With Brittney Kaiser’s public skewering of Facebook and the “own your own data” movement, there are renewed rumblings that an attack on “Big Social” might have legs.
The real argument for decentralized social.
A single company now controls the newsfeed of 2.4 Billion monthly active users on Facebook and another 1 Billion on Instagram. Another single company, Google, controls 2 Billion active YouTube accounts. This is an enormous, unprecedented amount of power concentrated in just two hands — enough to sway elections (2016) and force massive political change (Brexit).
Facebook has admitted to listening to phone conversations for better ad targeting. Thousands of key people in Facebook groups are bots or controlled fake users designed to influence opinion. Both Facebook and Google also have unprecedented censorship control. Just a few weeks ago, hundreds of prominent youtube crypto profiles were abruptly shut down. The world’s entire information funnel is now controlled by just a few companies.
This is the first and only defendable argument for (and principal of) Decentralized Social. We need a system of information that can’t be easily manipulated by a few central authorities. This is the 21st century version of the First Amendment, the new “freedom of the press”.
A second argument: re-distribution of profits.
When “blockchain mania” hit full steam in 2016, many companies were formed on the basis of “cutting out the middleman” by distributing all of the “excess profits” of big centralized tech companies back to the people. This of course never happened at all for a simple reason: the unit profits of these middlemen is exceptionally small on a net basis(at most $10 per user per year), and only reaches these levels because of all the intrusive practices. In the case of Uber and others, the companies are not actually even turning a profit, there is nothing to re-distribute.
Specifically in the case of Facebook or YouTube, you can’t simply offer the same service in a decentralized way and pass back meaningful profits to all end users. The math does not work.
However there is a caveat, and it’s a big one. In social platforms such as YouTube, the ratio of content creators to content consumers is high. By rewarding content creators and not end users “selling their data”, the math can indeed work.
So how can it work?
All social networks at the end of the day consist of content that can be replicated on multiple servers. This has to be the backbone of Decentralized Social.
- A set of posts (text, images, video) that can’t be censored centrally
- A graph of who follows who and who likes what — again stored in a decentralized way
- A way of paying for this with blockchain tokens
I left the third point deliberately vague — for now. There are a number of different schemes for using tokens to incentivize block producers to replicate data and the social graph. We’ll go into a bit of that later — the key point is that without crypto, the only monetization possible is centralized ad systems, in which we now have two clear category winners: Google and Facebook.
The core of modern social media is based on the idea of the feed. You follow people, and a feed is algorithmically designed to show you what’s new in your world.
As an aside, I had a social network, TagWorld in 2006, and we were one of the first companies to really experiment with this this concept, a year before Facebook used it for global domination. It wasn’t clear then, but it’s completely clear now that the feed is everything. People want all their information in one continuous stream. Facebook, Twitter, Instagram and new TicTok all are designed around the idea of the feed. Real decentralized social needs to address the feed.
The simplest way around this is to use the Twitter model: show all the content of everybody you follow. This decentralizes easily — machine learning algorithms to “suggest” content doesn’t.
Alternatives to the Feed: #tags.
A much less successful approach to browsing data in a feed is to hop around different hashtags or categories. This is effectively the reddit model (which evolved from the late 1990s company Digg). In this model every category like “btc” has a home page with posts that can be voted up or down.
Although reddit is going strong, this UX clearly feels outdated. You need to manually jump around between all the different forums that interest you in order to find content.
Steemit, founded by Dan Larimer was the first attempt to decentralize this model. The UX is very similar (and feels equally outdated).
In this model, the posts are stored on a decentralized blockchain, and users pay Steem tokens to promote a post to the top of the stack.
Voice, a project that is expected to get in beta soon, seems modeled after this metaphor. There is a strict list of tags, and you can use “Voice” tokens to push content to the top.
I would like to emphasize that both Steemit and Voice seem to be targeting Reddit much more so than Facebook or Instagram in their product design, although the marketing “Social as it should be” would suggest otherwise.
It’s important to note that Voice will initially launch without video. What that means is that people who wish to share a video link will have to host it themselves — a major consumer disadvantage.
Storage and Video Streaming Costs
Streaming video on the web has three distinct components
- the cost of ingesting the video (about $0.0085 per GB)
- the cost of streaming the video (about $0.085 per GB)
- the cost storing the video (about $0.023 per GB per month)
The big cost is streaming. To see this consider a Social app with 1,000,000 videos each of which has 1,000 views.
ingesting the 1,000,000 videos into a service like Amazon costs about $83 (9,765 GB of data , a cost of $0.0085 per GB)
Storing the 1,000,000 of these videos for a month costs about $224 per month (9,765 GB of data, at a monthly storage cost of $0.023 per GB)
Streaming these videos will cost about $83,000 (9,765,000 Gb of data, at a consumption rate of $0.0085 per GB).
Clearly, the big cost is streaming. Just the streaming cost for such a setup is likely to exceed a million dollars per year.
Storage and Streaming Providers
A completely decentralized social media platform could exist with little actual content replication. Several “streaming providers” could easily store the data at clouds like Google or Amazon and be paid by the decentralized social media platform in tokens.
From a consumer point of view the user could simply upload files as in Facebook or YouTube. These files would then be uploaded and either stored on multiple “streaming providers” with redundancy.
The notion of a “streaming provider” could be somewhat similar to a “block producer” or “validator” in Proof of Stake Consensus models such as EOS or Cosmos / TenderMint. This is a step to having a truly decentralized system.
The role of tokens
At the end of the day, there really is only two options for monetization in a Decentralized Social platform
- charging users to either “like”, post or read content.
- a form of decentralized ads (pay for position)
A portion of token proceeds could be used to cover the cost of storing the data (paid to “streaming providers”) . At the end of the day somebody has to cover these costs, although they might get absorbed by a company like dLive who has significant traction in the video streaming space.
A second use of tokens is to reward content creators whose content is “liked”. This is an important point, mentioned earlier. It’s been successfully tested in apps like Steemit and Karma and is the basis for Voice:
People who like content are fine with paying tokens as a way of expressing their like. These tokens can then be passed on to the content creators.
In the case of dLive, the token model is designed as a means of tipping video streamers. By attracting influencers such as Pew Dee Pie, dLive has managed to get 5 million monthly unique views per month — benefiting Pew Dee Pie directly, as well as the dLive and Lino platform.
Charging users to post or even just to access a decentralized social network is another untested use of tokens. A new platform Chirp, being developed by Kurt Braget’s team in Los Angeles is testing that payment use case.
Steemit and now Voice both have the concept of “surfacing” or “voicing” posts that have been liked with tokens. From a UX perspective, we are back to the Reddit model.
Chirp is experimenting with a right column, similar to Google’s “sponsored posts” where advertisers can promote content to the community.
We are at the beginning of something that may or may not turn into a significant threat to Facebook and Google. Clearly, both companies are in incredibly strong positions, with network effects the likes of which have never been seen in human history. Still, there is always room for new players, and with the right user interface, either feed type solutions or #tag type solutions could gain traction in the next 12 months.
Appendix: DSocial Companies Mentioned here:
Voice.com, the social #tag platform being developed by Block One
Steemit.com the original reddit competitor with blockchain rewards, also created by Dan Larimer
DLive.com a new twitch like streaming platform with cryptocurrency rewards (but not really any more decentralized than twitch)
Karma a social rewards “for good” platform which started on EOS and now is onthe WAX platform.
Chirp an upcoming social posting platform
this is a work in progress. Please message me at lynxfred on telegram if you want to add to this.