What-If Analysis in Excel for Ecommerce (2026)

A what-if analysis lets you test a business decision before making it. You can change inputs like price, conversion rate, or shipping cost to see how they affect revenue, profit, and other business results. Excel includes three built-in tools for what-if analysis: Scenario Manager, Goal Seek, and Data Table.

KPMG’s 2025 survey of 208 US C-suite leaders found that 52% of US organizations haven’t integrated risk and resilience capabilities into their planning, and about half regularly use advanced analytics such as scenario analysis and predictive modeling. What-if analysis gives teams a way to test how sudden changes or disruptions could affect the business before they have to respond in real time.

This guide walks through what a what-if analysis is, how to build one in Excel, and how to use it in ecommerce planning.

What is a what-if analysis in Excel?

A what-if analysis is used to model how changes in inputs or variables affect business outcomes. You set up a formula that links inputs like price, units sold, or conversion rate, to an output, like revenue or profit, then change those inputs to see how the output responds. There are three ways to do this in Excel: Scenario Manager, Goal Seek, and Data Table.

For an organization, this connects directly to scenario modeling. Pricing decisions, demand forecasts, inventory levels, and shipping costs all move together, and a what-if analysis lets a finance or ecommerce team see how a change to one input affects revenue or margin before that change happens in the real world. Instead of adjusting a price and waiting a quarter to see the effect, you can model it first.

A what-if analysis in Excel can answer questions like:

• When should inventory be discounted to maximize revenue during a peak season?
• What conversion rate is needed to hit a specific revenue goal?
• How many units need to sell each month to stay profitable?

Each of these comes back to the same underlying question: What happens to a business outcome if a specific input or key performance indicator (KPI) changes?

How to conduct a what-if analysis in Excel

Excel offers three ways to run a what-if analysis, each suited to a different kind of question:

Before building any of these models, pull real inputs instead of estimating them. Shopify Analytics reports provide average order value, conversion rate, and sales by product, all of which can be used in the formulas below.

Scenario Manager

Scenario Manager is best for comparing multiple complete scenarios at once, like best-case, worst-case, and expected-case models for pricing, demand, shipping costs, or inventory levels. Rather than testing one variable at a time, it lets you build several named scenarios, each with its own set of input values, and compare the resulting outputs side by side.

For example, to model how different discount percentages affect revenue during a sale, start with a table that defines the key figures:

• The number of products in stock
• The percentage of products sold at full price
• The full price
• The discounted price
• The number of products sold at full price (formula: B3*C3)
• The number of products sold at a discount (formula: B3*(1-C3))
• Total revenue based on those prices and volumes (formula: C7*D7+C8*D8)

You define the first four figures directly. Excel calculates the remaining three.

With that table in place, open the Data tab, select What-If Analysis, then Scenario Manager. 

Click Add to create a scenario, name it (for example, “60% sold at full price”), and set the Changing cells field to the percentage cell. 

Enter 0.6 as the value, then click OK.

Repeat this for each scenario you want to compare. You might also look at 70%, 80%, 90%, and 100% sold at full price.

Selecting a scenario and clicking Show updates the table with that scenario’s values. The most useful part of the feature is Scenario Summary: Click Summary, select the cell containing total revenue as the result cell, then click OK.

From there, Excel generates a new sheet comparing every scenario side by side.

Keep in mind that changing numbers in the source table after generating a summary won’t update that summary. To test a new set of inputs, like a different inventory size or discount depth, you’ll need to generate a fresh summary. This gives you multiple, comparable snapshots of how pricing and inventory decisions affect revenue.

Goal Seek

Goal Seek is a way to work backward from a target using one input. It takes a target outcome and works backward to find the single input value needed to reach it. 

Where Scenario Manager compares several complete scenarios, Goal Seek answers a single question, like what conversion rate, traffic level, or average order value (AOV) is needed to reach a specific revenue number.

To set this up, first build a table with the key metrics: website visitors, conversion rate, average order value, and a revenue formula that ties them together (for example, B3*C3*D3).

Use input values close to actual visitor counts, conversion rates, and AOV for the most accurate result. 

On the Data tab, click What-If Analysis, then Goal Seek. Set the Set cell field to the revenue cell, enter the target revenue in To value, and choose the input to change in By changing cell. 

For example, changing visitors to hit a $7,000 revenue target returns a figure of 2,333 visitors, holding conversion rate and AOV constant.

Run the same process against conversion rate or AOV instead of visitors to see which gives you the most realistic path to the same revenue goal. This helps prioritize whether your team should focus on traffic acquisition, conversion rate optimization (CRO), or upselling.

Data Table

Data Table maps outcomes across one or two variables at a time, unlike Scenario Manager, which handles up to 32 changing values in fixed, named combinations. 

A one-variable Data Table shows how an output changes across a single input, like margin at different price points. A two-variable Data Table shows every combination of two inputs, like unit sales and price, and how each combination affects a shared output.

This makes Data Table well suited to profit forecasting, similar to how the high-low method separates fixed and variable costs. It models how monthly profit moves as unit sales and profit-per-unit change, or how a discount and its resulting sales lift interact. 

First, set up the formula: Profit per month = (Units sold x Profit per unit) - Fixed costs, or (C4*C3)-C2.

Duplicate the profit-per-month formula into the top-left cell of the table you’re about to build. In this instance, it’s =C5. This will be the foundation for the Data Table feature.

Below that, list a range of values for units sold (for example, starting at 500 and increasing by 100 per row). 

Across the top, list a range of profit-per-unit values (for example, starting at $10 and increasing by $5 per column).

Select the full table, open the Data tab, click What-If Analysis, then Data Table. Set the row input cell to profit per unit and the column input cell to units sold, then click OK.

