Author of the North-Northwest Tri light
On November 10, 2021, video site Youtube announced that Dislike, a button that was used to visually quantify audience feedback attitudes, would remove the public display count.
In other words, how many "likes" a Popular Youtube video gets won't change. But how many times it has been "stepped on" has become a secret that cannot be announced. As a result, users quickly judged the standard habit of video quality and collapsed.
Although the "press" button itself has not been canceled, this behavior, which is easily misunderstood as "self-deception to cover the fig leaf of unwelcome content", still caused an uproar on the Internet:
In the BBC's view, even if Youtube officially claims that this measure can reduce the malicious behavior of "click bombing" and alleviate the pressure and discomfort of video producers, it may not work. After all, "stepping" still stays below the video playback window, how many people don't like their videos, the publisher can still see clearly in the background.
The criticism from Youtuber is even more intense: this seemingly inconspicuous change represents the complete deterioration of the Youtube website.
In the past, compared with those websites that even had to hide the "likes" number, Youtube had the courage to put "stepping" in a prominent position in the interactive area and unabashedly present statistics, which would have won a lot of goodwill. But that pride is now gone.
As a result, the real Youtuber was completely enraged: just the day after the official announcement, the short film received more than 50,000 "stomps".
Subsequently, Jawed Karim, the most senior Youtuber (not one of them), re-edited and uploaded the description of his first Youtube video, "Me at the zoo", and bluntly expressed his attitude:
"Me at the zoo" is Youtube's first video, and Javid Karim is best known as the co-founder of Youtube, and the content of this video description is no less influential than Youtube's official announcement. Coupled with the video's 10 million likes, Youtuber's attitude is self-evident.
Despite the near-one-sided opposition, from the official statement of Youtube, the "number of hidden points" seems to be an improvement that is expected. Today, the number of "steppings" has been officially hidden.
The question is, what exactly is the reason why Youtube is so stubborn? And what is the reason for it to directly choose to "step" one-size-fits-all? What is the practical significance of the controversial "stepping"?
Self-inflicted bitter fruit Of Youtube
It can be said that for the UGC content community, the interactive function of "stepping" that clearly symbolizes the opposition posture and efficiently conveys negative emotions is a sensitive design that must be treated with caution in any case.
YouTube officials have also suffered their own consequences many times. At the end of 2018, millions of Youtubers came together to top youTube's official annual retrospective video, "YouTube Rewind 2018: Everyone Controls Rewind," with a uniform "stepping on" column.
In YouTuber's view, "YouTube Rewind 2018: Everyone Controls Rewind" is a large-scale "disaster documentary" full of old rotten stems, subjective assumptions, avoidance of the important and neglect of the left and right, plus disregard for the public opinion of the community, which belongs to youtube's worst annual review.
Once the user's emotions are intensified, no amount of justification is in vain. In just one week, the short film successfully surpassed Justin Bieber's "Baby" MV to become the #1 bad Review Video on YouTube in the Guinness Book of World Records, and it remained until the moment when YouTube officially hid the dot count.
With nearly 20 million "steppings" and a dislike rate of 86.77%, YouTube Rewind 2018: Everyone Controls Rewind has become a sliced sample of the negative atmosphere of the YouTube community and the negative emotional output of YouTuber with an unshakable bad review champion posture.
In the face of overwhelming "stepping" and ridicule, YouTube naturally does not dare to slacken off. Not long after, in an official statement, YouTube expressed its attitude frankly:
Beating Baby on the 'stomp' leaderboard isn't our goal of the year; honest feedback can be terrible, but we're always listening, always thanking users for their attention; trying to summarize YouTube's platform charm with a video is tantamount to trying to catch lightning with a Leyton bottle — we're experiencing the difficulty of content creation again and a further reminder to respect YouTube creators.
However, how this "respect" is realized is really surprising. In a video released on January 30, 2019 by Creator Insider, a channel that shares insider information about YouTube's technical team, Tom Leung, YouTube's Director of Project Management, describes two "stay respectful" scenarios:
First, when the viewer clicks "step", it must submit the reason for not liking the current video through the checkbox, so as to feedback more detailed rectification suggestions to the creator; in addition, this multi-point one-step operation can also allow the audience who do not hesitate to step on it to reduce the impulse appropriately, effectively reducing the emotional color of "stepping";
Second, completely eliminate the function of "stepping" - simple and crude, but how effective it is another matter.
Although Tom Leung repeatedly stressed that these proposals are "only in the internal discussion stage", they immediately caused an uproar after the news was announced.
In fact, if you look back at the history of the development of YouTube's interactive review system, you will find that as early as 11 years ago, YouTube made a complete modification to the platform's content evaluation function:
The now critically acclaimed "like" and "step" features were officially launched in early 2010, after YouTube used a five-star rating system similar to Amazon's.
As for why YouTube is making this one-size-fits-all change, a graph from the official blog explains:
At first glance, the five-star system means less granular comment feedback, but according to YouTube's official statistics, most of the ratings this system receives are the highest five-star, and then a completely disproportionate one-star, as for the "intermediate state" of two, three or four stars, almost no one cares.
