Towards Human Distilled Judgement for User Adoption Funding1
By Eval.Science
Figure 1: User Adoption Dependency Graph
Abstract
This framework applies Deep Funding principles to user adoption, using machine learning models to allocate resources across a dependency graph of interconnected ecosystem components. By defining user acquisition as consistent monthly blockchain interaction over a six-month period, the model identifies and weights critical infrastructure elements—educational resources (30%), wallets (35%), on/off ramps (25%), and communities (10%)—that contribute to sustainable adoption based on a jury. The adaptive funding mechanism continuously refines allocation based on real-world adoption metrics building on top of metrics gardens framework.Table of Contents
1. Goal Definition
User acquisition is defined as a wallet address that has interacted with the Celo blockchain at least once every month for the last 6 months. The Deep Funding initiative aims to increase sustainable user adoption by identifying and funding the most impactful components of the ecosystem.
2. Graph
The dependency graph below illustrates the interconnected components that contribute to Celo user adoption. Deep Funding allocates resources based on machine learning models that predict the impact weight of each node.
User Adoption Dependency Graph
A wallet address that interacts with Celo at least once monthly for 6+ consecutive months
- Wiki - 8%: Celo Documentation Hub with ultra-light client explanations
- Tutorials - 12%: Interactive guides for cUSD transfers and governance
- Videos - 10%: "Celo Thursdays" series highlighting use cases
- UX - 20%: Valora wallet with recurring payment features
- Docs - 5%: Multilingual support documentation
- Infra - 10%: Light client implementations for low-end smartphones
- Liquidity - 15%: Integration with local mobile money providers
- Infra - 10%: USDT-based transaction capability for feature phones
- Events - 7%: Monthly "Impact Maker" workshops
- Resources - 3%: Impact Hub showcasing community projects
The percentages represent the predicted contribution of each node to overall user adoption, as determined by Deep Funding's machine learning models.
3. Metrics & Features by Node
Impact metrics that should be used by the Deep Funding ML models to evaluate and allocate resources across nodes:
Educational Resources
- Open Rate & Views: Engagement metrics for educational content
- Conversion Rates: Percentage of content consumers who become active Celo users
- Topic Activity Correlation: Increase in specific topic activity correlated with on-chain interaction
- Bootcamp Graduates: Number of users completing educational programs who maintain wallet activity
Wallets
- Wallet Growth: Increase in number of wallets over time
- Retention: Percentage of wallets remaining active month-over-month
- Documentation Metrics:
- Cross-chain user stats
- Number/percentage of transactions initiated through wallet
- Wallet stickiness (return rate)
- Feature adoption metrics
On/Off Ramps
- Total Funding Bridge: Volume of funds entering ecosystem
- Liquidity Metrics:
- Transaction count into the ecosystem
- Liquidity pool size and availability
- Diversity of pools (number of pools by type)
Communities
- Event Conversion: Number of attendees transacting after events
- Engagement Depth: Sustained activity following community touchpoints
ML Model Integration
The Deep Funding algorithm processes these metrics to:
- Identify which features most strongly correlate with sustained 6-month user adoption
- Adjust funding weights dynamically as new data becomes available
- Optimize for regional variations in adoption patterns and user needs
- Track the cascading effects of improvements in one node on dependent nodes
4. Data
The Deep Funding model ingests the following data sources to continuously refine its funding weight predictions:
- On-chain transaction data tracking monthly user interactions
- Wallet installation and retention metrics
- Educational resource engagement analytics
- Community event participation and subsequent transaction patterns
- Fiat-to-crypto conversion volumes across different on-ramp solutions
- Regional adoption trends correlated with local infrastructure developments
- Social impact metrics aligned with Celo's mission
An initial dataset will be created to train ML models here. The dataset will be made available to the public.
5. Governance
This framework and graph is available and is governed by the community on Github.
Community Weight Allocation
Allocate importance weights to each component (total must equal 100%)
Educational Resources
Wallets
On/Off Ramps
Communities
Your vote will be recorded on-chain and contribute to the next funding cycle.
- Transparent ML Model: Open-source machine learning algorithms with documented training data.
- Community Review: Quarterly reviews of funding allocations with community input.
- Impact Verification: Independent verification of reported user adoption metrics.
- Adaptive Thresholds: Dynamic adjustment of what constitutes meaningful blockchain interaction.
- Regional Balancing: Governance mechanisms to ensure global representation in funding allocation.