Idea: Sprocket (advertising + user rewards)

[mochet]

Sprocket

Sprocket is a concept that has two aspects--advertising space and community rewards. It is an idea first and foremost and doesn't delve too deep into technical aspects. It may be achievable directly on the Joystream platform or may have to utilize external smart contracts to achieve what is described.

This idea is based around video content, but could work for text, audio, images, GIFs and various other types of media.

• Adspace Sprockets are a purely commercial space, where a user or company can purchase advertising using a mechanism on a specific hashtag. This hashtag can be used in a variety of creative ways that expand beyond what is possible on a typical video platform and there are use cases that exist outside of Joystream itself and could possibly work in the field of physical displays.
• Community Sprockets are a hashtag where users can participate in informal communities centered around categories of content that they enjoy. In general, users pay some tokens to get their video into a pool of videos and the userbase then votes on a "winner"--the winner is rewarded with tokens and also with possible prominent placement of their video. This "winner" is also added to the feed of content, meaning that new "winners" are helping to build a highly curated list of interesting content. This idea is similar to a TCR (Token Curated Registry).

1.1 - What does Sprocket mean?

A Sprocket is a component that is integral to all manner of machinery, be it a bicycle, car, lathe, printing press or camera. Sprockets are what allow machines to be consistent and to abide by the laws of time.

In the world of film & photography, a Sprocket hole (https://en.wikipedia.org/wiki/Film_perforations) is the name given to the perforations that run alongside a strip of film. A projector uses these Sprockets to push through film at a constant pace, allowing the still images to form a consistent optical illusion that forms one of the most widely available forms of media today: movies.

On a rangefinder camera, a photographer has to use a film advance lever to manually pull new film into place before clicking the shutter and taking a photograph. With Sprocket, we take the same idea of advancing film and try to build a social and/or commercial (advertising) mechanism for this action to take place.

A YouTube video showing various film advance levers: https://youtu.be/NPTRfMUeXQU?t=91 (Title: Leica Camera Study 1: The Film Advance Channel: https://www.youtube.com/channel/UCi-NmFLJaEsJF29T2No02uw)

Sprocket would use the traditional idea of hashtags and categories seen on most social media websites and add a set of functions to allow for advertising, community rewards and hopefully a very effective mechanism for producing highly curated and interesting collections of content.

Specifically when it comes to community rewards for creating content, Sprocket would ideally offer a means for tokens to be distributed to a far larger range of users compared to using the proposal system.

1.2 - The growth of self-advertising

Generally on video platforms, there is no concept of groups. Users typically upload on their own channels and build engagement with other channels via commenting or through other platforms like Twitter. Uploading videos helps a channel to build a library of videos, but this doesn't usually help to build communal spaces as anyone is free to use tags when listing their videos. There is little disincentive for parties to not spam tags and there is no concept of these tags being curated by the users of the platform.

This also results in it sometimes being difficult to find content worth watching, if a user is interested in cooking videos there are simply far too many channels creating this type of content and search results are personalized and also influenced by an algorithm that means smaller, perhaps better channels find surfacing their content to a wide audience to be almost insurmountable task.

If we look at the methods available to video creators generally these are tied to metrics such as view counts which is a very limited perspective when trying to attract and reward a diverse range of content. Critically, content that may be of high quality and have a passionate userbase may not be financially sustainable which pushes many creators towards other means of income such as crowdfunding platforms and sponsorship deals. Major YouTube channels have also taken on an increasingly corporate nature and can be seen taking over large parts of the platform, this leaves less incentive for new and novel creators to participate as the chance of gaining significant viewership is largely diminished.

On platforms such as YouTube, advertising seems dominated by major companies and depending on the user's region there may be very few instances of self-advertising (meaning a regular user paying for their videos to be placed as pre-roll advertisements). Instagram and Twitter have had far more success in allowing users to self-advertise their content, these advertisements seen in a users personalized feed are marked as "Sponsored" but importantly this self-advertising content is not distinguished in any way from commercial users of advertising. Thus we can see that there is demand for users to be able to access features typically only available to large corporations for the purpose of self-advertising.

