Unpopular Opinion: TikTok's Success Is Not Just About Feed Personalization
UGC today is about creation first, period.
The “For You” page and the “algorithm” are constantly discussed as the magic behind TikTok’s success. Its personalization and AI are widely applauded and talked about as almost magic. Unquestionably, TikTok has figured out the right mix of new, viral, and “your type” of content to bring to your feed and that has catapulted it to massive engagement on the platform. But, I’m about to argue that the personalization was only the cherry on the cake that took it to the billion-plus users club and the hour-plus engagement club. Success started much before that, while the (western) world was less aware of it.
Prologue: The Summer Of ‘19
During the summer of 2019, as we were getting ready to launch Rizzle beta, we had four college interns in our US office, with the most diverse backgrounds you can imagine. They had one thing in common - they were all women (not intentional, but it did end up that way - two guys we made offers to didn’t show up!). That team had a white woman, an Asian woman, an African American, and an East Asian woman. One of them was studying political science, one journalism, one marketing, and another applied math.
There was one problem. None of them were on TikTok. They also did not know anyone who was on TikTok (other than one person’s middle school brother!). At the time, America was in denial and the American adults were busy ranting about the TikTok cringe.
Even those who knew of great TikTok videos would often say “I just watch TikTok on Twitter” (or Reddit or <their favorite social platform>). File this away in memory because we’ll come back to it later in this article.
As a social experiment, I mandated all our interns to be on TikTok and really understand what was going on in the short video world. These ladies would install TikTok on their phones during work hours and uninstall it when they left work (or when they were going to meet friends). I had never seen people going to that extent to avoid caught being “juvenile”, as they described it.
Make no mistake, by then, TikTok videos were blowing up on the Internet, racking up millions of views on Instagram and Twitter. But for those that had no interest in making the videos, the curated videos on other platforms were sufficient.
Content Spectrum: The 10000ft Level
At a high level, we can slice the content spectrum in the above manner, from the point of view of who conceives, creates, and curates the content.
User-Generated Content (UGC)
This is the most coveted type of content for social platforms, essentially because it keeps millions of people engaged and obsessed about their own creations, and the creativity and the content comes for free. The creator economy of today makes it not all free, but it is and will remain “almost” free because of its long tail nature. More about that in another post.
The only defining characteristic of UGC used to be that it is created by users of the platform - i.e., anyone can be a content creator. Today, UGC also often means that it is created on the platform itself. While it is not always true, it is getting closer to that.
Platforms like YouTube enjoyed the privilege of having UGC without offering any creation tools. Today’s platforms cannot get away with that. In some ways, today, the quality and quantity of UGC are directly proportional to the creation tools that the platform offers.
While Snapchat was among the earliest UGC platforms for vertical video, Snapchat’s UGC quality was extremely poor, especially in Stories - its primary vehicle for videos. I will speculate on the reasons for that in a different post.
Perhaps discouraged by the UGC quality in Stories, Snapchat swung the pendulum hard in the other direction, going for exclusively pro content in Discover and Snap Originals. Today, as Spotlight tries to compete to become a UGC short video play, this DNA crisis shows.
On the contrary, Instagram had developed a culture of glossy, high-quality UGC. When it became the destination for Stories, the UGC quality in Stories naturally got a lift, keeping with the DNA of the platform.
An aside: Today, there is a lot of discussion on “authenticity” and “anti-Instagram”. Apps like Poparazzi and Dispo are gaining traction based on this concept. I will reserve my detailed comments on this sentiment for another post but suffice to note that “authentic” is not always watchable at scale. Watchable videos are almost always performances and *sometimes* - only sometimes - authentic content that is deeply relatable and unique makes the cut. But, the content styles and genres considered watchable at scale changes over the years, and platforms that can’t keep up, end up losing. At the risk of being hugely unpopular, I will claim that Instagram’s crisis stems from this more than the lack of authenticity itself.
UGC brings the creativity of the masses and highly motivated creators at near-zero cost to a platform. Done right, it is the key to reaching huge scale.
Professional-UGC
P-UGC, or PGC for short, is paid content that is often meant to be an inspiration to UGC on a platform. For many platforms that emerged in the western world, this is a foreign concept.
PGC requires reasonably deep pockets to seed at scale and is not something a small startup can easily afford, particularly in western countries.
