How to Use Decomposition Tree in Power BI

Digging into your data to find out why a number is high or low can feel like a guessing game. The Power BI decomposition tree changes that, transforming deep-dive analysis from a chore into an intuitive, interactive exploration. This article will show you exactly how to build and use this visual to uncover the stories hidden in your data.

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 a Decomposition Tree?

A decomposition tree is an interactive visual in Power BI that allows you to break down, or "decompose," a metric into its contributing parts. Think of it as a tool for performing root cause analysis on the fly. Instead of needing to build a rigid, predefined hierarchy, you can explore different dimensions in whatever order makes sense to you, letting your curiosity guide the analysis.

For example, if you see that your total sales are $10 million, you can use a decomposition tree to ask follow-up questions visually:

• Which sales region contributed the most to that $10 million?
• Within that top region, which product category was the bestseller?
• For that specific product category, which salesperson closed the most deals?

The tree lets you click through each level, instantly visualizing the breakdown at each step. This process helps you understand how different factors collectively influence a key performance indicator (KPI).

Key Benefits of Using a Decomposition Tree

• Ad-Hoc Exploration: It's perfect for when you don't know the exact path of your analysis beforehand. You can explore different combinations of dimensions without having to reconfigure the visual.
• Built-in AI: The visual includes AI-driven "splits" that can automatically find the highest or lowest value in your data, helping you quickly identify major contributors without manual guesswork.
• User-Friendly: It’s one of the most intuitive ways for non-technical users to conduct detailed drill-down analysis, making data exploration accessible to a wider audience.

When Should You Use a Decomposition Tree?

While versatile, the decomposition tree shines in specific scenarios where you need to understand the composition of a total value. It’s ideal for diagnosing problems, pinpointing opportunities, and simply satisfying your curiosity about what drives your results.

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

Common Use Cases

1. Sales Performance Analysis This is a classic use case. You can start with "Total Revenue" and decompose it to answer critical business questions.

• Path 1: Total Sales → By Region → By Salesperson → By Product Category
• Path 2: Total Sales → By Product Category → By Customer Segment → By Month
• Path 3: Use AI splits to find the highest-performing salesperson, then see which products they sold the most.

2. Marketing Campaign Analysis Start with a key marketing metric like 'Website Sessions' or 'Ad Spend' and break it down to see what's really working.

• Path 1: Total Website Sessions → By Marketing Channel (Organic, Paid, etc.) → By Landing Page → By Device (Desktop, Mobile)
• Path 2: Total Ad Spend → By Platform (Facebook, Google) → By Campaign → By Ad
• Path 3: Begin with 'Total Leads Generated' and use the 'High value' split to instantly see which campaign drove the most conversions.

3. Financial Analysis Dive into financial statements to understand spending patterns or revenue sources without getting lost in spreadsheets.

• Path 1: Total Expenses → By Department → By Expense Category (Software, Travel) → By Vendor
• Path 2: Total Revenue → By Service Line → By Client → By Quarter

4. Operational Monitoring Identify inefficiencies or bottlenecks in your processes by breaking down operational metrics.

• Path 1: Total Shipping Costs → By Warehouse → By Carrier (FedEx, UPS) → By Shipping Service (Ground, Air)
• Path 2: Total Support Tickets → By Product → By Issue Type → By Support Agent

How to Create a Decomposition Tree in Power BI: A Step-by-Step Guide

Building a decomposition tree is surprisingly straightforward. Let's walk through the process using a sample sales dataset.

Step 1: Get Your Data Ready

First, make sure your data is set up correctly in Power BI. You'll need at least one metric to analyze (a number you can sum, average, etc.) and several categorical dimensions to explain it. For our example, we'll use a simple dataset with columns like Sales, Product Category, Region, and Salesperson.

Step 2: Add the Decomposition Tree Visual

In the Power BI report view, navigate to the Visualizations pane on the right-hand side. Find the decomposition tree icon — it looks like a branching diagram. Click it to add the visual to your report canvas. Once you’ve selected it, a placeholder for the visual will appear on your canvas. Now, we need to populate it with data.

