Being less wrong than MatPat about why PewDiePie became the most subscribed channel
Since 2013 the video MatPat created to explain why PewDiePie became the most subscribed channel has been used as an authoritative source to explain exactly that. However there is good reason to believe it is just a theory … an incorrect theory.
The main issue with MatPat’s hypothesis is that he made it entirely by using a single data point, namely that of PewDiePie. By simply mentioning this obvious fact it becomes clear that literally anything that could be perceived as making PewDiePie unique could have also be used to explain his rise, and would have been backed up by an equivalent amount of data points.
MatPat, however had knowledge about youtube regional penetration behaviour, citing that PewDiePie’s channel based in Sweden could easily get views in Sweden, and then when he moved to Italy, get them there as well, and since his videos were in English they could be viewed internationally and penetrated easily into the United States. So MatPat was not theorizing blindly, but the merit of a theory cannot be measured by how well it fits over the data points used to formulate it, but how much predictive validity it has for novel data points.
While his statement that video promotion in the algorithm works regionally is completely correct, he did not initial say this. At the end of his more recent video MatPat he did state it as such. However it should be noted I have no dispute about this phenomena as I can observe it myself. The problem is he did not show how this phenomena was specifically related to PewDiePie’s success. All he did was say that “this can boost a channel” and PewDiePie did this. Most of his game theories work like this, which is fine when something is largely for entertainment, but in other cases a higher level of rigour is required and it is necessary to attempt to use the theory to explain behaviour not used to create the theory in the first place.
If what MatPat was saying was correct we should have seen similar boosts for other youtubers who moved locations, regardless of the type of content they made (they don’t necessarily have to be gaming channels to see such a travelling boost, and most likely vlog channels that traveled, would be seeing a boost relative to vloggers who stayed in the same location) Though seeing as the algorithm preferred gamers in MatPat’s theory, the travelling is only necessary to explain why PewDiePie is that particular gamer which made it to the top of that genre.
However PewDiePie was by no means the only gamer who was located outside the United States, and he was probably not the only one who had moved. Into that second point, it should be noted that PewDiePie moved to Italy in 2012 after he already had 1 million subscribers.
Italy could not have possibly been a large enough market to have made a difference after he already had over 1 million subscribers and was already growing at an obscene pace gaining 10,000 subscribers per day.
Regardless of where you live, the top countries in English that you will likely first penetrate into on youtube will be US, UK, Canada, Australia, Germany, Netherlands, and Sweden, with the last three being high because they have high English as a second language speakers. (The Netherlands and Sweden actually have greater (90%) portion of English speakers than Canada (85%) because lots of Canadians only speak French. Italy has low English as a second language speakers (40%). Not only that but Italy has extremely low internet usage rates, with only 56.8% in 2011, where as the countries where there are a lot of English speakers (included those as a second language like Sweden) generally had very high internet access (~90%).
PewDiePie would have a local promotion advantage for Sweden, which would definitely helped him in the beginning stages of his channel by having a boost in Sweden, but he was by no means the only Swedish gamer, and even a German or Dutch gamer who would have had the same advantages. The question of why “PewDiePie” which is what MatPat’s hypothesis was trying to explain, is still left unanswered as there are simply too many people who would have had a Sweden-like advantage if we assume moving to Italy at the time he did would have been completely insignificant.
This is of course not a new criticism, as PewDiePie himself has stated it. I believe he slightly misunderstood matpat, thinking that matpat was saying this explained why he was growing so much in 2013, rather than him being given a slight edge in the beginning which caused him to snowball. By giving matpat the benefit of the doubt, we can interpret him saying it was PewDiePie traveling to Italy which gave him an advantage much earlier in his career rather than him moving much later when it was too late to have an effect.
PewDiePie did in fact visit Marzia in Italy in August 2011, but he specifically said he was not uploading videos during that time (besides the vlog saying so). He only had ~30,000 subscribers at this point, so by being charitable uploading a single vlog announcing he would be appearing on a Swedish radio program while in Italy MIGHT have been able to influence his growth.
However if simply travelling to a single other country gave him the advantages, then tons of other people must have had the same advantages as travel is much more common than moving. He may also have not changed his country location as a result of this travel and thus the entire point would be moot.
None of this is to say his Swedish advantage was not present, as I can attest based on my own limited data set of my own channel that home country advantages are a thing and thus MatPat’s hypothesis in a diminished form that is much less of a killshot and more a mundane everyday advantage still might be plausible.
Unfortunately for MatPat, this everyday advantage probably is not as significant as he thinks it is. Thanks to the latest video, I now have the ability to reconcile what he was saying with what I observed myself with my channel. He said
“At the same time I made that video I was actively working in youtube analytics […] as a consultant for some of the biggest channels on the platform. I saw the effects for being a successful international channel. […] English speaking international channels would cross into the US very easily, giving them huge views in their home country and in the US. But US channels trying to cross outward into a different country at a certain threshold of views in that country before something in the data just flipped like a switch and suddenly the views just started to pour in from that new part of the world”
My interpretation of how the algorithm works based on what MatPat is saying here is that the content recommendation system is region specific. You get recommended videos based on what people in your region are watching (the trending tab already works like this so why not other things) If nobody from that country is watching the videos, the algorithm won’t have the available information to know if it should recommend it, but once it has the information the flood gates can open.