Platforms like Instagram have enjoyed rapid proliferation of UGC right from the start and have not had to dabble with PGC - until now. Rather, the truth is that Instagram did not need to use money as a currency to woo the early professional creators - their Weekend Hashtag Project featuring the top creators and content, initially on the Instagram Tumblr and later on the Instagram official account on the platform itself, was sufficient to send professional photographers to hanging bridges and all corners of the world to bring amazing photos to the app.
That said, YouTube was an early adopter of PGC (even though that term was never used to describe it). Early on, with its YouTube Originals program, YouTube went on to seed content that would define what top channels aspired to bring to the platform. This is what set the standards for YouTube where great creation meant that one used 4K cameras and highly skilled editing to make amazing videos.
However, because the bar from PGC was set so high on YouTube, the fraction of users who turned into creators on the platform was very small. In an earlier post, I covered how YouTubers need to be masters of many skills to be successful.
In the last decade, PGC has been a huge part of the initial bootstrapping on short video platforms in China and India. These countries, especially rural India and China, need firm guidance with leading examples for people to follow. Platforms like Kauishou and TikTok (Douyin) have been excellent at seeding the right types of PGC at the right scale.
Douyin and TikTok successfully inspired amazing quality UGC at large scale, while Kuaishou has had to stick with PGC content for a good portion of their feed. The UGC quality in Kuaishou remains questionable, even after all these years. I will analyze the potential reasons why Douyin and Kuaishou had drastically different outcomes on UGC quality in another post.
Another reason PGC is so prevalent in India and China is that it is actually affordable to operate a fairly large-scale PGC operation, with moderately deep pockets. To execute a PGC strategy in the US, it would require massively deeper pockets!
But, that is precisely what TikTok pulled off - by paying for a diverse and robust PGC pipeline, TikTok rapidly bootstrapped content types of interest to huge parts of the population.
It is extremely rare for a platform that does not successfully make the transition from PGC to UGC to succeed big. Kuaishou is an exception, not the norm. This is primarily because the collective creativity of PGC producers is limited and will never reach the same level as creativity of the masses.
An aside: Eugene Wei correctly postulates that culture is a huge barrier for platforms that originate in one culture to massively succeed in another culture. However, he then, incorrectly in my view, goes on to observe that TikTok’s algorithm made it possible for it to succeed globally, in multiple cultures (China, India, and the west). In another post, he goes on to observe that there is no groundbreaking magic to TikTok’s algorithm (this I agree with). These are loosely contradictory viewpoints, although not immediately obvious. I will reserve my detailed notes on why I think TikTok succeeded globally in multiple cultures for another post (that’s an essay). But, it has largely to do with nailing the recipe to make that transition from PGC to UGC so very well. Kuaishou has repeatedly tried to infiltrate other markets and failed, mainly because they have not figured out the recipe to make that transition effectively.
Relying on PGC alone means relying on a few people’s abilities to pick winning content formats and videos time and time again. This rarely ever happens.
A platform that largely operates with PGC gets closer to Pro and OTT content. Even Jeffrey Katzenberg spectacularly failed in identifying winning content. Disney and Netflix have their share of losers. It is risky business.
Professional (Pro) Content
I distinguish Pro Content from P-UGC by its intent. Pro Content is not intended to be exemplary for UGC. Rather, it is all about the viewers. It is meant to be created by a selective few highly talented creators and consumed by millions.
But Pro Content differs from OTT and other content (also professionally produced) in that it is distributed on a social platform and typically has access to all the social features similar to UGC/PGC.
Because this content can be interacted with - commented on, shared, memed, etc. - the bar on quality is not quite at the OTT or film levels.
Interactions are a fair tradeoff for quality because the audience gets to bond with the content and creators in other ways.
Quibi really suffered the worst of all worlds with sub-standard (relative to OTT) content quality, lack of social interactions, and a subscription fee! More about paying for content in a different post.
Pro Content can be complex to create since it is not meant to be reproduced or imitated by anyone else. In fact, the harder to imitate it is, the better.
Starting at Pro Content and beyond, content starts getting referred to as “IP” (Intellectual Property). This comes with rights and distribution restrictions that are much more stringent than UGC or PGC.
YouTube has certainly brought Pro Content over the years as it continued its YouTube Originals program in some shape or form. Snapchat Discover is all Pro Content. Both Snapchat and YouTube have done incredibly well on this front. They’ve picked some good winners here!
Pro Content co-exists with UGC on the same platform and is often interacted with in manners similar to UGC.
OTT, TV, and Cinema
Here, I’m grouping the remaining types of content because, for this discussion, their differences are irrelevant.