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: Add Fields to the 'Analyze' and 'Explain by' Buckets

With the visual selected, look at the Visualizations pane again. You'll see two key input fields (also called buckets):

1. Analyze: This is where you put the primary metric you want to break down. This value must be a measure or a column that can be aggregated (like a sum, average, or count). For our example, drag your Sales field into this bucket. Power BI will automatically default to 'Sum of Sales'. 2. Explain by: This is where you add all the dimensions you might want to use to drill down. You can add several fields here, and the order doesn't dictate a fixed hierarchy — it only sets the suggested order for exploration. Drag Region, Product Category, and Salesperson into this bucket.

After adding your fields, your visual will show a single bar representing your total sales, along with a small plus (+) sign next to it. This is your starting point for exploration.

Step 4: Interact with Your Decomposition Tree to Find Insights

This is where the magic happens. Click the plus (+) sign next to the total sales bar. A dropdown menu appears, giving you options for how to split the data.

Using the AI Splits (High value and Low value)

At the top of the menu, you'll see options like "High value" and "Low value". This is the AI feature.

• If you select High value, Power BI will automatically search through all your 'Explain by' dimensions (Region, Product Category, Salesperson) and find which one factor and which value within that factor contributed most to your total sales. It might, for instance, open a branch showing that the "West" region drove the highest sales.
• If you select Low value, it does the opposite, showing you the lowest-performing branch.

This is incredibly powerful for guided analysis. Instead of clicking randomly, you can let Power BI point you directly toward the most significant drivers in your data.

Choosing a Specific Dimension

If you have a particular question in mind, you can ignore the AI splits and choose a specific dimension from the list below them. For example, if you want to know how sales break down by product first, simply click on Product Category in the menu. The tree will expand to show a bar for each category, sized according to its contribution to total sales. You can continue this process by clicking the plus sign on any of the new branches. For instance, after splitting by Product Category, you could click the plus sign next to "Electronics" and then choose "Salesperson" to see who sold the most electronics.

Locking and Unlocking Levels

As you explore, you might find an analysis path you want to save. You can "lock" a level by clicking the small lock icon that appears on the header of that level. Once locked, that level cannot be removed until you unlock it. This is useful for preventing accidental changes as you explore other branches of the tree.

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

Best Practices and Tips

To get the most out of the decomposition tree, keep these simple guidelines in mind:

1. Don't Overload 'Explain by': While it's tempting to add every possible dimension, doing so can make the drill-down menu cluttered and confusing. Stick to 5-7 highly relevant fields that are logical for the analysis.
2. Start with AI Splits: Always try using 'High value' or 'Low value' for your first split. It often uncovers insights or highlights areas of interest you might have otherwise missed.
3. Use Clear, Well-Defined Measures: Ensure the metric you place in the 'Analyze' bucket is unambiguous. A simple 'Sum of Sales' is great, but a more complex DAX measure like '[YTD Revenue]' works perfectly too.
4. Combine it with Other Visuals: The decomposition tree is fantastic for exploration, but it works even better as part of a larger dashboard. You can use it to filter other visuals. For example, clicking on a node (like a specific region or salesperson) can filter cards, tables, and charts on the rest of the report page to provide more detailed context.
5. Remember to Format: Don't forget the formatting options. In the Format visual pane, you can adjust colors, connector line styles, data labels, and more to enhance readability and align the visual with your brand's style guide.

Final Thoughts

The Power BI decomposition tree is an essential tool for anyone who needs to perform root cause analysis and explore data without restrictions. By allowing free-form drilling on the fly and incorporating helpful AI splits, it empowers you to move beyond simply reporting numbers and start truly understanding them.

While Power BI helps you dig deep into data you've already imported, the initial process of bringing together data from platforms like Shopify, Salesforce, Google Ads, and others can be a huge manual reporting burden. With Graphed, we connect all your scattered marketing and sales sources in one place. You can use simple, natural language to instantly generate the real-time dashboards you need, allowing you to ask questions like "Show me my ad spend vs revenue by campaign" and get a live dashboard back in seconds — no tedious setup required.