DNA Of Short Video Platforms - Part III
In Part I, I discussed how the usage patterns and content consumption models are impacted by the content duration and dimensions on various platforms and how short video platforms fit in. In Part II, I wrote about the creation process and how short video apps have disrupted creation in ways that other apps, including YouTube, had not imagined. Further, I covered how popular music has now become table stakes for short video platforms to be taken seriously.
In this final part directly focused on the DNA of short video apps, I will discuss the move from search to discovery and the associated changes these platforms have brought to social and content at large.
Self-Contained - Creation, Distribution, Monetization
Before I directly address the behavior of short video apps in this area, let’s explore two key aspects of existing platforms from the previous decade.
Creators need to be master of many trades
YouTube, until very recently, was a distribution platform, and that too, a limited one at that. Creation happened offline (as I discussed in detail in Part II) and content was uploaded to YouTube. If many things were done right, there is a small chance that a new piece of content from a new creator is picked up by YouTube’s recommendation algorithm. What are the many things?
Content Quality: Every piece of content is competing with exceptional quality videos coming from millions of creators a day, some of whom are exceptionally skilled in creating great content that is optimized for their own audiences and YouTube’s algorithm. As a new creator, it is important to survey and understand the landscape, pick your niche, and create the best possible content in that niche.
HD Equipment: This follows from content quality, but often producing high quality landscape content means owning camera and associated equipment that can help produce that high quality content.
Exceptional Editing: This is another corollary of the content quality, but it is worthwhile calling this out since it requires a creator to be well versed in professional editing tools. Jump cuts, background music, b-rolls, text, and other effects that improve the visual appeal of the content are essential for success.
Thumbnails: Since YouTube is a browse/search platform (I’ll talk more about this in the next section), the thumbnail plays a huge role in the potential of the video being watched. Creators often use Adobe tools to create amazing thumbnails that stand out and have the clickbait effect in order to maximize clicks.
SEO - Tags: In order for the recommendation algorithms to pick up a video, it needs to be properly tagged with the right topics and keywords. This often requires research to understand the search volumes and optimizing for the sweet spot that has enough search volume but does not have too many search results.
Actual Content: Needless to say, in order to hold the viewer attention for longer than a few seconds, the content itself needs to be highly captivating.
Content Licensing: For highly engaging content, creators use sound effects and background music, which may need to be licensed.
Growth Marketing: Even after having the right tags, thumbnails, and content, a creator has to bootstrap traffic to their channel. Other than telling friends and family, there is plenty of advice out there to like, subscribe, and comment on other YouTube videos with the hopes of being noticed.
In-Video Marketing: In order to get viewers to subscribe to the channel, creators need to often explicitly ask to subscribe (without sounding desperate or taking too long), get savvy with using the YouTube tools to add the right call to actions at various points of a video, and potentially in the description.
Sponsorships: Today, for many creators that are not monetized well on YouTube, third party platforms like Patreon is an option to get sponsored by fans. But, for this, the creators need to ask their viewers to sponsor them on another platform.
WikiHow lists 15 essential things to do to be a good YouTuber. Even so, that list is not exhaustive!
Follow-based networks require a lot of self-promotion for growth
Until TikTok changed the game, social networks were exclusively based on a follow model - which means, you had to grow your followers to be noticed. This puts a huge burden on creators to even get started. Some apps tried to make it simpler by ingesting contacts and Facebook friends to bootstrap the social presence at the expense of privacy. However, this assumed that people we know are among the interested audience for the content we make. This assumption is often false.
Creators not only needed to make great content but also shout from the rooftops about their profiles and channels, which does not come naturally for many creators.
Apps built “recommended follows” to suggest profiles to follow. Apps like Instagram based that on your existing social graph (using a ‘people you follow also follow’ logic), but that again assumes a people-centric graph rather than a content-centric one. Other than that, celebrities and popular influencers get recommended, thereby living up to the success breeds success philosophy.
