How to Make a Pareto Chart in Looker with AI

Creating a Pareto chart is one of the most effective ways to identify the biggest drivers behind your business outcomes. Based on the 80/20 rule, it helps you pinpoint the "vital few" causes that generate the majority of your results. This article will walk you through how to build one manually in Looker using table calculations and then show you a much faster way using AI.

What is a Pareto Chart and Why Is It So Useful?

The Pareto principle, or the 80/20 rule, suggests that for many outcomes, roughly 80% of the effects come from 20% of the causes. It’s a powerful tool for prioritization because it visually separates the most significant factors from the less significant ones, helping you focus your resources where they'll have the greatest impact.

You can see this principle everywhere in business:

Sales: About 80% of your revenue often comes from roughly 20% of your customers or products.
Customer Support: Around 80% of customer complaints typically stem from 20% of known issues.
Marketing: Roughly 80% of your conversions will likely come from 20% of your marketing channels or campaigns.
Project Management: About 20% of the tasks in a project will consume 80% of the time and resources.

A Pareto chart visualizes this rule using a combination of a descending bar chart and a cumulative percentage line. Here’s what each part represents:

The bars represent individual values (like revenue per product) and are ordered from largest to smallest. They show the direct contribution of each category.
The line represents the cumulative percentage of the total. It starts at the contribution of the first bar and climbs to 100% as you add the contribution of each subsequent bar.

By finding where the cumulative percentage line crosses the 80% mark, you can instantly see which few categories are responsible for the bulk of your results.

Data Prep: Getting Your Data Ready for Analysis

Before you jump into Looker, make sure your data is structured properly. For a Pareto analysis, you need a simple dataset containing at least two fields:

A dimension that represents the potential causes or categories you want to analyze (e.g., Product Name, Country, Complaint Type, Marketing Channel).
A measure that represents the effect or volume you want to measure (e.g., Total Sales, Number of Support Tickets, Web Sessions, Revenue).

Your goal is to have an Explore in Looker that can produce a clean table with your dimension and the aggregated measure, ready to be visualized.

The Traditional Method: Building a Pareto Chart in Looker Step-by-Step

Creating a Pareto chart in Looker "the old-fashioned way" requires using table calculations to derive the cumulative sums and percentages needed for the chart. It's a precise process, but once you learn the steps, you can apply them to any dataset. Let’s use an example of analyzing Sales Revenue by Product Category.

Step 1: Start Your Explore and Add Fields

In Looker, navigate to the Explore that contains your sales data. Select the dimension for your categories and the measure for your values.

Dimension: Product Category
Measure: Total Sales

Run the query. You should see a simple two-column table with each product category and its associated sales total.

Step 2: Sort Your Data in Descending Order

This is a critical first step for a Pareto chart. The bars must be ordered from largest to smallest. Click the header of your measure column (Total Sales) to sort it in descending order. The product category with the highest sales should be at the top.

Step 3: Calculate the Grand Total

Next, we need to calculate the grand total of your measure, as we’ll use this to determine the cumulative percentage. We’ll use a table calculation for this.

Click the Add calculation button in the Data tab.
Give your calculation a name, like "Grand Total Sales".
In the formula box, enter the following expression to sum all values of your measure:

Replace ${sales.total_sales} with the actual name of your measure field. Click Save. You’ll now have a third column where every row shows the same grand total.

Step 4: Calculate the Running Total (Cumulative Sum)

Now, we need to calculate the cumulative sum of sales as we move down the sorted list. This will power our cumulative percentage line.

Click Add calculation again.
Name this one something like "Cumulative Sales".
Enter the running_total() function in the formula box:

Click Save. This adds a fourth column that shows the sum growing row by row. The final row in this column should match your "Grand Total Sales".

Step 5: Calculate the Cumulative Percentage

With both the cumulative sum and the grand total calculated, we can find the cumulative percentage. This is the value that will be plotted as a line on your chart.

Add one more table calculation.
Name it "Cumulative Percentage".
The formula will divide your "Cumulative Sales" calculation by your "Grand Total Sales" calculation:

After saving, you’ll have a fifth column. Go to the calculation settings for this column and change the Format to "Percent" to make it reader-friendly.