TikTok, which is now one of the leading video platforms has a creator fund (https://newsroom.tiktok.com/en-gb/tiktok-creator-fund-your-questions-answered) but the method of distributing this large pool of money to creators is very limited and requires a creator to have more than 10,000 followers to apply for funding. These kind of arbitrary restrictions on distributing rewards to a userbase mean that novel users are disincentivized from participating.

Simultaneous to this transformation in advertising, there are now examples of large corporations using features intended for users for the purpose of advertising: https://www.bbc.com/news/newsbeat-53862091

In public spaces, there has been a significant rise in digital billboards, which are LED screens that are internet-connected and controlled by a company. Companies seeking to advertise on these billboards are required to contact the company which controls the billboard and ask for a quote and there are no public prices available nor any direct mechanism for submitting and paying for content (an example: https://www.outfrontmedia.com/).

In these ways, the line between corporate advertising and self-advertising has become increasingly blurred and the key method of retaining users (by rewarding users who create the novel content that attracts advertisers in the first place) has not seen any substantial changes. This is the reason why Sprocket seeks to provide functionality for both rewarding the community and also allow for advertising--these two aspects increasing share a lot of common ground and rather than make a system that caters solely to one side of the equation it is probably more beneficial to try and cater to both.

By adding a group concept to Joystream where likeminded users can gather and work together to curate content and reward its creators, it would hopefully spur a significant number of new users who contribute novel content and work to build Joystream into a sizable platform that is attractive to advertisers.

2.1 - Adspace Sprocket

An Adspace sprocket is a purely commercially intended place, where a user or company can pay tokens to purchase adspace for a period of time. There is no distinction in how a user or company would interact with this system so it can be used for corporate advertising (for example, to advertise a product) or for self advertising (for example, to advertise a users latest video)

2.2 - Adspace Sprocket components

• sprocket_adspace = This is the name of the Adspace Sprocket a user can select from. There may be several of these. It may cost users tokens to create or maintain them. This adspace could be shown in a variety of places, including preroll ads, embedded (such as on a section of the forum or a video creator's profile page), physical displays (such as an LED screen positioned in a place with high foottraffic).
• type
  • billboard = This type is linear and has a single layer, it is similar to a billboard or single TV channel in that there can only be one piece of content at one time.
  • registry = This type allows for simultaneous pieces of content to be registered at the same time. The user client would then randomly select one or more pieces of content to show in succession.
• rate = This is the rate at which advertising is charged to be shown. It may be a figure which is quoted in hourly terms. It may be set by a council or user group. It may be a minimum figure and users then bid higher amounts to actually win placement.
• burn_percentage = In order to have the concept be profitable for the network, there may be a percentage of tokens spent that are burned.
• minimum_duration & maximum_duration = These are the min/max durations that an advertisement can occupy adspace for. For example the minimum value might be 24 hours meaning a user paying for advertising cannot buy adspace for any period shorter than this number.
• start_blockheight & end_blockheight = The blockheight range a user would like their advertisement or content to be shown.

2.3 - Personas / Use Cases

• Michael owns a cafe and has a lot of foot traffic, in order to try and capitalize on this feature, he purchases and installs two LED displays which are connected to the Joystream network. He creates the hashtags #cafe_display_1 and #cafe_display_2 and lists the ideal formatting.
• Jimmy is a livestreamer, he has a decent viewership and wants to experiment with advertising. He buys a physical display which he features prominently in the background of his livestream and creates a hashtag for it. He also creates another hashtag which is embedded on his livestream output and livestream homepage. Users and companies pay to list advertisements and he recieves rewards from these advertisements.
• Alice runs a cooking channel on Joystream, she isn't happy with the terms of regular advertising, so she makes it so her channel runs ads from the #cooking_ads Sprocket as she feels the advertising is more relevant to her viewers.

2.4 - Adspace Sprocket limitations

• There is currently no concept of moderation, in order to prevent abuse, it may be beneficial to only allow content which has been approved by a working group to be allowed as advertising.
• There is currently no concept of ownership. It may be that a user can create an Adspace Sprocket and recieve rewards from users listing ads, how ownership would work would have to be decided.

