The Mathematics of Entrepreneurship: Modeling New Venture Creation Through Diffusion Control
- hihmtaj
- 16 minutes ago
- 4 min read
Startups are often portrayed as chaotic journeys of intuition and iteration. Yet a seminal Stanford study by Zhengli Wang and Prof. Stefanos Zenios (2020) reframes entrepreneurship as a quantifiable drift-variance diffusion control problem. In their model, a venture’s evolution follows a Brownian motion, where the entrepreneur strategically adjusts the drift (expected progress) and variance (uncertainty) through costly, controllable actions.
The Model in Plain Language
Wang and Prof. Zenios (2020) depict the state of a new venture as a stochastic diffusion process governed by two decision variables:
Drift (μ): the rate of progress toward success;
Variance (σ²): the volatility or unpredictability of outcomes.
The entrepreneur chooses between two controls, each defined by a unique combination of drift, variance, and cost. Each control has a constant per-unit-time cost hᵢ > 0. When the process reaches an upper boundary (U), the venture succeeds and yields a reward M; when it hits a lower boundary (L), the venture fails.
This abstraction parallels real-world startup milestones - such as customer-acquisition thresholds or regulatory approvals and shows how strategic intensity can shift as uncertainty resolves.
Dynamic vs. Static Strategy
The study’s central result, Theorem 1 - proves that the optimal policy switches between the two controls at most once. Entrepreneurs therefore operate within one of four regimes:
Always use Control 1 (safe, low variance);
Always use Control 2 (aggressive, high variance);
Use 1 then 2 as performance improves;
Use 2 then 1 when near failure (only under fixed boundaries).
This finding formalizes the intuition behind “pivot vs. scale” timing - entrepreneurs must identify a single, critical inflection point rather than continuously oscillate between strategies (Wang & Prof. Zenios, 2020, Theorem 1, pp. 10–12).
Visual Insight - Optimal Policy Map Figure 4 (Wang & Prof. Zenios, 2020) visualizes the single-switch policy. The four color-coded regions correspond to the strategies “Always Use 1,” “Always Use 2,” “Use 1 Then 2,” and “Use 2 Then 1.” It clearly depicts how only one switch between controls is ever optimal.
The “Triumph of More Expensive Controls”
A counter-intuitive result, highlighted in Section 4, shows that the costlier control may dominate when it carries a substantially higher effective drift (2μ / σ²). In practice, founders sometimes over-invest in expensive growth actions to avoid the lower boundary - akin to “buying survival probability.” Wang and Prof. Zenios (2020, pp. 16–17) term this the “triumph of more expensive controls.” This effect arises under fixed lower boundaries and disappears when the lower boundary is free.
Free vs. Fixed Boundaries: Knowing When to Quit
When the lower boundary is fixed - such as investor-imposed deadlines - entrepreneurs may behave inefficiently, over-exerting costly efforts merely to delay failure. However, when the boundary is free, the entrepreneur can choose when to abandon the venture.
Theorem 3 (pp. 19–20) defines the abandonment region: quitting is optimal when both controls are strictly inefficient, i.e., when their costs h exceed the “modest inefficiency” thresholds h⁺. This reframes failure not as cash depletion but as the point where no available control yields a non-negative expected value.
Visual Insight - Abandonment and Flexibility Figures 6 and 7 (Wang & Prof. Zenios, 2020) illustrate how boundary conditions influence strategy.
• Figure 6 (fixed boundaries): the curve divides regions where entrepreneurs abandon early versus continue after hitting 0.
• Figure 7 (free boundaries): when the boundary becomes flexible, the feasible region expands- showing how autonomy improves survival and reduces inefficiency.
Visual Insight - Sensitivity and Efficiency To validate these dynamics quantitatively, Figures 9 and 10 present sensitivity analyses.
Figure 9 plots how expected cost and stopping time change as the switching threshold S moves from its optimal value S*. The curves show that deviating too far from S* increases both cost and time, confirming the precision of the one-switch policy.
Figure 10 compares dynamic vs static policies, revealing that dynamic control consistently minimizes total cost across boundary conditions.
Together, these charts demonstrate the measurable economic advantage of adaptive control.
Managerial and Investor Implications
Section 8 of the paper translates the mathematics into actionable insight for entrepreneurs and venture investors:
Quantify Drift and Variance Early. Identify which activities accelerate progress (high μ) and which reduce uncertainty (low σ²).
Adopt Dynamic Control Policies. Optimize for one strategic switch - testing early, scaling later - rather than continuous toggling.
Allow Flexible Boundaries. Overly rigid milestones induce wasteful behavior; granting founders flexibility improves efficiency.
These implications mirror lean-startup logic but rest on formal stochastic-control foundations.
Why It Matters
This study connects entrepreneurial decision-making with stochastic optimization and operations research, providing a theoretical basis for venture strategy design. It bridges intuition and quantitative rigor, showing that success hinges not merely on speed but on adaptive control of uncertainty and cost.
In essence, entrepreneurship is a diffusion process - and every founder is a controller balancing drift, variance, and burn.
Publication: New Venture Creation: A Drift-Variance Diffusion Control Model By Zhengli Wang, Stefanos Zenios Operations Research, September2022 Vol. 70, Issue 5. DOI: https://doi.org/10.1287/opre.2021.2171
SSRN (full paper download available): Wang, Z., & Prof. Zenios, S. (2020). New Venture Creation: A Drift-Variance Diffusion Control Model: https://ssrn.com/abstract=3667132
Acknowledgements
I am grateful to Stanford University Graduate School of Business for providing exposure to the Research Tracks, through which I was introduced to Professor Stefanos Zenios's scholarship. His work has been a source of inspiration and learning for me.
Professor Stefanos Zenios is a renowned scholar of operations and innovation. He is the faculty director of Stanford GSB’s Center for Entrepreneurial Studies and the co-director of the Program in Ecopreneurship, a joint initiative between the GSB and the Stanford Doerr School of Sustainability. He designed and teaches Startup Garage, a popular and rigorous experiential course that each year helps hundreds of Stanford GSB students and executives learn and apply the innovation processes that are at the center of the Silicon Valley ecosystem. He also oversees the Stanford GSB Venture Studio: a vibrant learning facility for Stanford graduate students across all disciplines who want to learn about designing and creating viable, high-impact ventures by testing what they are learning in the classroom. He previously designed and co-taught Biodesign Innovation, a project-based course on designing and launching new medical devices, and is a founding senior author of the Biodesign textbook.
Comments