"If most videos get five stars, what's the point of this evaluation system?" Maybe it's more effective to give a thumbs up/down, or would it be better to just say "like" to endorse the content? These are the questions we ask ourselves as we look at statistics and think about how we can improve the platform experience. ”
This was the question youtube asked in 2009 through the official blog post Five Stars Dominate Ratings. Now, at least for YouTuber, the design of publicly displaying click/like data has indeed been popular. But from the perspective of the platform side, is this solution really the optimal solution?
Unauthorized voting rights
If I give a video a "step", what happens next?
In all the confusion of YouTubers at the moment, this question definitely qualifies for the top three - and even PewDiePie, a funny video owner who specializes in Vlog "Can this video get 1 million dislikes?" " to sensationalize, the result of nearly 6 million "stepping",
For ordinary video creators, does "stepping" really just mean emotional venting, not any negative effects?
First of all, it can be clear that the number of "likes" and "steps" can indeed affect the judgment of the recommendation algorithm. This conclusion can be clearly seen in the official YouTube Creater channel video " "Algorithms" - Understanding How YouTube's Search and Discovery System Works:
How can our algorithm show the right video to the right viewers around the world at the right time whenever possible? We rely on data — we'll know the title, thumbnails, and descriptions of each video, and how much other viewers like the video; we'll also know how much the viewer has watched the video, how many people commented, and how many times they "liked" and "stepped on" it.
Even so, it's still difficult to quantify exactly how much a single "stepping" will affect a video:
According to a paper published by Google at the RecSys 2019 conference, many of the parameters mentioned above will be divided into "engagement (such as views and viewing time)" and "satisfaction (including likes and clicks, etc.)", and then use the width model and deep neural network to generate predictions based on parameters for each "appropriate video", and then superimpose a separate sorting model for recommendation.
But in the eyes of most outsiders, YouTube's recommendation algorithm is basically equivalent to a black box, and it is almost impossible to accurately grasp the impact of the playback parameters of the uploaded video on the number of recommendations.
Even if the complete model and algorithm cannot be cracked, it is still possible to use the exhaustive method to estimate the weights of parameters that may affect YouTube's recommended rankings. In 2018, researcher Justin Briggs of briggsby.com crawled 3.8 million data points from 100,000 videos from 75,000 YouTube channels for analysis, and the final result was as follows:
The number of visible "steppings" has an impact on YouTube content, although far less than the number of video views and likes, at least in all parameters is enough to occupy the middle stream.
However, the weight of influence is only one aspect, and whether the influence is good or bad is the key. Back to the beginning of the question: under the influence of weights that cannot be ignored, what does it mean to harvest a large number of "stepping" on YouTuber videos?
It's clear that this question touches on sensitive areas of the YouTube content industry, and it's hard to expect an answer through the formal media. However, we can still peek into a conclusion worth thinking about through the caliber of the "professional team" engaged in related industries:
The point is that no video in YouTube's history has ever received 100% positive feedback; even the most popular videos have a bunch of "stomps" for a variety of reasons. If a video has thousands of "likes" without half a "stomp," the alarm goes off and YouTube is well positioned to detect such anomalies and act accordingly. So if you've ever bought a "like" for a YouTube video, you'll also need to buy a corresponding "stomp"; it's a trade-between a balance between positive and negative interaction feedback that makes the whole deal look like it doesn't have to go through.
Above is the veteran social networking provider MediaMister," an explanation of "why buy YouTube click"; in contrast, in contrast, in the more senior InstaFollowers, the answer is much more explicit:
Considering that "stepping" is a weighted metric that is critical to the ranking of content recommendations, and can convey the impression of "poor quality" to YouTube, you can buy "stepping" to undermine the performance of your competitors' content recommendations. As for quantity, if you buy "stepping" to balance the number of "likes", then 10% is a good percentage; however, if you buy to curb competitors, then you can leave "see good and take" behind.
Considering that both companies have about a decade of operating experience (MediaMister was founded in 2012 and InstaFollowers went online in 2010) and have not yet closed their operations to this day, as a seasoned team of senior gray production professionals, this statement is naturally credible.
In addition, Devumi, a "rising star" whose business level is too prominent and the results have led to the forced banning by the regulator, is even more from the standpoint of a professional gang, bluntly pointing out the essence of the "brush volume" business:
Social identity is a psychological concept that looks at group behavior. It contains the basic theory that when you see enough people doing something, you follow suit. For example, if you see a lot of people talking about new movies online, you'll be willing to check it out too — and the theory applies to YouTube as well.
The amount of brushing brings only a number that enhances social recognition. You buy these numbers to make your video work more active for real viewers to watch, comment on, and share.
Brushing the number of plays and likes can drive the live audience to actively watch and leave a "like", and vice versa - giving the "step" brush can also lure the real audience to follow the trend, which is the explanation from the "professionals". What "stepping" means for YouTuber today is clear.
So, since the "stepping" statement means that it has long been tarnished by gray suppliers, the negative impact of the recommendation weight that should not be underestimated, and the fact that this function has become an effective weapon for personal attacks and even Internet bullying, why did YouTube not delete the "stepping" function to solve the worries once and for all, as predicted in 2019?
