Proposal: Community-Enabled Analytics for Sushi

[flipdave]

At the request of @dangerousfood & @omakasebar we are presenting Community-Enabled Analytics (CEA) to the Sushi community.

CEA will develop an educated and passionate community of nearly 500 individuals that attacks analytic challenges in real time.

The program will drive on-demand data needs - as well as customer acquisition, ecosystem growth and token circulation.

Below is the proposal, what do y’all think?

Three questions in particular we are really interested in the community’s thoughts on:

1. You all are the experts - what are the most effective ways to find the questions that the community is most interested in addressing via analytics?
2. Any processes or forums that are in place already that you would suggest implementing/engaging with to keep a live pipeline flowing for the questions the community is most interested in?
3. Any nominations for who should be on the Oversight Committee?

Otherwise, any thoughts, criticisms, or builds are all very welcome - let’s make an awesome program together!

What is Community Enabled Analytics?

Community-Enabled Analytics (CEA) is a free analytics solution combined with an extensive bounty program, designed to reward community members for learning about, analyzing, and publishing blockchain data.

Notable features:

1. SQL query-able pre-modeled and labeled data with visualization studio - free to all end users.
2. API ready - each query automatically generates a configurable, read-only API endpoint that allows data to be piped to any supporting service (including Google Sheets & Research, Tableau, Python, R, and others)
3. All data is downloadable via CSV.
4. Multi-chain support including Polygon & Terra with Arbitrum, Optimism, and other EVM compatible chains coming soon.
5. Community Requested Queries - a format for the community to ask for specific queries to be developed and produced on-demand.
6. Community members receive bounty rewards in the native token to utilize the product, conduct complex analyses, and build compelling tools to support SushiSwap and the broader crypto space.

CEA Program Outputs

The power of CEA is the variety and depth of analysis that can be performed. From research reports, to custom dashboards (both built with API or visualization studio), to custom interactive tools like LP calculators, CEA analysts can deliver on whatever analytics the ecosystem needs most.

Some examples:

• Custom Interactive Dashboard for AAVE: As part of the grant for CEA, Aave was looking for an interactive dashboard comparing liquidity and other key metrics across v1, v2, and eventually Polygon (being adding shortly) – a CEA community analyst built this custom dashboard as a result.

• In-depth Research For Terra: At the request of Terra’s founder, our community analysts produced robust research on Airdrop campaigns across the ecosystem. 3 separate elite analyst teams tackled Terra’s top protocols respectively (Anchor, Mirror, Pylon plus the whole ecosystem) to assess Airdrop efficacy.

• LP Calculator for Uniswap: Despite V3s dominance, users were having considerable difficulty understanding how to deploy the best strategies on V3 - so as part of our UGP grant the V3 calculator was developed. In the first 3 weeks there were 1100 clicks to Uniswap, 8.9k unique users,13.8k sessions.

Ecosystem Acquisition & Retention

Flipside’s value proposition begins with analytics, but extends into community member acquisition and drives token circulation and retention:

• Acquisition: Utilizing a series of educational programs (see Terra scavenger hunt, SQL training solutions, AAVE Cheat Codes), and a hands-on community and analytics team, we activate new participants into the ecosystem. This teaches new users how to actually use a complex protocol, how to think about the underlying mechanics, and even how to start becoming adept at using SQL to build queries and public-facing insights. Across previous programs, nearly 69% of bounty recipients had never held protocol tokens before.

• Retention: This guiding metric is the percent of distributed tokens that are not sold or swapped after they are received. This reflects a traditional retention metric: the community member has chosen to remain involved in and supportive of the ecosystem. Across programs for other customers, 77% of bountied tokens remained within the ecosystem.

Grant Request: 1M USD Annually in $SUSHI

To deliver the full solution, we’ve discussed a simple annual grant with @dangerousfood and @omakasebar to fund the program. This grant may be renewed on an annual basis.

Fee Breakdown:

→ 50% of this will be delivered to ecosystem community members as bounties ($41,667 per month):

All bounties are delivered through weekly programs: Flipside sources questions from Sushi and the community, prepares a bounty construct, releases to the community, QAs all results, and distributes bounties to the most effective outcomes. This is a complex and very time consuming process, but the return is significant: extremely directed results that serve core team and community needs with immediacy. Bounties range from $40 to $650 each, determined by the complexity of the request.

