The good news for Republicans: Reaction to the Democratic convention was also historically weak. The margin between those who say Hillary’s big show made them more likely to vote for her versus less likely was just four points, 45/41. Never before since Gallup started asking this question in 1984 has the spread for a Democrat been that small. (Although the spread for Obama’s 2012 convention was very close, at five points.)

The bad news: Trump’s numbers are much worse. The first column is “more likely,” the second is “less likely,” the third is “no difference,” and the fourth is the margin between “more” and “less”:

gal

Trump’s -15 is the first time in 32 years of polling that a convention from either party has alienated more voters than it’s attracted. As you can see, no previous modern GOP convention has cracked 40 percent on voters saying they’re less likely to support the nominee, let alone 50. Among Democrats dating back to 1984, only Hillary managed to break 40. (She shattered another glass ceiling!) Similarly, no nominee on either side has polled under 40 percent in terms of making voters more likely to support him until Trump pulled it off in Cleveland.

Question: How is this possible when he got a bounce after the Republican convention, one strong enough to put him ahead in the RCP average for a few days? You don’t normally see candidates surge in the polls after they’ve held an event that’s getting thumbs down from the public. I can only assume that most anti-Trump sentiment is/was already priced into the polls before the convention. Add up Hillary’s, Gary Johnson’s, and Jill Stein’s share of the vote across various surveys and you’ll get a number in the ballpark of 50-51 percent. (It’s precisely 50.0 in today’s RCP average.) If you were with one of those candidates before the Republican convention, chances are it’s because you’ve already been exposed to Trump repeatedly on the news over the past year and have decided to pass on him for reasons ideological or temperamental. That is, much of the 51 percent that told Gallup they’re less likely to vote for him now was probably highly unlikely to vote for him in the first place. Whereas the 36 percent who are more likely to vote for him now might include a larger percentage of curiosity-seekers who haven’t paid much attention to the election so far, decided to tune into Trump to hear what he has to say, and liked what they heard enough to shift into his column. It’s not the raw numbers in each column that matter, in other words, it’s the number of persuadables. In a traditional two-way election, having 51 percent say they’re less likely to support you after your speech would be near-fatal. In a four-way race, it’s … not good, but not fatal.

That’s the best I can do by way of a sanguine pro-Trump spin. After all, it’s also possible that there are more persuadables lurking in Hillary’s “more likely” numbers than in the “less likely” column. She’s in the same position as Trump fame-wise: Everyone knows her, nearly everyone has an opinion, therefore it stands to reason that much of the negative sentiment about the convention is coming from people who disliked her going in. If persuadables reacted more favorably to the Democratic convention than to the Republican one, which Gallup’s numbers don’t tell us, then you may see her not only erase Trump’s bounce but widen her pre-convention lead.

Here’s one other possible reason why Trump’s convention numbers are historically low in Gallup:

For all the hype about Bernie fans revolting in Philly, Hillary’s convention was straight down the line liberalism. (Apart from all the flag-waving, natch.) Trump’s convention, and in particular his acceptance speech, was nationalist, the first time in 32 years that one of two parties featured a non-incumbent nominee who clearly doesn’t follow the same ideology as much of the party’s base. It may be that the difference between, say, 41 percent saying they’re less likely to vote for him and the 51 percent you’re seeing here is a chunk of disaffected conservatives grumbling about his nationalist authoritarian program. Importantly, though, that doesn’t mean they won’t end up voting for him in November: Most conservatives who disliked the convention will nonetheless hold their noses and stick with Trump for the sake of the Supreme Court, which helps explain how he managed to pick up a bounce in other polls despite faring poorly here. Simply put, even some of the voters who are now “less likely” to vote for him remain #NeverHillary more so than #NeverTrump. Even so, though, a Republican nominee pulling less than 70 percent of conservatives this close to the election suggests a big unity problem, and risks catastrophe insofar as the longer the holdouts remain uncommitted to him, the more likely it may be that they end up defecting to Johnson or Clinton in November. Step one to winning this election is locking up the right. It looked like Trump had done that in the weeks after Indiana, when everyone else dropped out. It looks less certain now. And the Khan business surely isn’t helping.

One more data point from Gallup. When asked how they liked the nominee’s speech in particular, not the convention generally, 44 percent said Hillary’s speech was excellent or good versus 20 percent who said it was poor or terrible. That spread of 24 points is in line with other recent nominees; Romney’s spread was +22 in 2012, for instance. Trump’s spread was … -1. Thirty-five percent said his speech was excellent or good, 36 percent said it was poor or terrible. My hunch, again, is that the minority of anti-Trump conservatives was the difference, but Gallup doesn’t give us enough data to know for sure.