Pricing Decision Lite — Robust Price Selection Under Uncertainty
Live demo (embedded)
- Open in new tab: https://pranavsharma-pricing-decision-lite.hf.space
- Hugging Face page: https://huggingface.co/spaces/PranavSharma/pricing-decision-lite
- GitHub repo: https://github.com/Sharma-Pranav/pricing-decision-lite
What you can do in the demo (30 seconds)
- Choose a data mode (Synthetic vs observational-mode workflow).
- Review the profit distribution across candidate prices.
- See the naïve optimum vs the robust (governed) recommendation.
- Inspect downside risk (e.g., low-quantile profit) and decision status:
- OPTIMIZE (deploy a robust price)
- HOLD (insufficient leverage / too fragile)
- NO-GO (no feasible price meets governance)
What this proves
- The price that maximizes expected profit is often fragile under uncertainty.
- Downside-aware governance changes decisions: it can shift the price or block deployment.
- Pricing decisions are regime-conditional (clean synthetic assumptions ≠ noisy observational reality).
Decision logic (high level)
```mermaid
flowchart TD
A[Inputs: price grid, cost, demand model] --> B[Bootstrap uncertainty
elasticity + demand]
B --> C[Profit distribution per price]
C --> D{Governance checks
downside quantiles + thresholds}
D -->|Pass| E[OPTIMIZE
choose robust price]
D -->|Borderline| F[HOLD]
D -->|Fail| G[NO-GO]
````
Downloads
Slides