3 - Community Sprocket

A Community Sprocket is a an area where users can submit content, have it voted upon by other users and potentially recieve a reward in the form of having their content highlighted or some amount of tokens.

Community Sprockets can be considered as social barometers, for each community that has a presence in this space, the most recent content would highlight the current state of the community and the historic content would act as a rich history of the community.

Some diagrams to visually explain the idea:

3.1 - Community Sprocket components

• community_sprocket = This is the name of the community Sprocket a user wants to participate in. There may be several of these. It may cost users tokens to create or maintain them. This Sprocket could be displayed in a variety of places but is not a commercial space.
• submission_deposit = This is the amount of tokens a user has to pay to submit their content. This may be in the form of stake, or it may burn tokens or a percentage of the tokens could be used to fund the Community Sprocket.
• submission_cycle = This is the length of time each "competition" takes place over. So if it is 24 hours, that means a winner is decided once every 24 hours.
• winner_payout = This would be the amount a winner recieves. This amount could be tied to the number of submissions and the amount of tokens used when submitting. In general, the amount a winner recieves would be not be a huge amount with the expectation that users recieve more rewards from having preferential placement of their content.

3.2 - Personas / Use Cases

• Tareq enjoys making videos for his gaming channel. He has a small, dedicated viewership and he has heard through a friend on the platform that #videogames would be worth trying to submit a video too. He pays some tokens to list his video. A few hours later, if enough users vote for his content it will be shown on a section of the Atlas homepage. and he will recieve a bonus amount of tokens. Due to the incraesed viwership he recieves some new viewers on his video.
• Matteo has been a longtime fan of another user's cooking channel on Joystream. He really enjoys the latest video and decides to spend some of his own tokens as stake to get the video on #cooking. He uses the comments and forums to let other fans of the cooking channel know about this, and the video ends up winning on cooking for the next 24 hours.
• A group of ~50 users on Joystream make an informal community and they post videos using public domain and popular content that is repurposed and remixed. Although this content does not usually have high viewership, this community occassionally produces videos that become virally popular. They are a highly discerning group and they run their Sprocket as a tight ship, only approving of videos that they find to be funny.
• A video game developer who creates small games and sells them on various stores wants to try and experiment with his latest game. In order to try and get people posting more content of his game, he funds a Community Sprocket with tokens. In time once the funding dries up, footage of the game remains popular and the people posting videos end up funding it further.
• Layla has never used Joystream before, she accesses the Community Sprockets interface and finds a diverse range of new content that has been curated.

3.3 - Social Gambling

Community Sprocket's revolve around Social Gambling. This is the act of a user submitting content and taking the risk that a niche community may love, hate or be indifferent to what they've created. There may be a financial payoff, or it may just be a case that the reward of having content highlighted is enough to satisfy.

3.4 - Community Sprocket thoughts & limitations

• Picking an initial category of content for this idea would be challenging, I think a perfect category would be public domain content which is free, easy to find and will also be altruistic in nature.
• The concept of ownership and how rewards are handled has to be figured out. This includes whether there are fees to create a Community Sprocket or whether the attributes, such as payouts and minimum stake requirements are flexible and can be changed after the initial creation.
• It may be that an Oracle role is needed for management of Community Sprockets and in order to prevent abuse, or it may be necessary to have some mechanism for disqualifying off topic content.
• A lot of this idea relies on the userbase being authentic, therefore it is assumed that the onboarding of members will be honest and that actors on the platform would not easily be able to perform a sybil attack. In general, it should cost more to vote for content than the reward is so that manipulating the system for profit is not worthwhile.
• Once there are a sufficient number of Community Sprockets, it may be possible to have some mechanism where interested users can invest in a Sprocket and somehow capture a percentage of rewards from winners. This has not been explored yet.

4 - Other potential Sprocket types

• Threshold based Sprocket - This type of Sprocket would be similar to a Community Sprocket but would have some threshold of votes after which a "winner" is decided rather than being on a fixed cycle of time. This would be a less regimented, more creative type as the timing of new content would be less predictable. This type of Sprocket could work well for categories of content where there are several submissions on an hourly basis, possibly for something like video games.