OTT and its class of content - series and movies - are produced professionally and generally consumed on large screens. It is also often behind a paywall and requires a subscription to watch. There are exceptions like IMDB and the Roku Channel which are trying to run a free, ad-supported program. Unfortunately, that typically reflects on the quality of content they are able to bring - the vast majority of content brought to these channels is second grade.
An aside: Competing with platforms that have billions of dollars to spend on content production becomes harder and these platforms have to rely on the population that is looking to consume content on the cheap, for growth. But, that population also has lower discretionary spending power, and hence, the ROI on ads is lower. This is a really tough spot to be, unless, like Peacock or Paramount TV, the OTT channel is a secondary business backed by a custom content pipeline that comes from a possibly related but different core business.
With Ratatouille the movie, TikTok is starting to get into Pro Content at a small level. Generally speaking, the economics of bringing this grade of content does not work out at scale for free, UGC platforms, even if they are ad-supported.
UGC Platforms Are All About Creators, Initially
A UGC platform succeeds when it brings the right mix of quality and quantity in creators and content. The scale of creators is critically important because that is the only thing that brings the creativity of the masses to the platform.
As long as content is conceived and controlled by a small group of creators, agencies, and the like, it is incredibly difficult to capture the imagination of billions or even hundreds of millions of people for prolonged periods of time.
Content that holds that level of interest takes a huge amount of time and money to produce (think Marvel or Disney) and does not fit the lifecycle of UGC.
Viewers Have Choices
The Internet has billions of videos worth watching today and there are well-established platforms that distribute much of this content.
Viewers have choices and a new platform is not one of those until it is. There is absolutely no reason for a non creator to get on a new platform to watch videos.
Remember I asked you to file the “I just watch TikTok on Twitter” comment away in memory in the early part of this post? Even if a great new platform emerges with some highly interesting content, until there is a critical mass, such content can always be consumed on an existing distribution platform, making the new platform irrelevant.
It takes relentless number of irresistible pieces of content shared by large groups of people over many months for viewers to eventually cave and opt for watching the content natively on the new platform.
TikTok is indeed such a viral content production machine that resulted in a continuous stream of viral videos shared to existing platforms, eventually amassing billions of viewers!
Early Adoption Order
You may have seen some version of this curve on Diffusion of Innovation.
But this curve only tells a vague story. It is still up to the platform builders to figure out who the innovators and early adopters are for their specific platform.
I revise the curve for UGC platforms as follows.
Let’s dive in to see who these users are and why they may be motivated to try a new platform at different stages of its growth.
Curious Creators
These are creators who will try anything new once - these are the “innovators” in the UGC world. They want to try new formats as soon as they are available to see if there is anything in them that fancies their imagination.
Undiscovered and Rookie Creators
This group may include talented creators who are yet to rise up to fame on other platforms, but usually, this group is mostly rookie creators. This is because, the talented creators, even if undiscovered, are still trying to make it on a popular platform rather than adopt some new platform. Some of these creators who are yet to find traction elsewhere may try out the new platform just to see if they find good traction and if the platform shows signs of rapid growth.
These creators will stay if they either find a core community they are able to vibe with strongly or if the platform catapults rapidly into popularity.
Talented Creators and Curators
As the platform gains popularity, more talented creators make their way over to it, even some of those who are already popular on other platforms. As this happens, curators flock to the platform because this group takes pride in being the first ones to find great content. They go where great undiscovered content lies.
This is the group that really launches the platform into mainstream growth at scale. Until this stage occurs, there is absolutely no incentive for viewers and lurkers to get on this great new platform!
There is a chasm between the rookie creator and the talented creator adoption that is hard to cross, but necessary for a platform to then succeed. More on this below.
Skeptical Creators and Viewers
As the platform blows up in popularity, the rest of the creators make their way to it, as do viewers. Not much magic here - it is just the cycle of growth.
Skeptical Viewers
This is the final stage where the remaining viewers decide it is time to let go of their stubborn claims of cringe or “nothing new here” sagas to join the herd and give in. Or, they finally accept that cringe is what they’ve loved all along!
Early Creators: Why They Come
Why and how fast creators flock to the platform depends on what the platform has to offer in creation tools and experiences. The more unique and outlandish the experiences are, the faster they will come.
Today, virality in UGC platforms is almost exclusively driven by stunning creation experiences that are not available on any other platform.
Such experiences drive platforms like Wombo and Reface to early popularity faster than say, TikTok. However, sustainable growth needs more - it needs to be able to keep the creators that come to the platform.
