Background
The data team at OP Labs has been working on tracking OP distributions across governance grants, partner funding, and other sources.
We first shared early insights in Oct 2022. Takeaways then:
- Projects were deploying OP too slowly (34% of approved OP was deployed).
- Proposals approved via governance were less effective than other programs.
This doc serves to update data and case studies, and begin open-sourcing the data so others can analyze & contribute.
A snapshot of program data was taken on Mar 13, 2023
Current Deployment Status - Growth Experiments
-
We project that 56% of allocated OP (30.7M) has been deployed (not in projects’ wallets)
- We’ve observed 38 growth experiment proposals launch or complete, with 32 to be launched.
- Live & Completed programs represent 80% (43.7M) of Allocated OP (i.e. 24% of allocated OP is “live” and to-be-deployed).
# Programs # OP Allocated (M) % OP Allocated Live 🔥 Subtotal 33 41.1M 75% Governance - Season 3 - - Governance - Season 2 10 7.1M Governance - Season 1 9 4.5M Governance - Phase 0 14 29.5M Coming soon ⏳ Subtotal 32 10.9M 20% Governance - Season 3 12 2.1M Governance - Season 2 13 5.2M Governance - Season 1 4 1.3M Governance - Phase 0 3 2.2M Completed Subtotal 5 2.6M 5% Governance - Season 3 - - Governance - Season 2 1 240.0K Governance - Season 1 - - Governance - Phase 0 4 2.4M Grand Total 70 54.6M
Source: OP Summer Programs
Stats by Season
- Aggregate by Gov Fund Season
Stats were measured at the Latest Date (Note: Many programs still ongoing)
| Source | # OP Allocated | Net OP Deployed | Net $ Inflow | Net $ Inflow / OP | Incremental # Txs | Annualized # Txs / OP | Incremental Gas Fee ($) | Annualized Gas Fee / OP |
|---|---|---|---|---|---|---|---|---|
| Governance - Phase 0 | 31.0M | 16.8M | 128.1M | $7.61 | 16,775 | 0.36 | 123,940 | 2.6872 |
| Governance - Season 1 | 4.5M | 2.0M | 111.6M | $54.46 | 3,378 | 0.60 | 16,887 | 3.0086 |
| Governance - Season 2 | 2.5M | 978.0K | 16.9M | $17.23 | 1,214 | 0.45 | 5,154 | 1.9234 |
| Multiple | 8.9M | 6.9M | 227.0M | $32.69 | 10,701 | 0.56 | 33,508 | 1.7607 |
Revisited Case Studies & Early Theories
Note: We are mentioning specific programs. Some were more successful than others, but the intent is to learn from their examples, not to accuse or blame.
-
Retention Problem: Separate “Usage Acquisition” vs “Longer-Term Impact”
Theory: Incentives are great at “usage acquisition” (transaction volume, liquidity, etc), but this is not a good predictor of longer-term impact.
DEXs: Uniswap Phases 1 + 2 (selected managers), Revert Finance, Rainbow
Lending & Borrowing: Aave & WePiggy
TVL measured as “available liquidity” (deposits - borrows)
- Aave had 18% retention from the local max before incentives turned off (+$431M) to 30 days post-incentives (+$78.5M)
- WePiggy had 6% TVL retention by the same methodology (+$2.8M to $165k).
-
Value Extractive-Resistant Design: Can someone create fabricated activity to maximize rewards?
Theory: When Rewards > Costs, value-maximizing actors will spend to eat up the rewards. Anything that can be gamed, will be gamed.
Aave - Oversized Emissions Led to Recursive Borrowing
- Aave ‘Deposit APY’ + Rewards > Aave ‘Borrow APY’, so actors borrowed and re-deposited the same asset over and over to maximize rewards
- Learning: Unless ****we can design a system where Rewards < ‘Borrow APY’ - ‘Deposit APY’, lending rewards may always be gamed.
Snapshot ~1 day in to the Aave Liquidity Program
Rainbow Wallet - Swap Volume Leaderboard Led to Inorganic Volume
- Rainbow Wallet incentivized bridging to and swapping on Optimism. Base rewards were partial gas rebates, but there was an additional 52K OP bonus to the top 100 addresses by trade volume (as of Mar 5).
- On the last day, trade volume spiked to $12M, likely by addresses trying to get in the top 100.
- Trade volume fell to $25k Trade Volume / Day post-program (vs ~$4k prior), showing that the increased volume did not sustain.
This was similar to Slingshot’s Flash Programs we observed last time: Rewards were offered either per trade or per $ of volume until they ran out. Transactions spiked up following program announcements and then return to normal levels afterward.
Elsewhere: Demand-Side Incentives Have Led to NFT Wash Trading
Wash Trading: People trading NFTs back and forth with themselves to create fabricated volume.
- With LooksRare and X2Y2 introducing tokens rewards for trading, we’ve seen a significant increase in NFT wash trade volumes (58% of the NFT secondary volume was wash trading in 2022).
