How to Do Sensitivity Analysis with ChatGPT

Sensitivity analysis is a simple way to test how your business or financial model reacts to change. By tweaking your assumptions - like 'what if traffic drops by 10% next month?' or 'what if our CPL doubles?' - you can pinpoint which variables have the biggest impact on your bottom line. This article will show you how to use ChatGPT as an intelligent assistant to perform sensitivity analysis, from brainstorming scenarios to generating the necessary formulas or code for visualization.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

What is Sensitivity Analysis, Anyway?

At its core, sensitivity analysis, often called "what-if" analysis, is the process of changing the value of an independent variable to see how it affects a dependent variable. In simpler terms, it helps you understand which of your inputs are the real drivers of your success (or failure).

Imagine you run an e-commerce store. Your profit is dependent on several key inputs:

• Website traffic
• Conversion rate
• Average Order Value (AOV)
• Cost Per Click (CPC) from ads

Sensitivity analysis helps you answer critical business questions like:

• "Which factor moving by 20% would hurt my profit more: a drop in conversion rate or an increase in CPC?"
• "If I want to increase my profit by $5,000, where should I focus my efforts for the biggest return?"

Doing this allows you to build more resilient business models, identify major risks before they happen, and make data-driven decisions with far greater confidence.

How ChatGPT Changes the Game

Traditionally, performing sensitivity analysis involved spending hours in a spreadsheet like Excel or Google Sheets. You’d have to manually set up data tables, remember formulas, build charts, and hopefully not make any calculation errors along the way. For more complex analyses, you might need specialized software or coding skills in Python or R.

ChatGPT doesn’t replace these tools, but it acts as an incredibly smart co-pilot that dramatically flattens the learning curve. You don’t need to be a data scientist or a spreadsheet wizard anymore. You can simply describe what you want to achieve in plain English, and ChatGPT can guide you through every step. It can help you:

• Structure your model: Figure out the core formula and variables.
• Brainstorm scenarios: Suggest optimistic, pessimistic, and base-case values.
• Generate formulas: Provide the exact Excel or Google Sheets formulas to build your analysis.
• Write code: Create Python scripts for more advanced visualizations like Tornado charts.
• Interpret the results: Explain what the numbers and charts actually mean for your business.

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Step-by-Step Guide: Your First Sensitivity Analysis with ChatGPT

Let's walk through a practical example from start to finish. We'll stick with our e-commerce business model to see how profit is affected by different marketing metrics.

Step 1: Define Your Core Model

First, you need a basic financial model. Don't worry, this doesn't have to be overly complicated. All you need is a simple formula capturing the relationship between your inputs and your desired output. Our goal is to calculate profit.

Example Model: Profit = (Monthly Traffic * Conversion Rate * Average Order Value) - Total Ad Spend

Let's also assume:

Total Ad Spend = Cost Per Click (CPC) * (Monthly Traffic * CTR) (where CTR is Click-Through Rate from ads).

For now, you don't even need to open a spreadsheet. You can just bring this concept to ChatGPT to help you clarify everything.

Sample ChatGPT Prompt: "I'm building a simple financial model for my e-commerce store to track profit. The main formula is: Profit = (Traffic * Conversion Rate * AOV) - Ad Spend. What are the key input variables I should consider for a sensitivity analysis? Help me brainstorm some realistic numbers for a base case, best case, and worst case for each variable. I'll need numbers for Traffic, Conversion Rate, AOV, and CPC."

Step 2: Generate Scenarios with ChatGPT

ChatGPT will likely come back with a structured table that makes your life much easier, helping you think through ranges that make sense for your business.

Example ChatGPT Response:

Of course! Here are the key variables and some realistic scenarios for an e-commerce store:

For this analysis, we can vary each of these drivers one by one to see their isolated impact on your overall profit.

This simple step already gives you a clear framework. You now have your variables and the ranges you want to test.

Graphed

Build AI Agents for Marketing

Build virtual employees that run your go to market. Connect your data sources, deploy autonomous agents, and grow your company.

Build Your First Agent

Step 3: Build a One-Way Data Table

A "one-way" sensitivity analysis involves changing one variable at a time while keeping the others constant (at their base case). This is the easiest way to see the direct impact of each factor.

Let’s test the impact of Conversion Rate on profit. Now, you can ask ChatGPT to help you set this up in your spreadsheet tool of choice.

