| Parameter | More is less | More is More | | ------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | DR (Market Stability) | Duo may prefer a more stable market to reduce coordination costs and disagreements between founders. A stable market allows for more predictable decision making. | Duo may be more willing to tackle dynamic markets because they can divide and conquer, with each founder focusing on different changing aspects. Duo has more total cognitive bandwidth to process a dynamic environment.<br><br> | | SR (Market Size) | Duo may prefer to target a smaller market initially to prove their business model with fewer resources. Starting small also reduces the risk of conflict between founders. | Duo may target larger markets from the start because they have more combined resources, connections and credibility than a solo founder. Larger markets provide more opportunities for the founders to pursue different customer segments if needed. | | BR (Optimism) | Duo may be less optimistic than a very confident solo founder because the duo has to constantly calibrate their expectations with each other, reducing individual over-optimism. | Duo may be more optimistic than a timid solo founder because the founders can reinforce each other's confidence.<br><br>Having a shared vision can increase the optimism beyond what either founder would have individually. It's unclear if duo would be consistently more or less optimistic than solo. It likely depends on the individual traits of the founders and their interpersonal dynamics. | | OB (Market Acceptance info) | Duo may have less information on market acceptance because each founder has limited bandwidth to gather external data while also coordinating internally. Some market information could get lost in the "telephone game" between founders. | Duo likely has more total information on market acceptance because they can tap into two separate networks and have two pairs of eyes and ears open to the market. They can cover more ground in customer discovery. | | ER (Experiment Opportunity) | Duo may have fewer experiment opportunities because some experiments could be vetoed by the other founder. With shared resources, each founder may be more conservative in deploying experiments. | Duo can run more experiments in parallel by dividing and conquering. They also have a larger combined pool of resources and mental bandwidth to deploy a greater number and variety of experiments. | | CT (Market-Product Confidence Ratio) | Duo may have a lower ratio (relatively more confidence in product than market) because the founders reinforce each other's confidence in their own product. Overconfidence in one's own product is a common pitfall. | Duo may have a higher ratio (relatively more confidence in market than product) because they can do more market research and the Product Confidence gets divided across two founders' visions. Reaching agreement provides a check against Product overconfidence. | ## 🧬⛓️⚙️ evol.ent | Category | Actionable Item | Description | Key Components | Citation in Introduction | Citation in Literature Review | |----------|-----------------|-------------|-----------------|---------------------------|-------------------------------| | Theory | 🧠 Develop a multi-level framework of entrepreneurial experimentation | Create a comprehensive model that integrates the hierarchical levels of agency with types of experiments and entrepreneurial lenses | - Five levels of agency (operational to perspective control)<br>- Three types of experiments (exploratory, move testing, hypothesis testing)<br>- Four entrepreneurial lenses (engineer, scientist, artist, designer)<br>- Illustrate how these elements interact in the experimentation process | "Building on Piepenbrock et al.'s (2009) insights on enterprise architectures and competitive dynamics, we propose a multi-level framework for entrepreneurial experimentation." | "Piepenbrock et al. (2009) highlight the importance of understanding firm-environment interactions, which we extend to the entrepreneurial context through our multi-level experimentation framework." | | Theory | 🔄 Propose a dynamic, reflective process model of entrepreneurial iteration | Design a process model that captures the iterative nature of problem understanding and solution development in entrepreneurship | - Emphasize the role of reflection at each stage<br>- Incorporate both qualitative and quantitative approaches<br>- Demonstrate how entrepreneurs move between different levels of agency<br>- Show how problem framing evolves through experimentation | "Drawing inspiration from Fine et al.'s (2022) work on operations for entrepreneurs, we develop a dynamic, reflective process model of entrepreneurial iteration." | "Fine et al. (2022) emphasize the importance of scaling tools in entrepreneurship. We extend this concept by proposing a reflective process model that captures the iterative nature of entrepreneurial experimentation." | | Empirics | 🔬 Conduct empirical studies to validate and refine the model | Carry out qualitative and quantitative research to test and improve the proposed framework and process model | - Case studies of entrepreneurs using different types of experiments<br>- Longitudinal studies tracking how entrepreneurs iterate through levels of agency<br>- Surveys or experiments testing the effectiveness of different entrepreneurial lenses<br>- Analysis of how reflection impacts entrepreneurial outcomes | "We empirically test our model, addressing the call for more rigorous empirical work in entrepreneurship research highlighted by Piepenbrock et al. (2009) and Fine et al. (2022)." | "Our empirical approach builds on the methodological insights of Piepenbrock et al. (2009) and Fine et al. (2022), combining qualitative case studies with quantitative analyses to provide a comprehensive test of our entrepreneurial experimentation model." | ## 🗺️🧭 bayes.ent | Category | Actionable Item | Description | Expected Outcome | Citation in Introduction | Citation in Literature Review | Cost | Information Value | |----------|-----------------|-------------|-------------------|--------------------------|-------------------------------|------|-------------------| | Theory | 🧠 Mind Meld: Unified Framework | Integrate concepts from statistics, computer science, and entrepreneurship into a coherent theoretical model | A novel, interdisciplinary framework that extends current entrepreneurship theory | "Building on Camuffo et al.'s (2020) scientific approach to entrepreneurial decision-making, we propose an integrated framework