reports of allegedly forged signatures during this election, most of them reported by voters on social media. Here is another example: https://www.facebook.com/groups/518955314959236/permalink/51...
Sure the validity of these reports need to be verified but I don't see why they should be dismissed beforehand.
Same for exit polls. The author notes that Clinton voters were more likely to vote early, which would logically cause them to be under-represented in exit polls. Also the "enthusiasm gap". Also lots of people have written about how poor exit polls are at predicting outcomes (http://fivethirtyeight.com/features/ten-reasons-why-you-shou...). Exit polls are notoriously bad, which is why they're basically only fodder for the cable news industry.