→ The other 50% ($41,667 per month) is the subsidization of operational and infrastructure costs.

A key provision is to ensure data remains free and accessible to Community Participants at all times.

This subsidy covers all aspects including pipelines of transformed analytics-ready data, an open API, and the structuring, management and quality control of the bounty program and analytical outcomes.

Additionally, the subsidy addresses infrastructure costs inclusive of node management (ETH), compute / storage needs (AWS), data warehousing (Snowflake) and a variety of other software / architecture tooling requirements to sustain and make Flipside available to analysts of all levels.

Year 1 Program Expectations:

Given previous programs, the expected results from a year of bounties would be as follows:

~ A community of 460 educated analysts to produce on-demand insights for Sushi

~ 27,000 $SUSHI retained in ecosystem

~ 350+ unique dashboards, queries, visualizations, & in-depth analyses, plus premium interactive solutions including a Trident concentrated liquidity calculator.

Sushi Program Oversight

It’s critical that assets are managed effectively and are delivering value to the ecosystem.

Two components provide effective checks and balances to ensure maximum accountability to the Sushi ecosystem:

1. Oversight Committee: whose purpose is to ensure that the grant proceeds are applied in a manner consistent with its design. The Committee will meet quarterly to review and ensure accountability, and take a vote annually to evaluate the continuation of the program. The Oversight Committee will be composed of 5 members chosen by the SushiSwap community. A Flipside representative will be present at the meetings to provide information and context when/if the committee asks for it, but will not have a vote.

2. Transparency Dashboard: Results of the program will be broadcast live via a custom dashboard for the Sushi community.

Summary: Community-Enabled Analytics for Sushi

With the requirement for analytics to grow rapidly with Trident, CEA will engage the Sushi community, by both sourcing data needs as well as offering extensive opportunities for analysts to earn $SUSHI for their efforts.

Additionally, CEA will bring new users to the protocol, motivating analysts to provide protocol insights, and drive token retention by distributing $SUSHI to highly engaged users.

• Dedicated online portal for community-sourced analytics requests, which will be turned into bounties.
• Direct opportunities for Sushi community analysts to produce and earn.
• Free historical state-ready, transformed SushiSwap data for SQL analytics - as well as visualization and API solutions.
• Educational content for analysts of all levels to activate within the SushiSwap ecosystem. Beginners can learn SQL, advanced users can optimize results to deliver sophisticated solutions, including development of a SushiSwap fee calculator
• Analytics requirements also sourced directly from Sushi Protocol Members.
• Structured bounty programs activating analysts to produce outcomes. Includes quality control of results to ensure the data is effective and correct.
• Bounty rewards (in $SUSHI) delivered into the hands of community members, activating new members and retaining participation.

[orionmuir]

I love this. There is so much rich data to be collected, organized and disseminated. As a former professional poker player I saw data collection through SQL in tracking softwares revolutionize the game. As a contributor to sushi in the form of co-hosting the community call, UX testing and even the mere act of voting I have felt the effects participation has on brand loyalty. I see this more as a result of time investment and creative process attachment rather than the monetary award, but money helps get more top quality candidates in our ecosystem.

I believe that this could benefit our marketing on various levels. From the data collected to the stories of people finding their way into the industry to being a force of shaping the future of how DeFi looks (as we have continued to be), this has power.

I can’t wait to see uniswap forking some of the advanced data analytics tools we produce. Maybe we should patent them

[massnomis]

Hi all! First of all, huge fan of Sushi, especially its community. I would love to delve deeper into sushi on an analytics side, especially after focusing a lot on Uniswap and money markets. Sushi is all inclusive in its market functions and bringing over a community of dedicated community led analysts is perfect for Sushi.

[Neal]

Great proposal @flipdave, thanks for taking the time to spell it out so clearly.

Fully supportive of this.

[3.1415-r]

@flipdave

→ The other 50% ($41,667 per month) is the subsidization of operational and infrastructure costs.

Hi @flipdave ,

Cost aside, I am very supportive of the proposal. Better analytics will definitely be valuable for everybody.