5 - Other ideas

• A Community Sprocket for the Joystream logo could be an interesting idea similar to Google Doodle.
• A component of both Community Sprockets and Adspace Sprockets is that they can be connected to physical displays (like OLED or ePaper displays). It could be worth investigating making a competition for this to spur innovation.

Acknowledgements

Thank you to both Martin and Bedeho for helping to guide this idea.

Changelog

Sprocket has been in closed development since approximately May 2019. In that time it has moved wildly between a competition for ASCII art displayed on node startup to postage stamps, fixed dimension content, novel display technology, typography, audio hardware concepts, musical patterns, physical rewards, virtual (NFT) rewards and just about everything inbetween. During that time several informal interviews have been conducted and the concept has been explored with great depth to decide how to best approach the idea.

• v0.1 - May 2019 - Initial idea with hardware node attached to ePaper display. The primary focus of this was distributing hardware to competition winners (the best public domain content) that they could use to run a node, this would include a display that features content "live" from the JS network and is also incorporated in the JS interface.
• v0.2 - May 2019 - Idea of user proof in form of an ASCII competition, the ASCII art would be uploaded to JS and also feature in node bootups. Proved to be technical infeasible.
• v0.3 - November 2019 - Idea changed to be about the idea of a central social barometer rather than public domain.
• v0.4 - November 2019 - Idea expanded to focus on user, group and community owned sprockets. Also included an element of physical rewards (like merchandise or hardware). Focused greatly on the idea of hashtags.
• v0.5 - January 2020 - Added initial Sprocket concepts beyond public domain, added concept of visualizing content via weighted vote data.
• v0.6 - June 2020 - Idea split up into phases. Mainly based on concept of UGC + rewards. Initial phase would just be the public domain Sprocket. Added element of council controlling overall control of sprockets.
• v0.7 - July 2020 - Idea begins to incorporate the idea of TCRs, hardware concept now expanded to be more broadly relate to hardware and software integrations. Added emphasis on user vote vs council voting. Added basic diagrams to explain aspects. Idea now will be backed by user voting on forums and out of scope section added to reflect what isn't technically possible on forums.
• v0.8 - August 2020 - Major rewrite and settled on 3 main elements (RAI = rewards, adspace, integrations). Expanded advertising element to be user accessible.
• v0.9 - November 2020 - Removed integrations
• v1.0 - December 2020 - Rewrote entirely and simplified the concept.

TCRs

These links cover TCRs and some of the ideas surrounding them:

• https://medium.com/paratii/token-curated-playlists-1-thoughts-on-staking-and-consumer-applications-2a50bc837a94
• https://relevant.community/
• https://github.com/Paratii-Video/tcr - Token Curated Registry
• https://medium.com/@ilovebagels/token-curated-registries-1-0-61a232f8dac7 - Token Curated Registry
• https://hackernoon.com/what-are-token-curated-registries-and-decentralized-lists-d33fa42ba167 - Token Curated Registry
• https://kauri.io/incentivizing-highquality-curation-with-a-tokencur/5d256b3a16c3430080718f29d6758366/a - Token Curated Registry
• https://github.com/miguelmota/awesome-token-curated-registries
• https://dogesontrial.dog/how-it-works - TCR of doge images, pays out remaining fees to "winner"
• https://github.com/gautamdhameja/substrate-tcr - Substrate - Token Curated Registry
• https://github.com/gautamdhameja/substrate-tcr-ui - Substrate - Token Curated Registry

┆Issue is synchronized with this Asana task by Unito

[bwhm]

Hey @mochet , Read this a couple of times now, and the current implementation (of the Community Sprocket) looks really promising!

I say Community Sprocket, as that seems like the obvious place to start, but please correct me if I'm wrong :)

I don't know how you prefer to proceed, but we would gladly fund the initial sprocket, and provide some resources for the organizer. I'll write up more details tomorrow, but an initial question from me:

What changes would be needed to the content directory, atlas and/or pioneer would be needed to make this work?

Seems like at least one new category and one new class in the former, and perhaps some minor tweaks in atlas depending on how they should appear. From previous iterations, it seems the forum could be enough to arrange?