Early Creators: Why They Stay
One trick creation experiences that invoke curiosity can attract a huge number of early creators to the platform, but for creators to stay, more is necessary.
The creation experiences must not only be unique, but they must also be positive and delightfully repeatable. Over and over again.
This is where platforms like TikTok (and Instagram) succeeded big.
But, are repeatable and fun creation experiences alone enough? The answer is no.
These early creators also need fame - the platform needs to create a perceived fame model that is fair and appears to reward quality of content. The platform itself also needs to go mainstream - no one wants to be famous on a platform that no one else they know have heard about!
So, in a platform where viewers have not yet come in or stayed at scale, how does fame come about? And how does a platform get to large-scale adoption at a pace that keeps the perception of growth at satisfactory levels?
A fair early substitute for fame is community. If creators feel a tight-knit community that they vibe with, that brings a level of satisfaction that can keep the creators going for a while. But, even with a community, if fame does not eventually happen, things will fall apart. For community alone to hold value over the long term, these people must be united by a common cause that is bigger than most things in their lives.
When we talk about fame, it is not that every creator who does not find fame walks away. As long as all creators can see that some creators are gaining fame and that the ones becoming popular do have better quality content, everyone start to aspire to get there and it all works.
This is American capitalism at its finest. As long as it is fairly believable that hard work can lead you to a picket fence, yard, and family, people are willing to take the blame for being unsuccessful (the “you get what you deserve” mantra) and keep working hard. Discrimination exploits the American dream causing a vicious cycle in the long run, but I digress here.
Talented Creators: Why They Come
When a platform offers shockingly unique creation experiences, even the talented creators become early adopters, triggering “virality” at scale. But, unless the platform has sustainably repeatable creation experiences, these creators will simply leave after creating a few pieces of content.
Instagram’s filters were a highly repeatable experience - with every photo you took, the filter offered a better version of the photo. As long as you kept taking photos, it was worth putting them through the Instagram filters. Creation was repeatable and sustainable.
TikTok achieved this with video, not with a single key tool, but with a plethora of creation experiences that seem never-ending. I have more to say about the categories of creation tools, but that deserves its own post.
The Chasm: Crossing Over To The Talented Creators Phase
It is not automatic that a platform crosses over from the early creators to the talented creators phase. It is entirely possible that the platform offers unique creation experiences that draw and even keep the early creator community for a while, but unless a portion of these experiences captures the imagination of highly talented creators who may already be on their way to popularity elsewhere, they won’t have a reason to adopt this new platform.
This is the most critical phase of growth in a UGC platform - one that must be crossed with organic traction among the talented creators in order to woo the viewers at scale.
No amount of paid activity is a substitute for this; platforms that pay top creators will barely see contractual obligations being fulfilled unless the platform’s capabilities are addictively fun and repeatable and they see themselves and the platform growing at a pace that they believe is impressive.
We’ve done deep analyses of multiple platforms that have brought celebs and top influencers at scale for paid activities and have repeatedly observed this behavior. We’re not yet ready to make those analyses public - but, this definitely strengthens my position that there is no substitute for organic activity from popular creators and that organic creation can only be inspired by all the factors in the note above.
What we are really saying is that a platform needs a mix of the wow factor of a Wombo or a Reface and a collection of creation experiences that are repeatable without fatigue in order to succeed at scale.
It is useful to look at this as a framework of creation experiences rather than one specific experience that can get boring after a while.
In the interest of keeping this post from becoming even longer than it is, I will keep the framework discussion short. Duets in musical.ly was a framework - every new video created provided the opportunity to create new duets with it. Rimix in Rizzle is similarly a framework that scales exponentially with new videos added to the system.
With the onslaught of talented creators, the curators come and then the viewers. Growth now crosses over into the high-scale territory.
We conveniently overlooked one factor thus far - if creators want fame and viewers don’t come until much later, how do creators get the fame they are seeking? Creators and curators alone can only provide a “community” - that scale is seldom sufficient to bring fame or viral moments, especially to the early creators. To understand this, let’s understand how the flywheel effect comes about for modern UGC platforms.
The Flywheel Effect For Today’s UGC Platforms
The flywheel for UGC platforms starts with creation tools, quickly followed by seeding PGC such that the creation tools spark the imagination of large groups of creators.
There is an argument to be made here that if the creation tools were good enough, the collective intelligence of the UGC creator community is sufficient to bring great content. However, this runs the risk of setting up a narrow DNA for a platform or taking a very long time to develop the necessary diversity and breadth.