That being said, the monthly cost of $41,667 for operational and infrastructure expenditure seems utterly high to me. Could you please provide a detailed breakdown to help us understand how you came up with this estimate?

[jacobkaka344]

我支持这个

Summary

This text will be hidden

[flipdave]

Thanks @3.1415-r

We appreciate the ask on the costs, and are happy to share details.

In order to support a recurring on-demand analytics & bounty program, we have a number of required internal resources. These include a Community Manager, Content Writer, Marketing Manager, Business Analyst, Financial Controller, Data Engineer, Data Analyst, Data Scientist and Product Manager.

All of these individuals are not allocating 100% of their time to this initiative, but even just ascribing a small portion of their time (10-15%) across 9 blockchain salaried individuals is not insignificant.

We also have ongoing infrastructure costs, requirements to sustain and make Flipside solutions freely available to analysts of all levels, inclusive of node management (ETH), compute / storage needs (AWS), data warehousing (Snowflake) needs and a variety of other software / architecture tooling, that will only increase with each new protocol (and wave of analysts) that we onboard.

As we’ve run programs across a number of blockchains (ie: Terra, Flow, Algorand) and a variety of protocols (ie: Alchemix, Compound, AAVE) we have focused on identifying and refining the funding structure as much as possible. This program would run at a low margin of between 5-14% depending on scale of submissions, etc.

While we are fortunate to have venture capital supporting our business, it remains critically important that we cover some of the costs we incur in order to maintain an open software platform for the communities we service.

[3.1415-r]

Thanks for the extra colour @flipdave, this is helpful. Appreciate that you are a privately owned company, but given you are dealing with DAO clients and the name of the game here is transparency, one thing that would make me more comfortable is you to commit to periodically share your consolidated financial statements with us. It would allow us to get comfortable that you, as a business partner, are charging overall reasonable prices compare to your costs. As a VC funded business and as you kind of hinted to, you are probably operating at a loss, so no issue with you, but if we were to receive a pricing proposal from a business that operates at 40%+ EBITDA margin for example, then we could objectively argue that they are trying to extract a disproportionate amount of value from our protocol.

I know this is definitely not how the industry has played historically, but this would be a very welcome change at the web3 era, levelling the transparency playing field a bit, and hopefully setting a standard for traditional private companies looking to work with DAOs going forward (I am going to start requesting that in similar situation).

Hope you will be open to this idea.

[flipdave]

Love this idea @3.1415-r and are happy to be transparent with program specifics. We’d want to be considerate of individual employee salaries and other private data, but otherwise support clarity amongst all parties.

As part of the transparency dashboard we will deliver with the program, we’ll include details on the operating budgets and particulars related to margin and results.

How does that sound?

[3.1415-r]

Glad to see you are on-board!
A dashboard specific to the project would be nice too, but I was taking about you sharing your company level consolidated financials (same as what you share with your investors on a quarterly basis, and the audited year-end accounts, so we can see that it ties to the reporting).
The benefit of this approach versus a dashboard specific to Sushi is that: 1) we will have some comfort that the numbers are real thanks to the auditors stamp, whereas you could put whatever you want in the dashboard in term of budgeting, we have no way to check if this is correct; 2) No issue with individual employee salaries; and 3) This does not require any extra work for you, just sharing existing documents.

By the way, that’s a bit off topic, but given you are a “blockchain” company, why not ditching the legacy corporate structure and become an on-chain organisation? You would not need to give away the control quite yet (though transitioning toward a DAO in the long run would be nice), but at least you will have all your finance on-chain with some nice real-time reporting like Yearn or Aave for example; that way your costs will be transparent and it will become significantly easier to justify your pricing. I would personally favour dealing with a transparent on-chain organisation a million time over a private corporation, and I am sure this sentiment will be shared broadly among DAO participants.

[flipdave]

Aha!

Don’t want to give away all the details just yet, but let’s just say we’re all singing from the same songbook. It’s a bit beyond the scope of this proposal today; would be great to get started, create positive valuable outcomes, and like minds will eventually align there!

[3.1415-r]

Great, looking forward for more news on that front!

Regarding your current proposal, it would be great the get the view of other community members.

[Unagi]

Thanks for submitting this proposal! Having something similar to the v3 LP calculator would be awesome for Trident users!