- Wash trade volume may disappear once the incentives become less attractive or profitable for traders (starting Sep 2022).
Source: NFT Wash Trading Dashboard (hildobby)
-
What drives long-term impact? (The real unanswered question)
Theory: We can bootstrap a network with supply incentives, but demand needs to follow, and that comes from natural product usage (need to be careful to not create fabricated demand)
Aave - Non-Recursive Borrowing had ~60% Retention
- While only 18% of Aave TVL retained, 58% of “non-recursive” borrow volume retained 30-days later (+$30.6M vs pre-incentives).
- Hypothesis: The “Non-Recursive Borrow” demand comes from other use cases on/offchain
Aave User Journey Mapping
OP Quests - ~8% of transactions come from addresses new to Optimism via Quests
- While Quests appeared to have driven a high-volume of fabricated activity, total Optimism daily transactions increased ~50% Post-Quests, and 15/18 apps saw increased transactions
- 8% of Transactions (Last 30 Days) came from addresses new to Optimism via Quests (18% of transacting addresses)
- Quests on Coinbase Wallet launched Mar 9 (requires Coinbase authentication per wallet)
Optimism Quests - App Growth on Optimism After Quests
- Quests on Coinbase Wallet launched Mar 9 (requires Coinbase authentication per wallet)
Breakdown by Program - Liquidity
Top Inflows - Acquisition Period
For Liquidity Inflows, we can segment programs by where the incentives were deployed (i.e. to the native app, to an external DEX pool).
Inflows Cutoff at Program End Date (Latest Date if still Live)
| App | Product Incentivized | Net TVL Inflows | Projected OP Deployed | Net Inflows per OP |
|---|---|---|---|---|
| Aave | App | $342.0M | 5.0M | $68 |
| Velodrome | App | $241.9M | 5.1M | $47 |
| Synthetix | DEX Pools | $120.6M | 2.4M | $50 |
| Rocket Pool | DEX Pools | $79.4M | 222.0k | $357 |
| Pooltogether | App | $56.4M | 842.5k | $67 |
| Beefy Finance | App | $32.2M | 172.1k | $187 |
| Stargate Finance | App | $27.0M | 469.7k | $57 |
| Beethoven X | App | $26.3M | 209.5k | $125 |
| Pika Protocol | App | $11.0M | 672.6k | $16 |
| Rubicon | App | $9.1M | 791.1k | $11 |
Top Inflows - Post-Incentives Period
Only Showing Programs which Have Ended
| App | Product Incentivized | Net TVL Inflows (End Date + 30) | Projected OP Deployed | Net Inflows per OP (End Date + 30) |
|---|---|---|---|---|
| Aave | App | $77.3M | 5.0M | $15 |
| Defiedge | Uniswap - Phase 2 | $2.1M | 25.0k | $85 |
| Revert Finance | App | $1.6M | 240.8k | $7 |
| Xtoken | Uniswap - Phase 1 + 2 | $1.2M | 41.7k | $28 |
| Gamma | Uniswap - Phase 1 + 2 | $372.0k | 41.7k | $9 |
| Layer2Dao | DEX Pool | $235.0k | 20.8k | $11 |
| Wepiggy | App | $166.8k | 300.0k | $1 |
Breakdown by Program - App Usage
Top Usage - Acquisition Period
For usage, we aggregate all incentive programs and observe the activity on each apps’ contracts. For a broader view, see the Project Usage Trends dashboard and project <> contract mappings.
Cutoff at Program End Date (Latest Date if still Live)
| App | # OP Allocated | OP Deployed (All Programs) | Incremental # Txs | Annualized # Txs / OP | Incremental # Txs After | Annualized # Txs / OP After |
|---|---|---|---|---|---|---|
| Velodrome | 7.0M | 5.1M | 8,045 | 0.58 | - | - |
| Uniswap | 1.0M | 150.0K | 5,666 | 13.79 | - | - |
| Pika Protocol | 900.0K | 672.6K | 4,782 | 2.59 | - | - |
| Rubicon | 900.0K | 791.1K | 4,110 | 1.9 | 1,584 | 0.73 |
| Synthetix | 9.0M | 4.9M | 4,092 | 0.3 | - | - |
| Aave | 5.0M | 4.8M | 3,123 | 0.24 | 4,744 | 0.36 |
| Hop Protocol | 1.0M | 152.6K | 3,003 | 7.18 | - | - |
| Beethoven X | 500.0K | 164.7K | 2,297 | 3.96 | - | - |
| 1inch | 300.0K | 300.0K | 2,101 | 2.56 | 390 | 0.47 |
| PoolTogether | 1.0M | 842.5K | 1,910 | 0.83 | - | - |
Top Usage - Post-Incentives Period
Cutoff at Program End Date + 30 days (Latest Date if not yet reached 30 days)
| App | # OP Allocated | OP Deployed | Incremental # Txs | Annualized # Txs / OP | Incremental # Txs After | Annualized # Txs / OP After |