If Instagram only remained about selfies and never inspired professional photographers to adopt the platform, chances are that the majority of the content on the platform will only be watchable by a small group of close-knit family and friends. Today, we have food Insta or travel Insta, etc., because, via the WHP efforts, the Instagram team actively seeded concepts that the community was not yet thinking about.
An aside: Twitter is the living proof of a lazy platform where nearly every innovation beyond the basics has been driven by the community. If not for hashtags, we won’t have the liberal Twitter, the conservative Twitter, the medical Twitter, and so on today. Even after all the innovation driven by the community, Twitter has done a fairly mediocre job in seizing and shaping these innovations. For having invented microblogging, an innovative Twitter could have been the home for a short video platform. Wait, it sort of was, even if only by acquisition - but, the lack of imagination actively plunged Vine to its demise rather than become the leading innovator in this area.
TikTok achieved the many sides by actively paying early creators (PGC) to bring their skills and create content that was beyond lip-syncing and dance. By simultaneously evolving the creation tools that can then allow all creators to create such content, UGC took over where PGC left off.
The final part of the flywheel is about building a highly addictive viewing experience (often achieved via personalization). But, to get to that stage, the platform needs large amounts of diverse content covering varied interests. So, it is necessary that a platform goes through a stage where it is incapable of keeping the viewers when it accumulates that diverse and large-scale content of high quality before it can keep the viewers.
So, Where Are The Viewers?
The leaky bucket theory in customer acquisition shows that as long as the input is greater than the leak, the platform keeps growing.
The trick to perceived success before real success is to keep the bucket increasingly full at all times. This results in a buzzing platform, with a rapid inflow of new users, even if a huge portion of them are leaking out.
During this phase, it is critically important to keep the creators from leaking out - a platform that leaks creators rapidly has huge problems since the lack of continuity and consistency in creation shows readily. Whereas, as long as the view counts are growing, content does not have to be viewed by the same people to achieve the perception of fame.
I’m sidestepping the issue that there is more to it than view counts. View counts can be forced on platforms where 2s constitutes a view by simply putting the content forcibly in front of people (on the feed). But, in a system that supports likes, comments, and other interactions, those interactions may not achieve optimal ratios, leaving the creators feeling empty. Only when all of it comes together optimally, the fame actually means something. How this can be achieved before the platform can be interesting to large groups of people is another thesis of its own and I’ll save it for another post. It is undoubtedly a chicken and egg problem that must be solved - enough interactions must exist for the creators to feel motivated and enough interesting content must exist for the viewers to interact.
With deep enough pockets, this perfect pace of user acquisition can be created, along with retargeting lost users for repeated brand visibility. Deep pockets also help in achieving great content creation through paid efforts, solving the chicken and egg problem to some extent.
This is exactly what TikTok achieved by spending astounding nine-figure dollars on advertising monthly to quickly get scale. It is very important to note that the creation experiences and personalization evolved at the right pace to both keep the creators happy and the viewers increasingly engaged such that eventually, the leak was smaller than the inflow.
This rapid flywheel timing led TikTok to become the fastest to join the billion users club, in a mere 2.5 years. To be fair, in the chart above, TikTok is the only platform that benefited from having both China and India as early adopters, the two most populated countries in the world. It also had the unfair advantage of the Musical.ly user base to bootstrap the US growth. These factors, combined with deep pockets and willingness to spend massive amounts on user acquisition, helped attain this unmatched pace of growth!
The TikTok level spend is unnerving to many. But, both TikTok and Kauishou have shown that spending $2B-$3B to build a $200B-$300B company is not a losing proposition. It is also lost to many that Snapchat has raised nearly $5B to build a $100B company. Again, not a losing proposition.
Conclusion
UGC platforms have no choice but to be a creation-first play. The creation experiences must be repeatable and scalable, in addition to being at least moderately unique enough to attract creators at scale. To accelerate the creation of diverse and unique content, enough PGC must be seeded to catalyze UGC creation. Lastly, to perfect the flywheel effect, enough viewers must be added to the system to keep the creators satisfied. If all these occur, personalized viewing experiences as the final step can lock in success.
In case it is not obvious yet, this lifecycle of a UGC platform is bound to require more than your average startup funding to create, making it very hard for small startups to compete.
But, it is not impossible. At Rizzle, we are at the brink of crossing that chasm. Stay tuned for the future!
So insightful!