Predictive Signals for Network Effects

Network Effects Drive Durability in Web3 -– Paul Hsu, CEO and Founder, and Abdul Al Ali, Venture Investor, Decasonic 

At Decasonic, we underwrite conviction with a core belief that network effects drive durable and defensible growth in Web3. The best institutional investors spot them early, before consensus, before charts, before hype. They rely on fundamentals, not unsustainable spikes of narrative mindshare. Market volatility is an opportunity to double down on frameworks that forecast outsized, venture-scale returns.  And as we embark on the AI era, speed in executing growth emphasizes the importance of deliberate execution of network effects.  

Internally, we rigorously battle-test our frameworks to sharpen our investment alpha in our capital deployment and portfolio management. These debates draw on AI insights and align our myriad of questions that dig deeper and deeper into why.  Externally, we have the privilege to share some insights with our ecosystem and bring forward ideas that will move the industry forward.  We share these models to build alongside founders and attract the next generation of signal-driven investors.

Crypto analytics often chase what’s already happened - in other words, they are great at lagging indicators. TVL, revenue, or token price action are lagging indicators. Liquidity incentives wear off. Short-term spikes don’t forecast long-term strength. That’s why we frame our assessment across three timeframes:

1. Lagging: Cumulative revenue since inception. Key example: ~ $2.7M of cumulative total revenue generated by Farcaster. This figure presents a laggard indicator, and does not provide insights on future potential revenue generated by Farcaster. 

2. Present: Weekly volume over a given current period. In this example, Base has generated ~$950k in total revenue during the present week of June 2, 2025. Despite the recent revenue uptick, this is not a ‘predictive’ indicator of future revenue streams. 

3. Predictive: Compounding growth rate. This presents the acceleration in a relevant metric actively being assessed, and this is often a key indicator of predictive signals for compounding network effects. An example of a predictive signal would be the MoM growth rate in the number of Active Addresses for a chain. 

   
   

These predictive signals are leading indicators that forecast the emergence of compounding, self-reinforcing loops for network effects. The framework that we deploy to identify those signals, enables us to identify projects and tokens whose design, engagement, and growth trajectories indicate an acceleration towards a tipping point. 

That inflection when a protocol, project, or token transitions from functional to indispensable is where network effects form. And that is where venture scale, 100x, opportunities lie.

Defining Predictive Signals

A predictive signal is an early, leading indicator that a project can build network effects. These signals could often appear to be subtle: A spike in GitHub commits, a rise in unique wallets, or a pattern of repeated, self-inforcing usage of a protocol or dApp. The key metric for active, durable assessment often boils down to: compounding growth. 

As it relates to the assessment of compounding growth, consider some of the following relative metrics for assessment:

1. Strength: How meaningful is the observed behavior? 1,000% growth off 50 users ≠ 100% MoM growth on 10,000 users.

2. Duration: Is the growth persistent? Timeframe matters. Look at compounding over 3–6 months.

3. Stakeholder Breadth: Does the growth span users, developers, and token holders? The wider the base, the stronger the signal.

   
   

Defining Network Effects

Predictive signals form the groundwork for identifying emerging network effects. As investors, our role is to analyze, forecast, and take calculated risks on projects that show early signs of entering self-reinforcing loops, or flywheel effects. Network effects emerge when a product or protocol becomes increasingly valuable as more users, developers, and participants engage. These are compounding feedback mechanisms: Flywheels that accelerate usage, distribution, and long-term value. 

The more engaged the ecosystem, the deeper the moat, and the more defensible it becomes. At Decasonic, we underwrite around three core investment parameters across both venture and liquid token strategies: Narrative, Product-Market Fit (PMF), and Execution. Each parameter contains distinct predictive signals that can reveal the early onset of durable network effects. We break down examples of each below. For each of the examples we highlight below, there are always exceptions: 

1. Narrative: A strong example of a self-reinforcing narrative feedback loop is found in platforms like PolyMarket and Bonk. Here, users, key ambassadors, are incentivized to create prediction markets or launch tokens for a shot at significant valuation or impact. Recent 24-hour data shows PolyMarket with thousands of active markets and Bonk with 25,000+ tokens launched, of which ~100 hit key milestones. Despite a 0.4% success rate, users keep participating, aiming for high returns. The loop runs: Users launch markets/tokens -> some hit milestones -> success drives more launches. This user-fueled, community-driven cycle thrives on viral momentum.

2. PMF: Take a DeFi protocol. The higher the TVL achieved and the higher the fee-efficiency (TVL/Fees), the more users benefit from using that protocol. In this example -> deposited TVL -> higher swaps -> higher liquidity -> greater deposited TVL. This is a self-inforcing feedback loop, as the TVL will grow if the platform notes higher usage over-time. 

3. Execution: Hyperliquid exemplifies stellar execution in crypto, amplified by powerful network effects. It was first to list $TRUMP perps, doing so in ~2 hours post-launch, driving trader influx. This fueled Hyperliquid’s ~70% perps market share. The loop is: First to list tokens on perps -> more traders -> more listings -> higher market share. Network effects, like traders boosting liquidity, strengthen this cycle.