The completed table shows profit at every combination of units sold and profit per unit. This is useful for setting discount floors, forecasting break-even points, or evaluating how a cost increase from a supplier changes the volume needed to stay profitable.

What-if analysis examples for ecommerce

Applying these tools to specific ecommerce situations means defining four things: 

• Which inputs change
• Which output metric you’re tracking
• What decision the model supports
• Where the biggest assumption risk lies

The table below summarizes three common scenarios. Each is broken down in more detail below.

Market shift scenario

Consumers increasingly research and shop through AI assistants and agentic search instead of typing queries into a traditional search engine, changing how traffic arrives at a site and how it converts.

Consider a retailer testing what happens if AI-assistant referrals grow from 5% to 20% of total traffic over the next year, with a conversion rate 15% higher than traditional search traffic due to more qualified, pre-researched visits:

• Input variables: AI-assistant referral share rising from 5% to 20% of traffic, paired with a 15% conversion rate premium over traditional search.
• Output metric: Against $40 million in online revenue, the shift adds an estimated $2.8 million as AI-referred traffic converts at a higher rate and displaces a portion of lower-converting search traffic.
• Decision: Whether to invest in structured product data and AI-discoverability now, estimated at $1.2 million, or wait until the channel represents a larger share of traffic.
• Assumption risk: The model assumes the conversion premium and early-mover advantage hold as AI-assistant shopping becomes more common. That premium is worth testing against actual referral data before committing the full $1.2 million, since it may shrink once AI-discoverability becomes standard practice across the retail sector.

Supply chain disruption

A supply chain what-if analysis should account for both direct costs and the effects on day-to-day operations. 

Ready.gov groups disruption effects into categories like lost or delayed sales, increased expenses, and delayed business plans, which reflect how a shipping disruption can affect an online business.

Model a scenario where a geopolitical event increases shipping costs by 30% and extends delivery times by two weeks:

• Input variables: A 30% increase in shipping costs and a two-week delivery delay.
• Output metric: For a business with $20 million in cost of goods sold, the shipping increase adds $6 million in direct expenses. If the delivery delay causes a 5% drop in sales on $50 million in revenue, that’s $2.5 million in lost revenue, on top of the direct costs.
• Decision: Whether to shift toward local suppliers. A 10% increase to cost of goods sold (COGS) from local sourcing could still save $4 million a year in shipping, for net savings of $2 million, though it would likely require a $3 million investment in supply chain infrastructure.
• Assumption risk: The 5% sales drop from delayed delivery is the least certain assumption in the model. Test it against past order data from a comparable delay rather than treating it as fixed.

Consumer preferences shift 

GlobeScan’s March 2025 survey of US consumers found that 49% had bought an environmentally friendly product in the past month, while another 36% wanted to buy a sustainable product but were held back by price, limited awareness, or availability. 

For a business with an existing shopper base, that unmet demand could provide a starting point for a what-if model:

• Input variables: Shifting 30% of the product line to eco-friendly alternatives, with an up-front $8 million investment in research, development, and marketing.
• Output metric: If eco-friendly products carry a 10% higher profit margin on top of a 20% baseline margin, the shift could add $3 million to annual net income.
• Decision: Whether the projected margin gain and the unmet demand GlobeScan identified justify the $8 million investment, or whether a smaller pilot line could test the assumption first.
• Assumption risk: The 10% margin premium is the variable to test most closely. It depends on sourcing costs for sustainable materials holding steady as the product line scales beyond a pilot.

Benefits of what-if analysis in ecommerce

At a small scale, your business can recover from a minor pricing or inventory decision that doesn’t pan out. But at higher revenue and SKU counts, that same decision carries more risk. 

A what-if analysis gives finance, ecommerce, and operations teams a shared way to test decisions before committing budget to them.

Decision quality

Modeling scenarios ahead of time replaces a guess or blind choice about pricing, margin, inventory, and channel options with a side-by-side comparison of projected outcomes. Instead of picking a discount depth or a new market based on instinct, a team can compare the projected revenue and margin of each option to make a more informed decision.

Stakeholder alignment

A 2025 research report on enterprise risk scenario analysis found that companies ran cross-functional tabletop exercises, drawing on senior managers across the business rather than having a single team model in isolation, and factored in indirect effects like morale, competitive response, and interest rates alongside direct costs.

Ecommerce teams can take a similar approach. Finance, operations, marketing, and leadership all have different assumptions about how a pricing or supply chain change will play out. A shared model brings those assumptions into the open before a decision is made, giving teams a chance to compare them before committing to a plan.

Risk-planning and response speed

More than half of US organizations have not yet integrated risk and resilience capabilities into their planning, even as about half regularly use tools like scenario analysis and predictive modeling. 

When you already have a what-if model in place and a disruption strikes, whether a shipping cost spike or a demand shift, you can update inputs to understand the potential impact without building a model from scratch.

Turn scenario analysis into decisions

A what-if model is only useful when it’s built on your actual data and revisited regularly. A few steps can help put the analysis into practice:

• Pull real data: Export average order value, conversion rate, and product-level sales from Shopify Analytics instead of estimating.
• Match the tool to the question: Use Scenario Manager to compare complete pricing or inventory scenarios, Goal Seek to find the input to reach a single target, and Data Table to map a range of outcomes across one or two variables.
• Assign an owner to each model: If no one owns the model, it’s unlikely to be used regularly, if at all. Assign someone in finance or ecommerce operations to refresh the inputs quarterly or ahead of a peak sales period.
• Revisit assumptions after each disruption: Whenever a disruption occurs, use data on the actual impact to update your model, then compare the results with your original assumptions.

Build a small library of these models for pricing, inventory, channel investment, or other decisions your team regularly models.