[mochet]

Hey @mochet , Read this a couple of times now, and the current implementation (of the Community Sprocket) looks really promising!

Thanks!

I say Community Sprocket, as that seems like the obvious place to start, but please correct me if I'm wrong :)

I would say it would be an interesting experiment for sure. The adspace one could be something for later on.

I don't know how you prefer to proceed, but we would gladly fund the initial sprocket, and provide some resources for the organizer. I'll write up more details tomorrow, but an initial question from me:

What changes would be needed to the content directory, atlas and/or pioneer would be needed to make this work?

Seems like at least one new category and one new class in the former, and perhaps some minor tweaks in atlas depending on how they should appear. From previous iterations, it seems the forum could be enough to arrange?

• It would require a playlist or category that one or more people can edit.
• Hopefully it would have some listing on Atlas that can be linked to easily
• If the forum was used to manage it, ideally the forum reactions would be introduced, but I'm not sure how far away that is from happening.

Since I'll be away for a while I'll leave it up to you on how to proceed, but I think two of the very promising ideas for an initial community sprocket are:

• A public domain sprocket (since its available freely it is very easy to manage this)
• A joystream logo sprocket (this might be still images, which I'm not sure Atlas supports, there is also the question of making unofficial logos early this on, but the general idea would be something like Google Doodle.

Probably the public domain one is the most realistic at this stage.

One unanswered question is how users would vote on community sprockets, this could be achieved by using the forum poll or forum reaction system, but obviously to automate the system would be a huge undertaking at this point, so maybe if someone who is interested manages the initial stage it can be an opportunity to see how the system might work if it were implemented in a manual fashion.

[mochet]

@bwhm I thought more about this and maybe one solution is to have the CLI be able to post/read the forum. I'm not sure if that is something already planned, but I could see it being a useful function for not only this but for a variety of other things like budgets (https://github.com/Joystream/community-repo/issues/68) and council reports.

If that functionality was supported then I guess there could be a bounty to build a script which creates a thread, and then the script is run again after some time period (however long the competition is) and it can automatically parse through submissions in some defined time period and see how many votes each entry has.

There could also be other bounties for creating scripts for things like creating forum threads once per week which list the top 10 most viewed videos and other ideas that could help to build more interest in parts of the platform.

[oiclid]

@mochet This is my take, in response to the "expansion of ideas" bounty (https://github.com/Joystream/joystream/issues/1966).

My current ideas are more about execution and not about developing the Sprocket

Model

The one ad model I have been most impressed with in recent times has been that of the [Brave Browser](https://brave.com). Essentially, users choose whether or not they want to see adds and get paid i(in crypto) for it. They are then shown ads from (I assume) a curated set, or have their landing pages display artwork and images from the social media accounts of certain users.

1.1 Formats

Joystream could implement this, with different rates for the different media, with video costing the most, followed by gif, images, hashtags, and standard text. Viewers could also earn more for every clickthrough, although a system would have to be put in place to prevent, or at least reduce, click fraud

1.2 Wallets

To be able to let users view and access their tokens, Joy, would need a wallet. This could be an existing solution, a customized one, or a mixture of both.

1.3 Tipping

Users should have the ability (and choice) to tip any content they like. While this would most likely need separate infrastructure, I believe that the fact that it would enable users to support their favorite creators would make it worth the trouble.

[Edited last paragraph to add further clarity]

[mochet]

The one ad model I have been most impressed with in recent times has been that of the Brave Browser. Essentially, users choose whether or not they want to see adds and get paid i(in crypto) for it. They are then shown ads from (I assume) a curated set, or have their landing pages display artwork and images from the social media accounts of certain users.

This is part of the idea, meaning that a user can elect which Sprockets to display or a specific UI could have some predetermined ones. The difference between showing content from specific users and what Sprocket aims to achieve is that Sprocket would be awarding a wider range of users via category than just an individual channel.