   
   

Predictive Signals -> Network Effects 

Network effects are defined by reinforcement and can evolve linearly, hit a sudden tipping point and then explode exponentially. I saw this at my prior company, Zynga, during the early days of social gaming on MySpace and Facebook.  That’s why predictive signals matter: They emerge before that reinforcement becomes obvious. 

Projects do not immediately launch and have sustained, durable network effects. Pump launched on January 19, 2024, with nearly $0 in revenue until March, 2024. The chart below highlights the daily revenue generated by pump from token launches. This is a pattern observed with some of the most successful project launches, with pump having one of the highest cumulative revenue generated from any crypto project, a lifetime revenue of $700+M. 

Source: Link.

Below, we provide examples based on our core underwriting parameters associated with predictive signals that could lead to sustained, durable network effects observed by projects, ecosystems, and tokens. 

I. Narrative Signals

Narrative is more than virality, its belief distribution. It creates communities, and drives them to rally usage around a product. The core assumption we make in identifying predictive signals based on Narratives is that the mindshare of a project’s growth will continue to compound over time, and it will not be a brief spike.

1. Mindshare Sustainability and Growth: An interesting example is Virtual ($VIRTUALS), which has constantly sustained its leading position amongst the AI Agent token projects on Cookie. Zooming out, and assessing both the rate of growth in mindshare, and the sustainability of a dominant position helps lead to signals for network effects in a Narrative underwriting.  

   
   

Source: Link

II. PMF Signals

These are signs of increasing, repeated usage of a project, which compounds overtime to more usage. 

1. DAU/MAU compounding growth, not spikes. The key is to assess the growth in usage based on a pre-defined time period. You could look at the growth of a project over a 3-6 month period, in order to estimate future growth.

   
   

III. Execution Signals

Execution signals are visible when roadmap delivery results in downstream behavior:

1. Ecosystem integrations post-launch. Taking a look at the growth of projects building on an L1. An example of this is Soneium, the joint venture network between Startale and Sony. When we published the Soneium market map back in October, 2024, we identified 140+ projects aiming to launch on Soneium’s main net. As of the most recent figures from Soneium, their market map indicates 210+ projects building on Soneium post main-net launch.

2. This ecosystem integration extends to Sui. Initially when tracking the projects building AI on Sui, we observed a continued growth leading up to our market map that was driven by the delivery of Mysten Labs on their core promises for Sui-related upgrades. This has increased the number of Sui AI projects to 50+ at the time of writing this blog. 

3. A similar trend is observed with Bittensor and the underlying subnet ecosystem. At the time of the dTAO upgrade on Feb 13, 2025, the number of Bittensor subnets were 65. A few months after the dTAO upgrade and in part due to the Bittensor Foundation’s commitment to solving the incentive-related emission structure, the number of subnets has nearly doubled to 120. We will be providing a comprehensive Bittensor ecosystem market map in our next Web3 x AI Series. 

Source: Link

Sui AI

Source: Link

The signal is not the shipping speed of the project, nor is it deviation from achieving a target roadmap or development phase, it’s whether the milestones achieved lead to sustained, durable usage. 

From Signal to Flywheel: Network Effect Maturity

Not all signals lead to network effects. The key to predictive signals and identifying network effects, is defensibility. 

For a signal to evolve into a defensible loop, it must:

1. Persist across time, be durable

2. Engage multiple stakeholders, provide opportunities for growth and value accrual for all identified key stakeholders of a project

3. Feed into future activity, milestones that lead to continued adoption of a project

We have written about these flywheel effects across other blog posts, most recently in this blog post about SocialFI and Network Effects.  

We use a three-stage maturity model to assess this evolution:

1. Signal Onset: A single signal emerges (e.g., wallet count increase)

2. Signal Reinforcement: Multiple signals overlap (e.g., wallet count + narrative)

3. Loop Formation: Feedback begins user behavior drives more of itself

The transition from signal to flywheel is where most investors hesitate. We lean in here, especially when network effects begin to reinforce across Narrative, PMF and Execution simultaneously.

Conclusion

Predictive signals offer a path to identifying network effects before they appear in the data rooms, dashboards, or token charts. By measuring not just what’s happening, but how fast and how broadly it’s growing, we generate an edge in both venture and liquid markets. The key is to identify predictive signals prior to sustained, compounding growth in network effects. 

Key takeaway: Network effects are the outcome. Predictive signals are the roadmap.

The content of this material is strictly for informational and educational purposes only. It is not intended to constitute investment advice, nor should it be considered a recommendation or a solicitation to buy, sell, or hold any asset. Decasonic does not endorse investments in any specific tokens, and nothing in these blog posts should be construed as legal, tax, or financial advice. Please consult with a qualified professional advisor before making any financial decisions. Decasonic provides no warranties, whether expressed or implied, on the content provided in these blog posts, including its accuracy, completeness, or correctness. The opinions expressed here are those of the authors and do not necessarily reflect the views of Decasonic. Please note that Decasonic may hold a position in some of the tokens mentioned, including Virtuals. Decasonic is not liable for any errors or omissions in the content of this material or for any actions taken based on the information provided herein.