|---|---|---|---|---|---|---|
| Rubicon | 900.0K | 791.1K | 4,110 | 1.9 | 1,584 | 0.73 |
| 1inch | 300.0K | 300.0K | 2,101 | 2.56 | 390 | 0.47 |
| Revert Finance | 240.0K | 240.8K | 218 | 0.33 | 247 | 0.37 |
| Aave | 5.0M | 4.8M | 3,123 | 0.24 | 4,744 | 0.36 |
| WePiggy | 300.0K | 300.0K | 39 | 0.05 | 12 | 0.01 |
| Aelin | 900.0K | 900.0K | 8 | 0 | -5 | 0 |
Top Gas Spend - Acquisition Period
Cutoff at Program End Date (Latest Date if still Live)
| App | # OP Allocated | OP Deployed | Incremental Gas Fee ($) | Annualized Gas Fee / OP | Incremental Gas Fee ($) After | Annualized Gas Fee / OP After |
|---|---|---|---|---|---|---|
| Synthetix | 9.0M | 4.9M | 91,133 | 6.76 | - | - |
| Velodrome | 7.0M | 5.1M | 32,018 | 2.31 | - | - |
| Hop Protocol | 1.0M | 152.6K | 17,055 | 40.79 | - | - |
| Uniswap | 1.0M | 150.0K | 9,268 | 22.55 | - | - |
| Beethoven X | 500.0K | 164.7K | 7,638 | 16.93 | - | - |
| Aave | 5.0M | 4.8M | 6,832 | 0.52 | 12,297 | 0.93 |
| Rubicon | 900.0K | 791.1K | 6,519 | 3.01 | 8,511 | 3.93 |
| QiDao | 750.0K | 342.9K | 5,729 | 6.10 | - | - |
| Stargate Finance | 1.0M | 469.7K | 4,774 | 3.71 | - | - |
| 1inch | 300.0K | 300.0K | 4,511 | 5.49 | -49 | -0.06 |
Top Usage - Post-Incentives Period
Cutoff at Program End Date + 30 days (Latest Date if not yet reached 30 days)
| App | # OP Allocated | OP Deployed | Incremental Gas Fee ($) | Annualized Gas Fee / OP | Incremental Gas Fee ($) After | Annualized Gas Fee / OP After |
|---|---|---|---|---|---|---|
| Rubicon | 900.0K | 791.1K | 6,519 | 3.01 | 8,511 | 3.93 |
| Aave | 5.0M | 4.8M | 6,832 | 0.52 | 12,297 | 0.93 |
| Revert Finance | 240.0K | 240.8K | 99 | 0.15 | 263 | 0.40 |
| WePiggy | 300.0K | 300.0K | 132 | 0.16 | 132 | 0.16 |
| 1inch | 300.0K | 300.0K | 4,511 | 5.49 | -49 | -0.06 |
| Aelin | 900.0K | 900.0K | 49 | 0.02 | -66 | -0.03 |
Key Takeaways
- Usage Acquisition Efficiency has improved Post-Phase 0
- Incentives have been effective at attracting usage, but not retaining it (yet).
- Game-able program designs will be gamed - how can we mitigate this?
- Open design space with how to drive longer-term impact post-incentives.
Analytics Resources
Things are still super messy, but a lot of the code and scripts powering our analysis are listed below! [Readmes & how-to-contribute writeups coming soon]
TVL Flows by Program
Flows are shown by token at the latest price (unless otherwise indicated) | Sources: Defillama & TheGraph APIs
- Time-Series Chart of TVL Flows by Program from Start to End + 30 Days
- Folder of charts specific to each program
Onchain Usage by Program
- Incentive Program Usage Summary from Start to End + 30 Days
Other Metrics & Resources
Dashboards that publicly sharable
- Optimism Popular Apps and Project Usage Trends - Dashboard
- NFT Marketplace Volume on Optimism - Dashboard
- DEX Volume on Optimism - Dashboard
- Overall Optimism Protocol Metrics - Dashboard
- OP Analytics GitHub - See Readme for more
Google Sheet Summary of Results
Token Distribution Transfer Mappings [WIP]
We can map token transfers involving known (or suspected) project addresses to determine when tokens are deployed (and to where).
- Intermediate Addresses <> Program Mapping (can help here!) | Mapping Scripts
Closing Notes
- Tracking this stuff is super difficult to do as a small group. Please help
There are also infinite more rabbit holes we could go down - We’re thinking about better metrics than raw transactions, volume, and TVL (i.e. app fees, transfer volume, incentivized vs native yield) and deeper-dive methods (i.e. segment by behavior type). Open to ideas!
- Splitting by grants and by season may get increasingly difficult over time, since protocols are re-applying for grants and using the same addresses.
- In a perfect world, every proposal uses completely distinct addresses, but may be infeasible.
- For simplicity: Thales and Overtime Markets were combined since they each used the same proposal address (we can’t easily tell the grants apart)
This post is coauthored with @MSilb7.