As a platform grows big, it gets progressively harder as a new creator to be discovered and rise to fame. A Pex study showed that 1% of the YouTube channels generate over 80% of the views and nearly 90% of videos don’t reach 1000 views.
Due to the monetization prospects that are unique to YouTube, creators continue to aspire to build a presence there. More about the creator economy in another post.
Over time, subreddits dedicated to YouTubers have become popular, where YouTubers support each other with tips to grow their channels.
In order to build a presence on YouTube, several external platforms, tools, and skills become essential - from Adobe tools to Reddit and more!
Short video platforms are self-contained
TikTok and Kuaishou changed the game by being self-contained platforms. Likee and others followed the same model as well, making this aspect now a part of the DNA of short video platforms.
First, these platforms provided amazing creation tools, as I discussed in detail in Part II of this series. This enabled any creator to start creating content on the platform without elaborate knowledge of filming equipment or editing tools.
The ability to discover trends, reuse audio tracks, create duets, and more allows creators to naturally become part of trends and create memes faster, even if they don’t have the originality to create amazing original content on their own.
Next, perhaps most importantly, these networks changed the follow model on its head to build real discovery. By bringing a content feed and learning quickly about content preferences, these apps, especially TikTok, allowed anyone to be rapidly discovered if they created great content. This has led to a phenomenon and a massive increase in the number of creators on the platform.
Personalization is handled differently by TikTok and Kauishou - the latter seems to ultimately rely on those you follow to populate some of your feed, while the former more closely follows content you are likely to watch. On TikTok, this puts a lot of pressure to create exceptional videos that are likely to make the FYP (For You Page). Recently, some creators have discovered that sticking to similar content as the one that went viral helps, albeit at the detriment of creator enthusiasm.
Personalization and relevance in short video apps is a topic that deserves its own blog post. There are many considerations including semantic matching of content being watched, creator relevance and affinity, music affinity, and even the viewer mood. A big problem with this model is also bootstrapping for a new user to determine the viewing preferences.
Given the complete lack of tolerance in viewing, a platform has on average about 10 seconds and 5 posts to really keep the user around to serve more posts.
Hence, these platforms must rely on mass popularity and conventional tastes to serve the right content in the early days. TikTok came under fire for having suppressed posts from “ugly” people, as recently as in 2020.
In a world where cameras include beauty filters, is it a real surprise that feed algorithms looked for conventionally attractive people? As these platforms get closer to show business, the answer is tough.
In follow-based networks, some guaranteed propagation of content can occur based on people that are following a creator. Platforms like Twitter and Instagram have had to shift to relevance based feed from chronological feed in order to serve more interesting content and keep the users around. Hence, when there is abundance of content, some method of ranking content becomes essential.
When that follow-based relevance is almost completely removed from the equation, the goal for these platforms is essentially no different from show business where the only thing that matters is that the content is good enough to be watched!
This oversimplifies the model a bit, since unlike OTT content, which is truly a product of show business, short video content is generally highly interactive. The interactivity is borrowed from social network behavior and this is generally why these platforms are grouped in the same category as social networks.
That said, everything starts with the content being watchable and builds on it. In actual social networks, the social relevance of the person posting the content plays a larger role.
Content driven platforms need to build a large enough repository of highly watchable content to retain viewers on the platform. This creates a chicken and egg problem where creators expect large audiences in order to continue creating while viewers expect exceptional content to stay.
This is a complex dynamic that leads to a leaky bucket that needs to be filled faster than it leaks in order to retain the creators while simultaneously growing the perceived audience at any given time. This very nature makes it a cash intensive operation starting from the very early days in order to grow to large levels.
Yet another deep topic for another day!