Your data table is now complete! You can hide the "Grand Total Sales" and "Cumulative Sales" columns by clicking the gear icon on the column header and selecting "Hide from Visualization.” You only need the original category, the original measure, and the final cumulative percentage.

Step 6: Configure the Visualization

It's time to build the chart itself.

Open the Visualization pane and select the Column chart type. Looker will automatically treat this as a combo chart if you configure it correctly.
Click the Edit button to open the visualization settings.
Go to the Series tab. This is where you configure how each measure is displayed.
By default, both your "Total Sales" and "Cumulative Percentage" will be shown as columns. Find your "Cumulative Percentage" measure in the list and change its Type to Line.
To make the chart readable, the percentage line needs its own axis. Find the Y-Axis setting for the "Cumulative Percentage" series and assign it to the Right.

Finally, go to the Y tab in the visualization settings. Configure your Y-axes. Your left axis should represent Sales, and your right axis (for the percentage line) should have a Minimum value of 0 and a Maximum value of 1. Label them "Total Sales" and "Cumulative % of Total," respectively.

You now have a functional Pareto chart in Looker!

The Simpler Way: Creating Charts with Natural Language AI

As you can see, the manual process in Looker is powerful but involves a lot of precise steps. You have to remember the right table calculation functions, sort the data correctly, and navigate multiple visualization menus to configure the combo chart. For teams moving quickly, this can feel slow and cumbersome.

Modern AI-powered analytics tools eliminate this friction entirely. Instead of clicking through menus and handwriting formulas, you can simply describe the chart you want in plain English. For example, creating a Pareto chart becomes as simple as asking:

“Build a Pareto chart of our sales revenue by product category for the last 12 months.”

The AI handles everything in the background:

It understands the intent: It knows a "Pareto chart" requires sorting, cumulative calculations, and a combo chart setup.
It performs the calculations: All the running_total() and sum() logic happens automatically. You don't write any code.
It builds the visualization: It generates the sorted bar chart, the cumulative line, the dual y-axes, and proper labeling in seconds.

This approach isn't just faster, it's more democratic. A marketing manager, a sales lead, or a founder doesn’t need to become a Looker expert to get the insights they need. Answering a follow-up question is as simple as typing another sentence: "Now filter this to only show results from the United States." This turns data analysis from a technical task into a conversation.

How to Read and Act on a Pareto Chart

Creating the chart is only half the battle. The real value comes from interpreting it correctly to make better business decisions.

Identify the "Vital Few"

Look for the point where the line graph starts to flatten. The steep initial slope represents your high-impact categories - the "vital few." Once the line crosses the 80% mark, you've identified the key contributors. For the categories beyond that point (the "trivial many" or the "long tail"), each one adds less and less value to the cumulative total.

Turn Insights Into Action

After identifying your vital few, the next step is to ask "why?" and "what's next?"

If analyzing sales revenue by product: Can you increase marketing spend on your top 20% of products? Should you feature them on your homepage? Are there supply chain risks associated with them that need attention?
If analyzing sources of website traffic: Double down on the few channels driving 80% of your conversions. Can you allocate more budget or create more content for these high-performing channels?
If analyzing customer support ticket types: Fix the few root problems causing 80% of your support requests. Could better documentation, a product update, or a new feature resolve these common issues and reduce your support team's workload?

The Pareto chart isn’t just for finding problems, it’s for finding opportunities. It gives you the clarity to stop spreading your efforts thinly across everything and start concentrating your resources on what truly matters.

Final Thoughts

Pareto charts provide an incredibly clear view of where to focus your attention for maximum impact. While you can certainly create them manually in a powerful tool like Looker by carefully crafting table calculations and configuring visualizations, the process requires technical precision and time.

We built Graphed to make this kind of essential analysis accessible to everyone, instantly. Instead of spending 15 minutes building calculations and adjusting chart settings, you can connect your data sources and simply ask for what you need in plain English. This allows your entire team to move directly from question to insight, freeing you up to make the strategic decisions that will actually drive your business forward.