Sample ChatGPT Prompt: _"Great! Now I want to create a one-way sensitivity analysis in Google Sheets. My inputs are in column B: B1: Monthly Traffic = 10,000 B2: Conversion Rate = 2.0% B3: AOV = $50 B4: CPC = $1.00

My profit formula is in cell B6: =(B1B2B3) - (B10.05B4) (assuming a 5% click-through rate from traffic).

Please explain how to set up a data table showing me how the profit in B6 changes when I adjust the Conversion Rate in B2 from 1.0% to 3.0% in 0.25% increments."_

What Happens Next: ChatGPT will guide you step-by-step. It will tell you to list your conversion rate scenarios (e.g., 1.0%, 1.25%, 1.50%, etc.) in a column (e.g., column D). Then, it will tell you to place your output formula =B6 at the top of the adjacent column (e.g., cell E1). Finally, it would explain how to use your spreadsheet's "Data Table" feature (more common in Excel) or simply provide a drag-and-drop formula that references the scenarios, locking certain cell references with $ signs.

(Example Formula for cell E2 you might drag down): =(B1*D2*$B$3) - ($B$1*0.05*$B$4)

AI Agents for Marketing

Crash Course

Graphed

Free PDF · the crash course

AI Agents for Marketing Crash Course

Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.

Download the PDF

Step 4: Go Deeper with Two-Way and Tornado Charts

Sometimes you need to see how variables interact. A "two-way" analysis lets you test two variables simultaneously. For example, what happens when traffic is low AND CPC is high?

This is where Tornado Charts become incredibly useful. A Tornado Chart is a special type of bar chart that helps you visually compare the impact of different variables side-by-side. It immediately shows you the "winners" and "losers" — which inputs have the most and least effect on your bottom line. Creating these can be tricky, but it's a perfect task for ChatGPT.

Sample ChatGPT Prompt for a Python Script: _"I have completed my one-way sensitivity analysis for three variables: Conversion Rate, AOV, and CPC. I calculated the total profit swing (best case profit minus worst case profit) for each:

• Conversion Rate swing: $3,000
• AOV swing: $4,000
• CPC swing: $1,500

Can you write a Python script using the Matplotlib library to generate a simple Tornado Chart that visualizes the impact of these variables on profit? Please make sure the most impactful variable is at the top."_

ChatGPT will provide a complete Python code snippet. You can then copy this code, paste it into a free environment like Google Colab, and run it to produce a professional-looking chart for your reports or presentations. It bridges the gap between a simple question and a technical execution, saving you immense research and debugging time.

Best Practices and Common Mistakes to Avoid

While ChatGPT is a powerful assistant, its analysis is only as good as the instructions and framework you provide. Here are a few tips to get the best results.

• Ask Good Questions First: Before you even start a model, think about what decision you're trying to make. Are you exploring a new marketing channel? Trying to justify a price increase? A clear goal makes your analysis more focused and useful.
• Keep it Simple at the Start: Your first model shouldn't have 20 variables. Start with the top 3-5 drivers you believe have the most influence. You can always add more complexity later. This makes the results easier to interpret and act on.
• Be Explicit with Your Prompts: General prompts lead to general answers. For more precise results, feed ChatGPT the details it needs. Mention the software you're using (Google Sheets vs. Excel), provide cell references, detail your exact formula, and state your desired outcome clearly.
• Check the Work: ChatGPT, like any language model, can make mistakes. It’s a fantastic accelerator, but you are still the pilot. Always do a quick sanity check of the formulas or code it provides. Make sure the logic makes sense and test it with a few simple numbers you can calculate by hand.

Final Thoughts

Sensitivity analysis helps you transform a static, best-guess financial model into a dynamic tool for strategic decision-making. By using ChatGPT as your partner, you can get past the technical hurdles of formulas and code, and focus on asking the right 'what-if' questions that truly drive your business forward.

While ChatGPT excels at helping you structure these kinds of analyses, the process still requires manual setup with static data from spreadsheets. We believe insights should come directly from your live business data, without the need to export files or build models from scratch. That's why we created Graphed. It connects directly to your marketing and sales platforms like Shopify, Google Analytics, and Facebook Ads, allowing you to ask complex questions in plain English and get answers as auto-updating dashboards. You can explore how your conversion rate affects revenue or track which ads drive the most profit in real-time, turning hours of modeling into a 30-second conversation.