I’d love to learn more about your experience in running past bounty programs. Could you share some more numbers on previous campaigns. Eg:
How many questions where submitted per project per month?
How many submissions got paid? And how much per submissions?

Feel free to explain with the Uniswap example, and go over their respective submission and winner numbers.
What’s the difference between Elite questions vs other questions, and Grand prizes vs other prizes?

Also on the process side, how do you usually gather questions? Where would you start, beyond this post? What help would you need from the Sushi team?
From your previous experience, do more questions come from your community or the project’s community?
How do you curate them and assess the quality of the questions?

How do you handle change in number and quality of questions over time? Eg. if the monthly fee pool is fixed, while number of questions goes down vs the previous month, will you increase bounty reward?

To get an idea of specific outcomes:

Super pragmatic: How many Uniswap v3 LP dashboards would your community be able to build with this proposed plan per month?
Or would the outcome of this proposal be the SQL queries only, and the Sushi community would need to build a respective web app on top, if we wanted a LP dashboard?

A key provision is to ensure data remains free and accessible to Community Participants at all times

Besides the questions and queries, what will be Open Source or readable by the Sushi team?
Eg. is the frontend and backend for the v3 LP dashboard Open Source?

The subsidy addresses infrastructure costs …

Do you mind sharing infrastructure costs of previous programs?

On Acquisition:
Did you see Sushi Samurai’s similar plans?

On Retention:
Did you track other metrics besides token hodling by bounty receivers?
Eg. might be interesting to see if data analysts stay active in the Sushi Discord.

[flipdave]

Hi @Unagi, thanks for the questions.

Excited to have you digging in a bit and thinking about how these programs are constructed. We’ve addressed your questions below. The input and questions are valuable for helping define the right outcomes, so please let us know if you require further clarity anywhere.

How many questions where submitted per project per month?

Submissions per project are a direct outcome of the size of grant. With an ultimate goal to educate, activate and retain as many community members as possible, it’s important to generate multiple submissions for each bounty [which, of course, enables best analytics outcomes].

Submissions do vary greatly based on factors such as question difficulty, total # of questions requested, etc. There’s a lot of art here vs pure science.

For the size of program we are discussing here, we might expect an average of 400-450 per month.

How many submissions got paid? And how much per submissions?

Bounties are distributed to the top results for any submission set. There are a variety of models we run > sometimes only 1 bounty recipient. Sometimes multiple (if there are multiple great efforts) and often we’ll run ‘collaborative’ requests so multiple individuals can receive a bounty on one question.

The payment per submission varies but ranges from $50 - many thousands. Depends on the difficulty of the task; we have a number of models that optimizes the bounty request to directed outcomes.

Feel free to explain with the Uniswap example, and go over their respective submission and winner numbers.

The initial UNI program was a pilot grant with a total of $25k in bounties from the UGP. That’s what produced the UNI calculator; it also delivered 140 unique bounty submissions (queries; dashboards) covering over 30 different protocol-specific topics and attracted 34 engaged, active analysts (of which 70% were new to the Uniswap ecosystem). Note that Flipside self funded much of this to prove the model at a larger scale.

To understand a scale program, it may be simplest to describe via our case with Terra. You can read that here.

What’s the difference between Elite questions vs other questions, and Grand prizes vs other prizes?

Elite questions are harder prompts whose outputs are intended to be powerful, memorable tools and analyses. We keep them open for two weeks, pay double the bounty of Hard questions, and typically ask community members to go above and beyond in their answers, so that the results become high-quality pieces with “staying power” for the ecosystem.

One recent Elite question produced this [for AAVE].

Grand prizes are awarded to the best submission for each question. They’re on the order of 5x-8x the baseline payment for the question. The goal of offering a winner’s prize is to encourage motivated members to put in the extra work, i.e. to create something amazing or dive deep into the analysis.

We do our best to publicize grand prize winning submissions, in order to benefit both the makers (raise their profiles and get them some attention for their hard work) and the community (who gets to employ the results in their own decision making). We’d be happy to provide a full list of the grand prize winners for each bounty program.

In the meantime - two quick examples. Here’s one of the grand prize winners for Uniswap: an assessment of impermanent loss for some of the most prominent Uniswap pools.

Another grand prize winner for one of our AAVE questions produced this metrics dashboard.