that incorporates advanced analytical methods." | "Our work extends Zellweger and Zenger's (2021) pragmatist approach by synthesizing it with statistical learning theory in the entrepreneurial context." | 💰💰 | 🔮🔮🔮 | | Theory | 🛠️ Tool Time: Practical Guidelines | Develop detailed guidelines and examples for implementing the theoretical framework in real-world scenarios | Increased practical relevance and actionability of the research, bridging theory and practice | "We address the gap identified by Camuffo et al. (2023) by providing concrete guidelines for implementing scientific decision-making in diverse entrepreneurial contexts." | "Our practical implementation guidelines build upon the scientific intensity measure developed by Camuffo et al. (2020), offering a more granular approach to entrepreneurial experimentation." | 💰 | 🔮🔮 | | Empirics | 🌍 Global Lab: Multi-Site RCT | Design and execute a large-scale, multi-site randomized control trial to test the proposed framework | Comprehensive empirical evidence supporting (or refuting) the proposed applications across diverse settings | "We empirically test our framework using a multi-site randomized control trial, expanding on the methodological approach of Camuffo et al. (2023)." | "Our empirical strategy builds on the large-scale replication approach of Camuffo et al. (2023), while incorporating additional measures of entrepreneurial decision-making processes." | 💰💰💰 | 🔮🔮🔮 | | Empirics | 🔍 Zoom In: Heterogeneity Analysis | Investigate how the effectiveness of the proposed approach varies across different types of entrepreneurs and business contexts | Nuanced understanding of when and for whom the proposed approach is most effective | "We extend Camuffo et al.'s (2023) analysis by examining heterogeneous treatment effects across various entrepreneurial contexts and founder characteristics." | "Our analysis of heterogeneous treatment effects complements Camuffo et al.'s (2020) initial findings on the differential impacts of scientific approach adoption, providing a more granular understanding of its efficacy." | 💰💰 | 🔮🔮 | pozen | Actionable Item | Theory | Empirics | Citation in Introduction | Citation in Literature Review | | ---------------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Integrate pre- and post-entry learning into a unified model of entrepreneurial decision-making | Develop a theoretical framework that connects pre-entry learning, entry decisions, and post-entry performance | Analyze longitudinal data on entrepreneurs' beliefs and performance before and after market entry | "Building on Chen et al. (2018), we propose a unified model of entrepreneurial learning that spans both pre- and post-entry phases." | "Chen et al. (2018) demonstrate that integrating pre- and post-entry learning can explain several empirical regularities in entrepreneurship without invoking behavioral biases." | | Model the impact of behavioral biases on entrepreneurial learning and decision-making | Extend the Bayesian learning model to incorporate overestimation and overprecision biases | Conduct experiments or surveys to measure entrepreneurs' biases and correlate with entry/exit decisions | "We extend the work of Chen et al. (2022) by examining how behavioral biases interact with learning processes in entrepreneurial teams." | "Chen et al. (2022) show that certain combinations of biases, such as optimism and underprecision, can lead to near-optimal performance in entrepreneurial settings." | | Examine the dynamics of bias in entrepreneurial populations over time | Develop a model of how selection processes (entry and exit) affect the distribution of biases in the population of entrepreneurs | Collect panel data on entrepreneurs' biases and track changes over time as firms enter and exit the market | "Following Chen et al. (2018), we investigate how the distribution of behavioral biases in the entrepreneurial population evolves due to market selection." | "Chen et al. (2018) predict that the prevalence of certain biases among entrepreneurs may change over time due to entry and exit dynamics." | understanding the behavioral biases as computational rationality ## 🏭 prob.prog # Computational Rationality, Marr's Levels, and Pivoting in Startup Investment | Level | Computational Rationality | Marr's Three Levels in Investment | Three-Level Pivots | | ----------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Computational | **Objective:** Maximize E[U(i,c,M,s,f)]<br><br>**Example:** Founder chooses SAFE terms (i=$500k, c=$5M cap) to maximize expected future ownership (60%) given market conditions (M), seed investor (s), founder (f) | **Objective:** Optimize founder's share × price per share<br><br>**Example:** Founder aims for 20% ownership at $100M valuation, balancing dilution with growth potential | **Objective:** Maximize E[payoff] across pivots<br><br>**Example:** Founder evaluates pivots on valuation cap ($5M vs $10M), investment amount ($500k vs $1M), and investor selection (VC A vs VC B) | | Algorithmic | - Metalevel decision-making for resource allocation<br>- Approximate Bayesian inference for belief updating<br>- Switching between model-based and model-free strategies | - Perceiving: Break down SAFE terms into components (investment, cap, etc.)<br>- Probabilistic Reasoning: Model scenarios of future valuations and dilution<br>- Planning: Develop negotiation strategy based on modeled outcomes | - Test2Choose1 strategy for efficient exploration<br>- Sequential hypothesis testing for pivot evaluation<br>- Multi-armed bandit algorithms for balancing exploration and exploitation | | Implementation<br>(probabilistic program) | - Neural networks for valuation modeling<br>- Monte Carlo simulations for scenario analysis<br>- Reinforcement learning for strategy optimization | - NLP models for parsing term sheets and investor communications<br>- Bayesian networks for modeling interdependencies in SAFE terms<br>- Visualization tools for scenario comparison | - A/B testing frameworks for pivot experiments<br>- Rapid prototyping tools for implementing pivots<br>- Analytics dashboards for tracking pivot performance | Note: The principles in the second column (Marr's Three Levels in Investment) can be applied to both the first column (Computational Rationality) and the third column (Three-Level Pivots). For example, the perceiving, probabilistic reasoning, and planning steps can be used to break down and analyze both the utility maximization problem in computational rationality and the pivot evaluation process in the three-level pivots approach.