Decision tools: Monte Carlo simulations

Plan under uncertainty

You run a retail operation and demand jumps from week to week. Shipping delays shift by the day and your forecasts look good…. until they do not.

You need to know the range of what might happen next.

Monte Carlo simulations can help. They give you a way to model uncertainty. Instead of betting on a single outcome, you run the numbers across thousands of possibilities.

You get a distribution, not just a guess.

This is part of “Decision tools,” a series that breaks down analysis techniques for leaders. They will help you allocate resources, build realistic forecasts, and test strategy before investing.

Monte Carlo is one of the most versatile.

---

In simple terms

Monte Carlo simulations use randomness to explore what might happen. You define your model, assign distributions to the inputs, then simulate outcomes over and over and over and over...

Instead of saying, “Revenue next month will be $2M,” you say, “Given what we know about price, volume, and seasonality there is an 80% chance revenue lands between $1.7M and $2.3M.”

The power is in the range. The insight is in the risk and confident.

Let’s say you sell 1,000 units a week. Your cost per unit also fluctuates. With Monte Carlo you can simulate profits across all combinations of demand and cost.

That gives you a full picture. Not just best case or worst case.

---

Real world

Inventory planning

You manage inventory across 20 stores. Stockouts hurt revenue and overstock kills cash. You model demand as a range. You factor in lead times and supplier behavior.

Monte Carlo shows you the reorder points that minimize both risks.

Budgeting

You are setting next quarter’s forecast. You know your top-line drivers are variable with ad spend returns, hiring ramp, supply costs. Simulate thousands of scenarios.

Now you know your 95% confidence interval.

Capital investment

You are evaluating a $10M expansion. Material costs are volatile and construction timelines are uncertain. Monte Carlo helps you model ROI under a range of inputs.

You get a full risk curve. You make a better call.

---

Start simply

Pick a problem with uncertainty and sales, supply chain, pricing. Define the inputs that vary.

Play around with it, you do not need to be exact. Use historical data where you can. Use judgment where you can’t.

A few common ones:

• Normal: for things like daily demand or time on site.

• Uniform: for equal-likelihood ranges.

• Triangular: for expert estimates.

Now simulate. Draw random values from each input’s distribution. Plug into your model. Repeat 1,000 times.

Here’s a Python example:

import 

numpy as np  
demand = np.random.normal(1000, 200, 1000) 
cost = np.random.uniform(5, 8, 1000) 
price = 10  

profit = (price - cost) * demand

This runs 1,000 scenarios. You now have a distribution of profits.

In Excel, use RAND() or NORM.INV() to build the same.

---

Why this matters

You do not need perfect forecasts. You need to understand the range with the risk and spectrum of outcomes.

Monte Carlo gives you that. You stop anchoring on averages. You start making decisions based on distributions.

---