How to Do Sensitivity Analysis in Google Analytics

Wondering how a small dip in your conversion rate could impact your quarterly revenue? Sensitivity analysis lets you answer "what-if" questions like this by changing key variables to see the effect on your bottom line. This article will show you how to perform a sensitivity analysis using your Google Analytics 4 data and a simple spreadsheet.

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?

Sensitivity analysis is a modeling technique used to determine how different values of an independent variable will impact a particular dependent variable. In simpler terms, it’s a way to understand which of your metrics have the biggest impact on your ultimate goals.

For a business using Google Analytics, this translates to tangible questions like:

• "If our website traffic from organic search increases by 15%, how much additional revenue can we expect?"
• "How sensitive is our overall lead generation to a 0.5% decrease in the landing page conversion rate?"
• "If our average order value drops by $5, how much more traffic do we need to hit our monthly sales target?"

By exploring these scenarios, you can identify which levers are most powerful for driving growth, which risks you're most exposed to, and where to focus your effort and budget for the best return.

Why GA4 Doesn't Do Sensitivity Analysis (and the Workaround)

Google Analytics is an incredibly powerful tool for understanding what has happened. It tracks user behavior, campaign performance, and conversion events, giving you a detailed historical record. However, it is not designed to be a forecasting or modeling tool.

There is no built-in "Sensitivity Analysis" button in GA4. You can't ask it to project future revenue based on hypothetical changes to your session count or conversion rate. GA4 excels at reporting on the past, not simulating the future.

The good news is, all the raw data you need for the analysis is right there in GA4. The standard workflow to perform a sensitivity analysis is to:

1. Identify the key variables in GA4 that drive your desired outcome (e.g., Sessions and Conversion Rate drive Revenue).
2. Export the historical data for these variables into a spreadsheet.
3. Build a simple model in a tool like Google Sheets or Excel to run your "what-if" scenarios.

This approach combines the robust historical data from Google Analytics with the flexible modeling capabilities of a spreadsheet.

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 1: Get Your Baseline Data from Google Analytics 4

Before you can model different scenarios, you need a solid baseline. This means pulling your average performance data from GA4 for a recent, representative time period (like the last 90 days).

For an e-commerce website, our core model will be:

Revenue = Sessions &times, E-commerce Conversion Rate &times, Average purchase revenue

So, we need to find these three metrics in GA4. Here’s how.

1. Navigate to your GA4 property and go to the Explore section.
2. Start a new Blank exploration.
3. Set a date range. Choose something that represents your typical business operations. The "Last 90 days" is often a good choice.
4. In the Variables column, click the "+" sign next to Metrics.
5. Search for and import the following metrics:
6. Drag these four metrics from the Variables column over to the Values section under Tab Settings.

Your exploration canvas should now display a simple table with the total values for your selected metrics over the chosen date range. This is your baseline. For this example, let's say our data looks like this:

• Sessions: 150,000
• E-commerce conversion rate: 1.75%
• Average purchase revenue: $85.00
• Total revenue: $223,125

Step 2: Build Your Sensitivity Analysis Model in Google Sheets

Now that we have our baseline numbers, we can build the model. You could export this GA4 data, but since it's just a few numbers, it's easier to simply open a new Google Sheet and type them in.

Setting Up Your Baseline Model

In a clean sheet, set up a small section for your inputs and the calculated output. This makes your model easy to read and manage.

1. Label your baseline inputs. For example:
2. In cell A6, type "Calculated Revenue". In B6, enter the formula that multiplies your inputs:

=B2* B3* B4

This formula should return $223,125, matching the Total Revenue from your GA4 report and confirming your model is correct.

Your spreadsheet should look something like this:

[No visual, but layout is as described above]

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

Creating a One-Variable Analysis Table

First, let’s test how sensitive our revenue is to changes in a single metric: the e-commerce conversion rate. We’ll build a table to see how revenue changes if our conversion rate fluctuates from -20% to +20% of its current value.