One problem with the content-based feed is that there is really no need to follow anyone. Many viewers are simply there, asking to be delighted automatically. In an anecdotal survey of Gen Z teens, I learned that the bar for following a creator on TikTok is extremely high and that they rarely do it. Since this survey pool is relatively small (~10), I won’t yet draw any conclusions from it, but it does logically make sense that the compulsion to follow creators will be naturally lower in a feed of infinitely amazing content that keeps showing up.
These short video apps have also brought some monetization opportunities for creators, making it a one stop platform for creating, distributing, and monetizing videos.
TikTok and Kauishou provide monetization options on the Live side of things. I will not include the creator funds that have recently become the highlight of these platforms - that is very targeted monetization that lacks transparency and sustainability. I will treat that as a topic of its own in another post.
At Rizzle, we went one step further and integrated fan sponsorships into the platform with micro-payments. This brings a Patreon-style model integrated into the app itself for creators to build loyal sponsoring audiences.
The future of these platforms is definitely self-contained, with the goal of making things easier for creators and allowing them to focus on creativity!
Discovery, Not Search
Search is great when you know what you’re looking for, but for those moments when you just want to be entertained or informed about what is popular, discovery is key. YouTube is sometimes described as a search engine - which has some truth to it with it being the third largest website that makes up for about 3% of all searches on the Internet (Google Search serves 68% and Google Images serves another 20%, making Google home to nearly 90% of all searches).
YouTube approaches discovery with a browse approach, where a feed of content is presented and left to the user to choose to start playing. However, YouTube’s autoplay recommendation can continue the play with related content once the user starts watching something.
Facebook and other social media apps already changed this model with a discovery centered feed. However, by relying on the user to turn up the volume and having volume off by default, these platforms did not solve the discovery problem for video. Even if a video was presented, it needed to rely on visual stimulation to get the user to turn on the audio. Studies have shown that over 85% of the videos on Facebook are watched without audio!
TikTok and other short video platforms changed the game here by leaving audio on by default. Being music driven platforms, this was essential to the success of these platforms to a large extent.
During the early days of Rizzle, when the US was still getting used to acknowledging that TikTok is beyond cringe (or that cringe is in fact what we wanted to watch all along), we received a number of requests to mute the audio on Rizzle by default - or at least provide an option to do so. The video behavior that the US population was used to at scale was from Snapchat, which also chose to leave audio off by default.
As is necessary in product design, that was one feedback we chose to leave out, for good reason :).
The battle of default audio status is also a cultural one. In China and India, where talking in crowded places or generally loud behavior is part of the culture, the adoption of a platform that left audio on by default was automatic.
Despite cultural differences, TikTok did leave audio on by default for the US market - this is because the discovery and content value will be severely hindered without it, undermining the success of the platform itself. At Rizzle, we had little doubt that we needed to adopt the same principle.
Today, Instagram Reels and Snapchat Spotlight are struggling with this behavior, since there is a DNA conflict between what things were in the past and what they should be now!
YouTube has always been a platform where the video was explicitly chosen by the user to play and hence, the audio turned on when the play started. YouTube Shorts has continued this behavior by turning short videos into a browsing experience. More about that at some other time.
But, I want to draw attention to a crucial difference between the browsing nature of YouTube and its follow-up recommendation model and the TikTok style discovery. YouTube’s recommendation tends to take users into a rabbit hole, finding videos similar to the one they just watched. The TikTok discovery model presents a range of videos that are not necessarily related to one another, but rather related to what the user likes to watch more broadly.
The length and nature of content brings other differences which need to be explored in depth to really do justice to it. For example, shorter the content, the less it makes sense to spend time searching for it. If we’re about to watch a 30 minute show, it makes more sense to spend a few minutes looking for it or being particular about what we’re watching.
These patterns bring some expectations on short form content, in that it is increasingly important to get to the point and present something that instantly grabs the attention. This makes short videos more of a performance, requiring the creators to be performers of high caliber. More about this when I write about what constitutes good content.
But suffice to say that we are squarely in the age of discovery and there is no turning back now!