This quote provides me with the final piece of evidence I needed as I finally understood what he was referring to in the original video. Where I think MatPat’s issue stems from is that he was working with the biggest channels on the platform, which is to say the channels he was working on were likely given running starts rather than grinding their way up the ranks like normal people do. Due to this running start the videos were likely already getting lots of views before the information was available in the other countries to determine if they should be recommended, and since foreign countries are much smaller English speaking markets than the US, it took considerably longer for enough people to stumble upon them to get enough information. In contrast since the US is so massive it is relatively easy for enough people from the US to stumble upon the videos to provide the algorithm with the proper tribute to bring a rain of views down to America(perhaps through direct means if social media is used), but bringing the rains down to Africa is nigh impossible. When you do manage to unlock a new country, due to your running start, when it rains it pours.
In contrast if you are a normal person who is growing organically, you will periodically unlock new countries in an unimpressive manner. I remember that the first countries I unlocked as a Canadian channel were US, UK, Canada, Australia, Germany, Netherlands, and Sweden, and for awhile this ranking remained the same, but suddenly because of specific videos I did I managed to unlock Turkey and Spain, and then eventually all of a sudden I unlocked a lot of more exotic countries such as India, Indonesia, Philippines, etc. Currently my ranking for most views by countries goes US, UK, Canada, India, Turkey, Germany, Philippines. Now other factors might be at play with myself, but I distinctly remember that I suddenly saw the phenomena MatPat described of views pouring in from a new region. As I have only one data point, this testimony should be taken with a grain of salt, but even just from MatPat’s testimony I believe that an achievement unlock model of regions where some regions are more difficult to unlock than others can be used to best describe how it works. If you unlock the regions organically as you grow you will likely unlock a region before you would have seriously missed out on any growth as a result of having it be locked.
Thus I likely think that even though international channels have an easy time getting into the US, they have to unlock all the other countries just the same as an American. As a Canadian channel, the US has always represented about 40% of my views, but I used to get more Canadian views than UK views, but with time UK surpassed Canada as you would expect based on population. Canada still punches above its weight in terms of population with the UK because of the local area promotion, but while my local promotion is on Canada, there is likely extra local promotion in the US for American channels, so overall if you are a normal channel that does not fall outside the range of that the algorithms were tested for (AKA organic growth rather than trying to immediately try to launch a new big channel), it is difficult to determine if there would really be an advantage either way, and to what effect if any it could have on growth down the line.
Therefore if you are a channel grinding away, subscriber by subscriber, the rate in which you unlock countries will be much more organic than what MatPat saw with massive channels with running starts. Due to this gradual nature, PewDiePie had probably already unlocked every single country that could provide a significant number of views and subscribers long before a small Swedish advantage could have caused him to get a runaway lead ahead of Americans.
Not to mention that even if Sweden-like bonuses existed for foreign country based gaming channels, that would only give him an advantage over American gamers, and so would all foreign gamers have a leg up on Americans. In such a case American gamers and other channels that grew organically like PewDiePie and others did should consistently get swamped by hordes of foreign gamers. While Jacksepticeye exists, so does Markiplier, and to really see if there is a foreigner advantage as would be predicted by MatPat’s theory, a thorough analysis of many channels would be needed to be able to tell if there was some disadvantage for American channels. (Which would extend to ALL genres, not just gaming)
Even the explanation that PewDiePie’s rise to the top was brought upon because he was a gaming channel is brought into question by the near simultaneous rise of Hola Soy German, a Spanish language comedy vlogger, to the second most subscribed channel on the platform.
I used to accept MatPat’s hypothesis because I had no reason not to, but when I decided to look back and determine who were the historical most subscribers channels I saw that PewDiePie’s rise up the ranks to the top coincided with Hola Soy German’s rise up the ranks to second place, I realized that none of the explanations used for PewDiePie fit him, and yet he experienced the same phenomena.
There had to be other factors at play. For one thing, although Chile would likely have the same advantages Sweden would have in the Spanish market (with the largest Spanish speaking country being mexico with 128 million a good US analogue, Spain with 47 million a good UK analogue) so the single small country model applies, but he did not move until July 2013 when he went to an English speaking city (wild guess says Los Angeles) At such a point he was already the 8th most subscribed channel at 7.4 million, just 900k away from PewDiePie at 8.3 million. (Just as an anecdote all the youtubers above these two which were Smosh, JennaMarbles, RayWilliamJohnson, and Nigahiga were American). Therefore I do not think him moving to most likely Los Angeles significantly impacted his channel’s success. Again I agree with PewDiePie that its extremely unlikely that moving significantly impact a channel that already had millions of subscribers.
Hola Soy German was not a gaming channel, he made about 5 minute comedy vlogs, and yet he was blowing up hot on PewDiePie’s tail. My theory as to why this was happening was that the demographic situation in Latin America combined with those countries connecting to internet en mass created a boom in new Spanish language accounts. I go into greater detail in this video.