1. Create headings for your table. For instance, in D1, type "Change in CR", in E1, type "New Conversion Rate", and in F1, "Projected Revenue".
2. In the "Change in CR" column (column D), list the different percentage changes you want to test (e.g., -20%, -15%, -10%, ..., +15%, +20%).
3. In the "New Conversion Rate" column (column E), calculate the resulting conversion rate for each scenario. In cell E2, enter the formula and drag it down:

=$B$3*(1+D2)

(Using $ around B3 locks the cell, so it always refers to your baseline conversion rate as you drag the formula.)

1. In the "Projected Revenue" column (column F), calculate the revenue for each new conversion rate. In cell F2, enter the formula and drag it down:

=$B$2*E2*$B$4

This formula uses your baseline sessions and average revenue but multiplies them by the new conversion rate from column E.

You now have a table showing exactly how much revenue you stand to gain or lose with small changes to your conversion rate. You can instantly see that a 10% improvement in conversion rate (from 1.75% to 1.93%) would generate over $22,000 in additional revenue.

Creating a Two-Variable Analysis Matrix (The Real Power)

This is where sensitivity analysis truly shines. A two-variable matrix lets you see the combined impact of changes in two metrics at once. Let's analyze how revenue changes with fluctuations in both Sessions and E-commerce Conversion Rate.

1. Set up a grid. Across the top row (starting in B9, for example), list the percentage changes for one variable, like Sessions (-15%, -10%, -5%, 0%, 5%, 10%, 15%).
2. Down the first column of the grid (starting in A10), list the percentage changes for the second variable, your Conversion Rate (-15%, -10%, etc.).
3. Now, we'll fill the grid. In cell B10, we need a formula that calculates the revenue based on the session change in its column header (B9) and the conversion rate change in its row header (A10). The formula is:

=($B$2*(1+B$9)) * ($B$3*(1+$A10)) * $B$4

Let’s break down the $ usage, as it's very important:

• $B$2, $B$3, $B$4 are anchored to your baseline values so they never change.
• B$9 locks the row, so as you drag the formula down, it always refers to the session change percentage in row 9.
• $A10 locks the column, so as you drag the formula across, it always refers to the conversion rate change in column A.

1. With that formula in cell B10, you can now drag it across the entire row and then drag the whole row down to fill your matrix.

For even greater clarity, select your matrix of revenue figures and apply conditional formatting (found under Format → Conditional formatting). Use a color scale from red to green. This instantly visualizes your "hot" and "cold" zones, showing you where the biggest opportunities and risks are.

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

How to Interpret Your Analysis

The matrix you just built is more than just a grid of numbers, it's a strategic map. From our e-commerce example, you can now answer complex strategic questions with numbers to back them up.

You can see how a 5% drop in traffic (perhaps from a Google algorithm update) could be offset by achieving a simultaneous 5% increase in your conversion rate (perhaps from a site optimization project). You can also see the exponential impact of improving both at the same time — a 10% increase in both Sessions and Conversion Rate doesn't just increase revenue by 20%, but by 21% from the combined effects.

The goal is to move your performance from the center of the chart (your baseline) towards the green area in the top right. This analysis shows you which of those two axes - traffic or conversion - gives you the most powerful lift for the effort involved.

Final Thoughts

Sensitivity analysis is a powerful way to move from simply reporting historical data to modeling future possibilities. By pulling your baseline metrics from Google Analytics and using a spreadsheet to run "what-if" scenarios, you can uncover which drivers have the biggest influence on your core business goals and make smarter, data-informed decisions.

We know that manually exporting data and rebuilding spreadsheet models can be tedious, taking you away from the strategic thinking that actually drives growth. That friction is why we built Graphed. After a one-time setup connecting your Google Analytics account, you can skip the spreadsheets entirely and ask questions in plain English — like "model the impact of a 15% increase in ad spend on revenue" or "show me our most profitable campaigns" — and get instant dashboards and answers. It handles all the data wrangling so you can spend less time pulling reports and more time acting on insights.