Sankey diagrams are one of the most powerful tools in the data visualization toolbox. With their wide, flowing streams, they have an almost magical ability to tell stories about complex systems — from a country’s budget to the journey of users on a website.

At Datasketch, we’ve worked with many people who want to tell stories with data, and we’ve seen how a well-designed Sankey can make all the difference. But we’ve also identified the 5 most common mistakes when designing them — and more importantly, how to fix them.

1. Inaccurate Proportions

This one might seem obvious, but it’s more common than you’d think. If the flows don’t match their values in width, the chart is misleading. It loses its meaning. The core principle of a Sankey is that the thickness of the streams reflects their magnitude.

The fix: double-check the link between your data and your graphic. Make sure there are no formatting errors in your value column, and verify in your visualization tool that you’ve selected the right column to define the flow widths.

2. The spaghetti effect

A Sankey overloaded with thin, crisscrossed, overlapping flows quickly turns into visual chaos. Instead of seeing a system, you see a mess — like someone spilled a plate of pasta across the screen.

The fix: less is more. If you have many minor categories that don’t add much to your main message, group them under a single node like “Other” and highlight the more relevant flows. This cuts down on visual noise and keeps the reader focused on the core story. Need to show full detail? Use an interactive version where users can explore parts of the diagram at their own pace.

3. Illogical node order

One common mistake is letting nodes (the columns that define each stage) appear in default order — usually alphabetical. The result is immediate disorder: flows cross chaotically throughout the chart, making it hard to follow any meaningful narrative.

The fix: start with your data. Before visualizing, sort your source and destination columns from highest to lowest value. This small step reduces unnecessary line crossings and results in a cleaner, more natural flow that guides the reader’s eye through the story.

4. Color without purpose

The problem? Colors that are too similar, too vibrant, or totally random. Or on the flip side, using just one color for everything. Either way, it becomes difficult to follow and interpret the flows.

The fix: assign colors by category or flow type, and be consistent. Use soft tones for general content and reserve bold colors for highlighting key paths. Tools like ColorBrewer are great for choosing palettes that are both harmonious and accessible.

5. Labels That Are Invisible or Overwhelming

A Sankey with poor labeling is like a map with no street names — frustrating and unhelpful. If the text is hard to read or missing altogether, the chart’s communicative power falls flat.

The fix: keep it simple. Label your nodes clearly, and only label key flows. Use abbreviations or icons to keep text short. If your diagram still feels cluttered, make it interactive — let details appear on hover to offer a clean yet rich experience.

Why It Matters

A well-executed Sankey is more than just a beautiful graphic — it’s a powerful tool to reveal patterns, expose inequalities, or explain decisions. Done right, it can turn a dense spreadsheet into a compelling, visual story.

And if you’re ready to start exploring this type of chart, you can use tools like those from Datasketch, where building a Sankey is just a few clicks away.