YouTube changes
It is not difficult to imagine that, like the "stepping" that has deteriorated, under the influence of some business elements that have no precision, the actual operation of the YouTube algorithm recommendation system is not so idealistic.
According to a 2018 report in the New York Times, former Google algorithm engineer Guillaume Chaslot claimed that in order to improve platform stickiness and user retention, YouTube has tended to recommend aggressive video content to users through algorithms from a long time ago. Chaslot said it was because he was "shocked" by this platform strategy that he was swept out by YouTube in 2013.
According to Chaslot's subsequent research, for a long time (at least until 2016), YouTube has shown no obvious willingness to adjust to the content-oriented recommendation system that has become more aggressive – under Chaslot's efforts, the traditional media represented by the Washington Post finally agreed that "YouTuber is being led astray by the lack of moderation in content recommendation systems."
While not denying the conclusions head-on, YouTube has made a series of adjustments in recent years. The adjustment around "stepping" is only a conspicuous but not core part of it, and the invisible algorithm model modification is the highlight:
According to Google's official introduction, with the continuous optimization of the algorithm and the continuous adjustment of the parameter weight, at present, YouTube can automatically detect and flag 94% of the illegal content, 75% of the illegal content will be removed before the number of views reaches 10; as of Q4 2020, the VVR (Violative View Rate), which represents the clarity of the platform's content, has dropped to 0.16-0.18%, which means every 10,000 Only 16 to 18 times of user viewing behavior clicked on objectionable content; compared with the same quarter of 2017, it decreased by more than 70%.
The purification of platform content means a reduction in radical content. However, as long as those brush "stepping" businesses that ignore content and are purely rhythmic still exist, then expecting the user group to find rationality and let the "stepping" truly reflect their own attitude is doomed to be a matter of fate.
Similarly, the behavior habits developed by platform users for many years are obviously impossible to change easily, and directly deleting the "step" is bound to cause great disgust, so it is necessary to retreat to the second place, retain the function but hide the count, and fundamentally block the environmental conditions of the "step" of the follow-up brush, it seems to become an effective program that is not perfect, but it is indeed reasonable.
Judging from the current feedback from all sides, YouTube's obviously deliberative adjustment strategy has not won the recognition of users: in the eyes of many YouTubers, only retaining the function of "stepping" and canceling the count of "stepping" means that the actual quality of the click video cannot be clearly identified from now on - then, is there a scheme that can visually display the number of "stepping" and accurately reflect whether the "stepping" is a follow-up with the rhythm, or the real emotional output?
An example comes from Steam:
In response to the growing number of malicious bad reviews, in Q3 2017, digital gaming platform Steam made eye-catching adjustments to the user review interaction system: in addition to continuing to display the number of interactive reviews, and visually presenting the content quality through 9 gears from "good reviews" to "bad reviews" in percentages, the "good reviews" and "bad reviews" submitted by players can now be clearly presented through the timeline, and through simple clicks, you can directly jump to display the "bad reviews" at a specific point in time. Player reviews.
As a result, Steam players can easily examine whether the controversial content is due to quality or deliberately has a rhythm, which not only makes it possible to have a number in mind before placing an order, but also effectively reduces the probability of following the trend of bad reviews.
So, is it possible that a system that can guarantee transparency in interactions and truly achieve the ideal goal of self-responsibility is achieved on YouTube? The answer is not so optimistic:
Although it topped the digital gaming platform industry with 120 million monthly active users (2021 statistics), Steam's volume and popularity are still far from Being compared to YouTube, which has more than 2 billion monthly active users (also in 2021 statistics). Considering the huge difference between the content ecology and consumption mode of the two, it is purely expected that the user evaluation and interaction system will not be able to be used.
Not only that, today, 62% of Internet companies have regarded YouTube as an essential channel for publicity, and platform users spend more than 1 billion hours a day watching time; as a result, just like a series of related industries derived from traditional TV channels, countless peripheral industries have also emerged around YouTube - then, we may guess that in addition to formal and legal marketing services, those who rely on YouTube to make a living "professional teams", How many are in the gray or even black zone?
Enter "YouTube Bots (robot programs that interact automatically according to instructions, including but not limited to brushing "stepping" and "like") on Google, and from the first page, the results returned on the entire ten pages are almost all related ads, and there are almost no media reports that face this phenomenon and investigate.
At the end of the day, whether the "unreal" component of the YouTube spin-off industry is "an elephant in the room" or a "dragon in the garage" is known to everyone, but everyone is reluctant to break it: because everyone knows that the consequences of revealing this truth are likely to exceed the maximum carrying capacity of this website, the entire industry, and even Web 2.0 — after all, what it means for the Internet to puncture the bubble was witnessed by a whole generation 20 years ago.
When we can't find the perfect solution to the problem, the least bad answer is probably the optimal solution we have to accept — and for YouTube, it's the only version of the answer.
Not without regrets. But perhaps, this is also not perfect Web 2.0, leaving the era inevitable.