Whether or not a user wants to see ads is an interesting question though... since half of this idea is commercial advertising and half of it is "community spaces". But there is no concept of being paid to view ads and I don't think such a system is worth doing because of elements like fraud. In itself the typical advertising world isn't working at all because of things like fraud and because online advertising in itself is not proving that effective anymore. As an example:

https://thecorrespondent.com/125/the-non-sense-of-online-advertising-when-the-numbers-dont-add-up/267513125-ebeb97f2 P+G which has one of the largest advertising budgets on earth, reduced ad spending by $200 million and saw sales increase.

That's why when looking at Sprocket it is important to consider that it is not looking to replicate the traditional online advertising that other platforms have, instead its more like a system "by the users and for the users" and where users can compete for advertising space alongside major companies. So in short the idea of rewarding users for viewing ads is something I do not think is at all worth looking at just because online advertising is already going through rapid transformations that mean people are really starting to look at how effective the money spent on ads is actually is (and this includes paying users to view ads)

There could be a bidding system separate to sprocket for prominent placement of each community's sprocket though, which is an interesting idea (this would only be worth investigating if the initial idea works)

Joystream could implement this, with different rates for the different media, with video costing the most, followed by gif, images, hashtags, and standard text. Viewers could also earn more for every clickthrough, although a system would have to be put in place to prevent, or at least reduce, click fraud

The thing is that this system (depending on how it was implemented) wouldn't directly reward for clickthroughs or viewcounts at all. Viewcount rewards might be a separate thing on Joystream but it would be its own system completely separate to this system. AFAIK nothing has been officially communicated from @bedeho or the team at Jsgenesis about how the ad system/view counts might work in terms of rewards.

A viewcount system definitely has a place for rewarding users, but for a healthy platform, it should be one of many options available for users to get rewards.

To be able to let users view and access their tokens, Joy, would need a wallet. This could be an existing solution, a customized one, or a mixture of both.

Since the platform is built on substrate there may be some 3rd party "wallet" solution. There is currently polkadot-js which a user can store keys on (https://github.com/polkadot-js). One big point of Joystream that makes it interesting for me is that a user doesn't actually need a wallet, they can just load up the UI and connect to the network and manage their own keys without installing anything on their own system.

Users should have the ability (and choice) to tip any content they like. While this would most likely need separate infrastructure, I believe that the fact that it would enable users to support their favorite creators would make it worth the trouble.

I agree, but tipping will be a separate thing to Sprocket. Sprocket would be an indirect method of tipping categories of creators though. I'm not sure how tipping is planned to work, but I am sure that it is potentially being developed.

[mochet]

Very relevant:

• https://vimeo.com/480331322
• https://vimeo.com/473732449
• https://rubenfro.com/dissolving-realities-shibuya-takeover
• https://mobile.twitter.com/YOON_AMBUSH/status/1375276278576607237
• https://www.youtube.com/watch?v=vzZApqoOLxg
• https://www.instagram.com/neoshibuyatv/
• https://www.paoolea.com/portfolio/neo-shibuya/
• https://www.pablorochat.com/recent-work/shibuya-crossing
• https://zeroframe.art/

[mochet]

A major part of this was always intended as physical display technology, however a lot of work was left out of the original description. There have been some major developments in this space:

• https://techcrunch.com/2021/05/27/dapper-labs-backs-art-hardware-startup-infinite-objects-in-6-million-seed-raise/
• https://niftygateway.com/collections/beepletwo
• https://www.instagram.com/p/CPTkt9ajT-J/

There have also been dozens of new consumer devices coming out with embedded screens (which was the original idea with Sprocket - to combine a node with an ePaper display):

• https://rog.asus.com/laptops/rog-zephyrus/rog-zephyrus-g14-series/

• https://www.razer.com/streaming-microphones/razer-seiren-emote/RZ19-03060100-R3U1

• https://www.engadget.com/xiaomi-mi-11-ultra-price-launch-date-134516642.html

• https://roboads.com/roboads-introduces-the-worlds-first-mobile-advertising-robot-for-displaying-nft-art-and-live-cryptocurrency-pricing/

[mochet]

Metaverse Billboard: Nested, community-governed, multi-resource NFTs in action, by RMRK https://www.youtube.com/watch?v=wMpwGo6wi_0

[bedeho]

Nothing is actionable in this issue.