Have you ever released a video where there seemed to be a disproportionate amount of "dislikes" compared to your other videos? Did you ever wonder what that meant? Well, having seen a lot of videos over the past few years, that like/dislike bar can tell you a lot about how your video is perceived, but a disproportionate amount doesn't necessarily mean that this particular video is "worse" than any others. It certainly can mean that, but sometimes, people dislike a video for something completely outrageous. And I also believe that there's no such thing as "bad engagement" in most cases.
A Few Extreme Examples of Disproportionate Hate
I started thinking about this with this video from Minute Physics:
Here's the likes/dislikes bar for this video:
Minute Physics is a well-respected YouTube channel where people go to learn, and many of these "whiteboard" videos are excellent. People who watch these videos are used to some complex information being transmitted in an easily-digestible format. When this video came out describing the mathematical FOIL (First, Outer, Inner, Last) method, a pretty basic concept for most Minute Physics viewers, they turned on the channel in a big way.
The content wasn't bad, it just wasn't something MP fans were accustomed to seeing. It seemed like this video was being "phoned in." And the like/dislike bar was pretty even at first, like an unheard-of 1:1 ratio for a video with this many views.
It forced this video to come out:
Since then, the first video has gotten more of a "benefit of the doubt."
In the category of "no engagement is bad engagement:"
I think Rebecca Black learned a lesson about YouTube. First, you need thick skin, and second, bad engagement can still turn into major views. That 53 million views number, by the way, isn't nearly as many views as this video actually has, because it was pulled on its first go-round (probably from all the hate it generated). But someone tell me how a video that seemingly no one likes shot up to 53 million (and definitely more than that)? And even with over a million people hitting that dislike button, there's still over 250,000 who like it. There are probably people who are liking it ironically on one side, and people are disliking it for fun on the other side. But you're doing something "right" if this many people want to watch the video.
People are so crazy about music, anyway. It seems like every major music video has a hard time being universally loved. Look at this mega-popular video for Rihanna's "Diamonds:"
Way more liked than disliked, Rihanna still has nearly 60,000 people who watched this video and said, "I'm going to express my displeasure with this." It's got a ratio of 17:1, which is good. I think if you get anywhere past 10:1 or more, you're doing well. But maybe you're used to your videos being 20:1 or more, and you get that one that drops to 10:1. What then?
I was talking about a channel that I help run where we have to deal with trolls from time to time. We have videos that range from 70:1 to a couple that are 3:1. We have a lot of subjective content in our videos, opinions. If people don't like the opinion, or think you're being disrespectful to a certain topic of discussion, they'll dislike it immediately. It's my sneaking suspicion people will dislike a video as soon as they see a title they don't like. I've seen comments where people enjoyed the whole video only to dislike it because of one stupid thing. It's comments like these that make me not take that dislike bar too seriously. To me, it's like politics. People may like everything about a candidate except for one issue, and decide to vote for another guy they mostly dislike but agree with on that main issue.
When Dislikes Matter
But if you're doing music or comedy, straight-up, two things I believe people get more excited about/more disappointed in than most genres, and you see a video that is disliked more than usual, that might tell you something about the content. If you've built a channel that regularly has a ratio of 20:1 or similar, and then you make on that drops to 9:1...that might be a sign that you did something that didn't quite come together. You have fans who like you no matter what, but you didn't convince the people who take an objective look at your videos on each upload. You may have 10,000 subscribers, or 100,000...in that group you have a core group of fans who like everything you do, and you have a core group of fans who mostly like everything you do, but will let you know when you've done something to displease them.
So I think you look at your averages and if you see a noticeable drop, then you've probably done something wrong. Take a look at Freddie Wong, who regularly gets great scores. He expressed his displeasure at the Harlem Shake phenomenon and came out with this video on Node:
It has a 32:1 like-dislike ratio, which is phenomenal.
Everyone pretty much loved that take on a very quick fad that swept the nation for about a couple of weeks. But then he decided to do another one:
As you can see, that one is about 2.5:1. Personally, I think it's as irreverent of the Harlem Shake phenomenon as the one he did on Node. But the presentation of the first one has a cathartic element of surprise to it, whereas the other one is like every other Harlem Shake video...only way creepier.
But FreddieW has such a consistent fan base, they can easily forgive him for one video like this from time to time. And remember, it's well more liked than it is disliked. But more people decided to click "dislike" on this one more than normal. They told Freddie, "Please, no more like this."
So what your dislikes and likes mean is probably going to be more important down the road. If you're starting out, and you don't have that many views, and it's 3 likes to 2 dislikes, it can't tell you much. You can learn some things here and there from the comments and try to make something better next time, but not until you get that larger sample size will that bar matter. I'd say it's probably when those likes/dislikes get into the double digits that you can start assessing how much (like, I think 30 to 30 would tell you a little something), but an even larger sample is probably needed.