What I go into during the video is similar to what MatPat said about India and T-Series rise being the result of many millions of people coming online, with the exception that Latin America was slightly ahead of India in internet adoption, so the large increases were occurring during the early 2010s rather than late 2010s like in India. (It should be noted that since India speaks multiple languages, likely only music channels which can be appreciated by people who speak a different language will be extremely popular, where as youtubers will be split among the languages like how currently movies in India get split between Hindi speaking Bollywood and Telegu speaking Tollywood etc making it more difficult for them to get to the top)Thus Youtube trying to attract Latin Americans coincided with the algorithm switch to value watch time rather than strictly views, but was probably unrelated, thus Hola Soy German rose because of many Latin American kids logging on to youtube for the first time and he was growing quickly because he was well situated to capture these new users as subscribers, while MatPat was probably correct in saying the algorithm switch favouring gaming led to PewDiePie having a slight advantage, however Hola Soy German’s simultaneous rise means we need to find another common factor.
That image is what the front page looked like when MatPat uploaded his original PewDiePie video on July 20, 2013, and it looked like that from January 1, 2013 to April 3, 2014.
While there were numerous new Spanish speaking people to drive Hola Soy German’s subscriber numbers up, within the English speaking world Youtube was hard at work at trying to get existing casual users to make accounts and subscribe to more people. “Channels For You” was a feature that drove more subscribers than ever before. If you go up and look at the January 15, 2013, graph you will find that Smosh had under 7 million subscribers in the beginning of 2013, but even though by the end of the year they got dethroned they still more than doubled their subscriber count during that year.
It should be noted that it was quite controversial at the time that in late 2013 Google Plus was integrated into Youtube, which only accelerated the number of people making accounts (this was when I had made my first account even I had been using the sites for years as I remember watching nigahiga when I was just one of the great masses of kids who would watch their favourite videos over and over again) Also quite controversial was the extent to which Youtube Spotlight was promoted with the “Channels For You” feature. Many claimed youtube was biased towards its own channel by promoting it to everyone. As we can see this was enough that Youtube Spotlight was able to take … the spotlight away from PewDiePie for a brief period of time when they released the first Youtube Rewind. PewDiePie might have benefited from the feature greatly, but Youtube Spotlight did even more, which simply attests to the power of the feature if youtube wanted to use it.
Everybody was growing like crazy in 2013, and that meant that the fastest growing creators were made to grow that much faster. While snowballing your way up the ranks is normal for growing youtubers, because PewDiePie and Hola Soy German were snowballing at this particular time, their snowballs rolled that much faster than everyone else.
Not only that, but they continued to grow incredibly fast even after PewDiePie became number 1. He continued to gain over 1 million subscribers per month until April 2014, exactly when “Channels For You” was removed. A few months after that happened PewDiePie started complaining that the Youtube Sub box was broken and MatPat described what he called “Subscriber Burn”. Likely what had happened was Youtube had flown too close to the sun and in their efforts to increase the amount of subscriptions everyone had, they had inadvertently made the subbox useless. Up until just before the rumor of the broken subbox had started circulating, the number of subscribers everyone was gaining more than offset any subscribers that were burned and thus the issue was masked, and thus these late 2014 issues were akin to a Youtube subscriber bubble popping.
To summarize, the reason both PewDiePie and Hola Soy German made it so quickly to the top was that a period of time where they were gaining subscribers faster than anyone else in their respective languages coincided with a time youtube was implementing features intended to increase total subscriptions overall, thus leading to what would have otherwise been a rather normal rise through the ranks at any other time period the algorithm briefly favoured a certain type of content, transformed it into a rocket ride all the way to the top to the point that they developed a lead so far ahead of everyone else that no one else could compete with them in subscriber numbers for years until Youtube engineered another subscription bubble for an entirely different sub-continent.
I am unable to say why PewDiePie and Hola Soy German were the people growing the quickest in their respective languages at that time period, as to do so would require data I just don’t have (although since MatPat asked PewDiePie to tweet him to arrange to see and discuss data, I offer the same thing to MatPat should he read this, https://twitter.com/vandegraaph), but an important thing to consider is that somebody is always going to be the on the top in terms of fastest growth, and in an unpredictable system like youtube, being on top on the right time can mean the difference between a meteoric rise and a slow drudging grind. The best solution to answer this question in this situation is perhaps to just ask PewDiePie himself.
PewDiePie covers the history of his channel using the wayback machine and states that 2012 was the year he blew up, going from 100k to over 3 million. He says that the reason it happened in 2012 was that it was the year he dropped out of school to pursue youtube full time, and that the work he put into that year lead to the success. What ended up happening is that when he hit his stride the algorithm and youtube layout features converged to favour exactly what he was doing, and thus the story of youtube success is as always the story of putting in the work to place yourself in the right place, so that when the right time comes and the algorithm picks you up you are ready to ride the wave. Numerous other people might have hit their stride at inopportune times, and they still did well, but things were just harder for them.
T-Series has been uploading music videos for years since 2006, but it is only recently that they have been blowing up like they have. The fact that they existed for so long meant they were already on top of Indian music on youtube enabled them to ride Indian user growth straight to the top.