Also on the process side, how do you usually gather questions? Where would you start, beyond this post? What help would you need from the Sushi team?

We perform all of this work, focused on engaging the community. We will provide an open submission form for the Sushi community where anyone can recommend a question for bountying. Moreover, our team maintains an open notion board that presents all of the completed and open bounties.

We also activate within Discord and other communities to gather community needs to bounty, and of course anyone on the protocol team may ask for something specific.

From your previous experience, do more questions come from your community or the project’s community?

Each protocol/blockchain is different. We tend to optimize for the community; the more the community engages the better.

How do you curate them and assess the quality of the questions?

We have a team of nearly a dozen data scientists and analysts who (as one part of their role) evaluate questions; we often lean on community members themselves to validate complexity.

How do you handle change in number and quality of questions over time? Eg. if the monthly fee pool is fixed, while number of questions goes down vs the previous month, will you increase bounty reward?

Sometimes. We model out what is the likelihood of getting a quality response, and will adapt the bounty fees and number of questions to optimize. We also factor in variables such as how many bounties the community has participated in on a total and individual basis, as that impacts efforts.

Super pragmatic: How many Uniswap v3 LP dashboards would your community be able to build with this proposed plan per month?

The Uniswap v3 LP dashboard [we call these ‘calculators’] are highly complex and require multiple community members aligned with internal data scientists at Flipside. They also require significant bounty amounts to deliver.

For this size of program we would deliver at least one calculator for Sushiswap, but likely several depending on the community’s needs.

One way to think about this would be to consider program intent. While analytics are the cornerstone, the metrics we focus on our community growth, education and retention. We onboard new members into the ecosystem, using analysis as a solution for learning > and then motivate them to maintain assets [ie: 97% of LUNA remains in the Terra ecosystem].

While the short term benefit is the calculator development, the long term benefit is a trained, optimized group of analytical minds who are empowered to continue to deliver value for Sushi.

To that end: additional calculators are solely a mechanic of demand and how much bounty is available. If the community determined it would be most important to develop only calculators, then there would be less onboarding of new users, that’s the tradeoff (ie: one big project vs. many opportunities for users of all levels to get involved in the ecosystem).

Or would the outcome of this proposal be the SQL queries only, and the Sushi community would need to build a respective web app on top, if we wanted a LP dashboard?

We provide access to the SQL layer, as well as a full suite of visualization solutions - and every SQL query generates an API endpoint that can produce outcomes in any format (ie: Tableau, Notion, Medium, etc.)

Besides the questions and queries, what will be Open Source or readable by the Sushi team?

Eg. is the frontend and backend for the v3 LP dashboard Open Source?

We make a few things open source, and will continue to build more in public as we go.

Currently you can view the SQL for models producing the views in our public data layer here:

And the code that produces our UniswapV3 calculator here

The subsidy addresses infrastructure costs …

Do you mind sharing infrastructure costs of previous programs?

As noted in one of the responses above, we are happy to provide infrastructure costs for this program in real time as part of the transparency dashboard; each program does has a variety of different requirements and costs to deploy [ie: some protocol/chain data is more complex to manage than others] so the calculations will be different.

Did you track other metrics besides token hodling by bounty receivers? Eg. might be interesting to see if data analysts stay active in the Sushi Discord.

We track many metrics of wallet behavior; if the community seeks this type of insight, we can certainly bounty that out. Discord keeps their data closed to the owner of the community; so, we can identify data for Flipside’s Discord, but not for others, unless it is shared with us.

We look forward to continued dialogue - and, of course, to getting started at enabling clear growth of the ecosystem, which will benefit all.

[Unagi]

For this size of program we would deliver at least one calculator for Sushiswap, but likely several depending on the community’s needs.

While the short term benefit is the calculator development, the long term benefit is a trained, optimized group of analytical minds who are empowered to continue to deliver value for Sushi.

Thanks a lot for the detailed answers and the additional examples! I feel like this paints a good picture of what the Sushi community can expect to get out of your proposal:
Incentivizing smart minds to join the community, and supporting existing Sushi users with specific queries and dashboards.

[avimeyers]

Hi all, Snapshot is now live for voting!

[skyhawk13]

Interested to see how the vote plays out